WO2002048966A9 - Three-dimensional image streaming system and method for medical images - Google Patents
Three-dimensional image streaming system and method for medical imagesInfo
- Publication number
- WO2002048966A9 WO2002048966A9 PCT/US2001/048933 US0148933W WO0248966A9 WO 2002048966 A9 WO2002048966 A9 WO 2002048966A9 US 0148933 W US0148933 W US 0148933W WO 0248966 A9 WO0248966 A9 WO 0248966A9
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- resolution
- server computer
- dimensional
- transform
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/007—Transform coding, e.g. discrete cosine transform
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/62—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding by frequency transforming in three dimensions
Definitions
- the invention relates generally to the field of medical imaging and, more specifically, to transmission of medical imagery via a relatively narrow bandwidth client-server computer network such as the World Wide Web.
- CT computed tomography
- MR magnetic resonance
- PET positron emission tomography
- SPECT single-photon emission computed tomography
- Data compression techniques can be roughly divided into two classes: lossy and lossless. Lossless compression techniques allow exact reconstruction of the original image, while lossy compression techniques achieve higher compression ratios, because they allow some acceptable degradation.
- lossy compression is acceptable in many cases, lossless transmission is required in medical applications because lossy compression of images can lead to errors in diagnosis by introducing artifacts. Furthermore, there exist several legal and regulatory issues that favor lossless compression in medical applications. Although precise diagnosis require lossless representation of the study, lossy compression can be extremely useful for medical applications in which quick image browsing is desired. Therefore, the combination of lossy and lossless representations in a single workflow is desirable, with progressive lossy streaming of 3-D data terminating at lossless quality.
- 3-D image data can be represented as a set of two-dimensional (2-D) images, it is possible to code these 2-D images independently.
- 2-D lossless compression algorithms such as the LOw Complexity LOssless Compression of Images (LOCO-I) on which a lossless image compression standard called JPEG-LS is base, the Context-based Adaptive Lossless Image Codec (CALIC) algorithm and Compression with Reversible Embedded Wavelets (CREW).
- LOCO-I LOw Complexity LOssless Compression of Images
- CALIC Context-based Adaptive Lossless Image Codec
- CREW Compression with Reversible Embedded Wavelets
- the present invention relates to a 3-D image data streaming system that obviates storing a compressed version of the original data by streaming region-of- interest (ROI) data using the original stored 3-D data.
- ROI region-of- interest
- the system also avoids the computationally intensive task of compression of the full data.
- pixels-on- demand may be used in this patent specification to describe the process of serving the requested ROI and progressively transmitting the ROI data in real-time, where the quality of the view improves towards lossless quality as the transfer process progresses.
- the system allows lossless progressive streaming of 3-D images over the
- a server performs a fast preprocessing step, after which the server quickly responds to requests for viewing specific regions of interest.
- Each request for an ROI triggers an encoding algorithm, but such that does not involve the full 3-D image. Since the size of the ROI is limited by the size and resolution of the viewing device and by number of frames that can be simultaneously viewed at the client and not by the size of the full 3-D image, only a portion progressive coding computation is performed for a local area of the original 3-D data. The same is true for the client computer that performs decoding and rendering only for an ROI and not the full 3-D image.
- the imaging system described herein preferably uses wavelet bases of best performance in streaming ROI data.
- a system and method for transmitting a three-dimensional digital image over a communication network A 3-D digital image is stored on an image storage device accessible via the network.
- a client computer coupled to the network can generate a request for interaction with the 3-D image stored on the image storage device, the request identifying a region of interest within the image.
- the system includes a server computer coupled to the communication network and the image storage device, wherein the server computer performs the steps of: i) performing 2-D sub-band transform decompositions in x-axis and y-axis directions upon the three-dimensional digital image; ii) performing a 1-D sub-band transform decomposition in a z-axis direction upon a portion of the two-dimensionally transform- decomposed digital image; and iii) progressively transmitting to the client computer data representing the identified region of interest within the 3-D image.
- the system can transmit to the client computer data representing thumbnail-resolution images resulting from the 2- D sub-band transform decompositions.
- the system can, in one or more embodiments of the invention, perform a one-dimensional sub-band transform decomposition in a z-axis direction comprising the steps of: a) computing transform coefficients using integer arithmetic and not floating-point arithmetic; and b) correcting the computed transform coefficients by performing integer arithmetic operations with one or more predetermined correction factors.
- the steps of performing sub-band transform decompositions can comprise computing transform coefficients for each of a predetermined plurality of resolutions.
- the transform coefficients for an Nth resolution of the plurality can be computed in response to a one transform formula if N is odd and another transform formula if N is even.
- the server computer can scale the transform coefficients by alternating factors of 2 and 1 along the z-axis, i.e., the resolution axis.
- a system and method for transmitting a 3-D digital image over a communication network in which the request that a client computer generates for interaction with the 3-D image stored on the image storage device specifies a quality threshold and includes a request list specifying data blocks that define a region of interest within the 3-D image.
- the system includes a server computer coupled to the communication network and the image storage device, wherein the server computer, in response to the request list, transmits to the client computer a number of data blocks corresponding to the specified quality threshold.
- the server computer can map the luminance of the three-dimensional image stored on the image storage device from a predetermined image bit depth to a predetermined screen bit depth in accordance with a monotonic luminance mapping function.
- the server computer can do so by, for example, computing a root mean square (RMS) increasing factor defined by a maximal derivative of the luminance mapping function.
- the server computer can provide more data blocks (i.e., more than some predetermined normal number) if the RMS increasing factor is greater than one and a fewer data blocks if the RMS increasing factor is less than one.
- the request generated by the client computer can specify a resolution.
- the server computer provides a number of data blocks corresponding to not only the specified quality threshold but also the specified resolution.
- Figure 1 is a block diagram of an embodiment of the system architecture.
- Figure 2 illustrates an embodiment of the imaging system workflow.
- Figure 3 illustrates subband decomposition and tiling of subband coefficients at given resolution.
- Figure 4 illustrates subband decomposition types.
- Figure 5 illustrates a multi-resolution structure of transformed image data.
- Figure 6 illustrates multi-resolution decomposition of image data.
- Figure 7 illustrates scaling factors of 3-D wavelet coefficients for different sub- bands.
- Figure 8 illustrates tile dimensions and the addition of half-bit information planes for lossless subband decomposition types.
- Figure 9 is a flowchart illustrating a server encoding algorithm outer loop.
- Figure 10A illustrates a server encoding algorithm bit plane significance scan flowchart.
- Figure 10B illustrates a server encoding algorithm, bit plane significance scan flowchart.
- Figure 11 illustrates a server encoding algorithm bit plane significance scan pseudo-code.
- Figure 12A illustrates a least significant bit plane scan pseudo-code encoding algorithm.
- Figure 12B illustrates a half bit plane scan pseudo-code encoding algorithm.
- Figure 13 illustrates a least significant bit plane scan encoding algorithm flowchart.
- Figure 14 is a flowchart illustrating a client decoding algorithm outer loop.
- Figure 15 illustrates a client decoding algorithm bit plane scan pseudo-code.
- Figure 16A illustrates a least significant bit plane scan pseudo-code decoding algorithm.
- Figure 16B is a half bit plane scan pseudo-code decoding algorithm.
- Figure 17 is a flowchart illustrating client "progressive by accuracy” workflow.
- Figure 18 illustrates a curve defining the mapping from Original_Image_Depth- bit image to Screen_Depth-bit.
- Figure 19 is a flowchart illustrating the composition of a "Progressive by accuracy" request list for a ROI according to an embodiment of the invention.
- Figure 20 is pseudo-code for multi-resolution march in the client rendering function.
- Figure 21 is a flowchart illustrating the subband tile extraction function according to an embodiment of the present invention.
- Figure 22 is a flowchart illustrating the server workflow of the present invention.
- Figure 23 is a flowchart illustrating a server preprocessing function of the present invention.
- Figure 24 is pseudo-code for the memory constrained subband tile preprocessing function of the present invention.
- Figure 25 is a flowchart illustrating the low resolution encoding function of the present invention.
- Figure 26 is a diagram illustrating an imaging protocol of communication between client and server.
- Figure 27 is a flowchart illustrating the encoding algorithm principal steps of the present invention.
- Figure 28A is a flowchart illustrating the bit plane significance scan client decoding algorithm.
- Figure 28B is a flowchart illustrating the bit plane significance scan client decoding algorithm.
- Figure 29 is a flowchart illustrating a method for transmitting a 3D digital image in accordance with one aspect of the present invention.
- Figure 30 is a flowchart illustrating a method for transmitting a 3D digital image in accordance with another aspect of the present invention.
- a client computer 110 is coupled to a server computer 120 through a communication network 130.
- both client computer 110 and server computer 120 may comprise a PC-type computer operating with a Pentium®-class microprocessor, or equivalent.
- Each of the computers 110 and 120 may include a cache, 111 and 121 respectively, as part of its memory.
- the server computer 120 may include a suitable storage device 122, such as a high-capacity disk, CD-ROM, DVD, or the like.
- Server 5 computer 120 is programmed with software that effects the processes described below. In view of the description below, persons skilled in the art to which the invention relates will be capable of writing or otherwise providing software suitable for such programming.
- the software can reside in whole or part on a hard disk or similar data medium (not shown) within server computer 120 and can be 0 downloaded onto the disk or otherwise retrieved in whole or part from a remote source (not shown) via a communication network 130.
- the client computer 110 and server computer 120 may be connected to each other, and to other computers, through communication network 130, which may be the Internet, an intranet (e.g., a local area network), a wide-area network, or the like.
- communication network 130 which may be the Internet, an intranet (e.g., a local area network), a wide-area network, or the like.
- Any suitable web browser- type application running on the client computer 110 can be used to connect to the server computer 120.
- the browser software can likewise be stored for execution on a hard disk or similar medium (not shown) within client computer 110 in the conventional manner.
- a client computer 110 begins by requesting to interact with a specific 3-D image at step 201.
- the server computer 120 preprocesses the image at step 202, and notifies the client computer 110 that it is ready to serve the image.
- the client computer 110 continues by requesting a specific ROI at step 203.
- the server computer 120 encodes this ROI and sends its encoded data at step 204.
- the client computer 110 performs progressive rendering of the received data at step 205, and may iterate back for another ROI at step 203.
- the system workflow is now described with reference to Fig. 1.
- the user operating a client computer 110 selects, using common browser tools, a 3-D image residing on the image file storage 122.
- the corresponding URL request is received and processed by the server computer 120.
- the server computer 120 performs a fast preprocessing algorithm (described herein below as step 2201 with reference to Fig. 22).
- the result of this computation is inserted into the cache 121.
- the goal of the preprocessing step is to allow the server computer 120, after a relatively fast computational step, to serve any ROI specified by the user of the client computer 110.
- server software processing step 202 of a sequence of 100 slices 512 by 512 pixels, 16-bit Grayscale each, will typically take 1-2 seconds, running on a computer with a Pentium® processor and the Windows NT® operating system.
- the server computer 120 notifies the client computer 110 that the image is ready to be served.
- the server computer 120 also transmits the basic parameters associated with the image such as dimensions, number or resolutions, etc.
- the client computer 110 can select any ROI of the 3-D image, as indicated by the user through some GUI.
- the ROI is formulated (step 203) into a request list that is sent to the server computer 120. Each such request corresponds to a data block (as described herein below with reference to Figs. 3 and 8).
- the order of requests in the list corresponds to some progressive mode selected in the context of the application. For example, the order can be of increasing accuracy.
- the server computer 120 Upon receiving the ROI request list, the server computer 120 processes the requests according to their order. For each such request the server computer 120 checks if the corresponding data block exists in its cache 121. If not, the server computer 120 then computes the data block, stores it in the cache 121 and immediately sends it to the client computer 110. Once a data block that was requested arrives at the client computer 110, it is inserted into its cache 111. At various moments in time during the transfer process, a decision rule invokes a rendering of the ROI.
- the rendering of the ROI will be of lower quality. But, due to the progressive request order, the rendering quality will improve with each received data block as specified.
- the user can order a change of the ROI, in which case the rendering task at the client computer 110 is canceled.
- a new request list corresponding to a new ROI is then sent from the client computer 110 to the server computer 120 notifying the server to terminate the previous computation and transmission task and to begin the new one.
- the imaging protocol serves as a common language between the server and its clients. It is described with reference to Fig. 26.
- the imaging protocol 3050 is mainly used by the client computer 3010 to efficiently stream to the server an ordered request 3020 list of data blocks required for progressive rendering of a given ROI, and by the server computer 3030 to efficiently compose and send to the client computer 3010 a stream of (requested) data blocks 3040.
- Both client computer 3010 and server computer 3020 use the same distributed database associated with the protocol, such that they are using the same tagging system for the various data blocks. Using this approach it is also possible to cache the data blocks at both ends.
- the server computer 3020 is able to cache any computed data block, such that this data block can be sent to any other client that requests it.
- the client computer 3010 is able to cache any data block received, such that for any subsequent ROI rendering operation requiring the data block, it will be extracted from the client cache and not ordered from the server computer 3020.
- the protocol is designed such that the imaging system will have efficient ROI control and rendering in any mode such as: progressive by accuracy, progressive by resolution and progressive by spatial order.
- the client computer 110 (Fig. 1) generates a request for interaction with the 3-D digital image stored on the image storage device 122. As noted above, the request identifies an ROI within the image.
- a 3-D image can be conceptually described as a sequence along the z- axis direction of x-y frames, analogous to the manner in which a movie is described by a sequence of frames.
- the image is decomposed in the x-y plane and along the z-axis direction.
- the server computer 120 performs 2-D sub-band transform decompositions in x-axis and y-axis directions upon the 3-D image at step 2930.
- each frame is decomposed into four parts: ⁇ HH,HL,LH,LL ⁇ . This is repeated with each resolution step or level, where the LL part of resolution level N is the source image for resolution level N+l.
- Decomposition along the z-axis direction averages pairs of frames together, so with each resolution level the number of f ames is reduced by half. For example, if there is a sequence of frames ⁇ 1,2,3,4,5,6... ⁇ , at the first resolution level there will be ⁇ 1+2,3+4,5+6, ... ⁇ , where "1+2" represents the averaging of frames 1 and 2 together.
- the server computer 120 acquires thumbnail-resolution images that were created as a result of the 2-D sub-band transform decompositions.
- thumbnail-resolution images or, simply, "thumbnails” are frames presented at a resolution less than or equal to that of the original image and at a quality less than or equal to that of the original image.
- thumbnails are presented as a sequence of the same number as in the original image ⁇ 1,2,3,4,5,6 ⁇ , so each thumbnail corresponds to a frame.
- Thumbnail resolution images are a by-product of the LL part. After decomposing the original image a few resolution levels, the LL part at that level is used as the thumbnail. Note that in embodiments of the invention in which thumbnails are not acquired, the process can proceed directly to step 2950.
- 1-D sub-band transform decomposition in a z-axis direction is performed upon a portion, such as ⁇ HH,HL,LH ⁇ , of the two-dimensionally transform-decomposed digital image.
- the 1-D z-axis decomposition is not performed upon the LL portion in order to preserve its thumbnail qualities.
- the computed transform coefficients are corrected by the performance of integer operations that utilize one or more predetermined correction factors.
- the coefficients are ideally or theoretically to be scaled by a factor or 2 at each resolution level. Nevertheless, the floating-point computations required to scale coefficients by 2 can impact processing speed.
- Employing integer operations, such as multiplying by integer correction factors, to compensate for the omission of true floating-point operations maintains processing speed while sacrificing image accuracy at only some of the resolution levels.
- the server computer 120 transmits data to the client computer thumbnail-resolution images that were the result of the 2-D sub-band transform operations of step 2940.
- the server computer 120 progressively transmits data representing the identified ROI to the client computer 110.
- Figure 30 illustrates the method of operation associated with a further aspect of the invention, in which the client request can include not only a request list specifying data blocks that define the ROI but can also include a specified quality threshold.
- step 3110 indicates the client computer 110 (Fig. 1) transmitting such a request that includes the request list and a quality threshold. Additionally, the client request can include a specified resolution.
- the server computer 120 (Fig. 1) can map the luminance of the 3-D image stored on the image storage device from a predetermined image bit depth to a predetermined screen bit depth in accordance with a monotonic luminance mapping function.
- the server computer 120 computes a root mean square (RMS) increasing factor defined by a maximal derivative of the luminance mapping function.
- RMS root mean square
- a determination as to whether the RMS increasing factor is greater than one is made at step 3140. Accordingly, the server computer 120 provides more data blocks, i.e., more than some predetermined normal number, if the RMS increasing factor is greater than one at step 3150 and fewer data blocks, i.e., less than the predetermined number, if the RMS increasing factor is less than one at step 3160. If the RMS increasing factor is equal to one, the server computer 120 provides a number of data blocks that is responsive to the specified quality threshold at step 3170.
- the basic components of the protocol are accuracy layers associated with tiles of subband/wavelet coefficients.
- l(x,y,z) be a given single component 3-D image such as a grayscale image. It is explained herein below, in reference to Eq.l, how the protocol is generalized to arbitrary color spaces.
- FWT(l) (Forward Wavelet
- Transform be a subband/wavelet transform of the image.
- a basic introduction to the FWT is not provided in this patent specification because the FWT is in and of itself well known in the art.
- Transforms used herein are required to be separable.
- a 3-D separable transform can be computed by three ID steps: subsequently in the X, Y and Z directions.
- the wavelet transform typically creates for each given resolution j eight subbands denoted as III,, llhj, Ihlj, Ihhj, hllj, hlhj, hhlj, hhhj, while only the subband /// is passed to the next resolution level processing.
- the coefficients of FWT ⁇ are associated to the multiresolution structure shown on Fig. 3. Observe that each coefficient belongs to a certain resolution j .
- each coefficient "influences" only a compactly supported region of the image.
- High resolution/frequency coefficients are associated with a small region and low resolution/frequency coefficients are associated with a larger region.
- the coefficients are subdivided into the tiles as shown in the Fig. 3 for the purpose of efficient rendering. Several subgroups of coefficients are grouped into tiles, associated with the same spatial area, corresponding to the subbands at the resolution j that are not passed to the next resolution.
- the sub-tiles can have, for example, a length and width of 32 and height of 1, so that the tile contains «x32 2 xl coefficients, where n is the number of sub-bands that are not passed to the next resolution.
- n is the number of sub-bands that are not passed to the next resolution.
- each subband tile is further decomposed into "accuracy" data blocks.
- Each data block is associated with the "bit plane" interval [0,2) of level 0 or the intervals [2 ! , 2 /+1 ) at level / > 1 , so each data block of the subband tile of Eq. 1 will have a 5D coordinate
- maxBitPlane(t_resolution) 8 is suitable.
- the data block of Eq. 2 contains the following data in encoded format:
- each data block of Eq. 2 is associated with the bit ⁇ me _ ⁇ 2'- bitPlane ,s2 t - bi!Pla " e+l ) .
- the invention can provide a quickly generated thumbnail view of all the slices of the study, the 3-D image. This feature is particularly useful in medical imaging.
- the system can generate a low resolution representation of a 2-D image at every coordinate Z of the 3-D image.
- a multi- resolution decomposition of the original 3-D data is employed for this purpose.
- Figure 4 describes three types of subband decomposition. Each type merits different handling as follows:
- Case (A) is named herein the HHX-Subbands type.
- Case (B) is named herein the XXH-Subbands type.
- a wavelet transform in the Z direction only is performed first, and then the low-pass result of the transform is passed to the next resolution for each X and Y.
- Case (C) is named herein the All-Subbands type of decomposition.
- a 2-D wavelet transform is first done for each Z, and then the result undergoes the wavelet transform in the Z direction. Only the /// subband of this separable 3-D wavelet transform is passed to the next resolution level.
- An embodiment of the present invention for medical applications minimizes the response time for ROI request from the client by evaluating the multi-resolution decomposition in accordance with the following description in reference to Fig. 5.
- the embodiment then further employs XXH-Sub- bands type decomposition (steps 5050, 5060 and 5070) to reduce resolution in the Z direction alone.
- the resulting DC block (step 5070) may be further optionally decomposed by the All-Subbands type decomposition.
- Decomposition such as depicted in Fig. 5 allows quick calculation of all the thumbnail views of every slice of the study and also swift ("near real time") response to any ROI chosen by the client computer no.
- the multi-resolution decomposition mentioned herein above in reference to Figs. 4 and 5, is shown schematically in Fig. 6 for further clarity.
- the decomposition of original 3-D image 5010 is executed in two stages, the first stage is depicted by dividing line 6010, and the second stage is depicted by dividing line 6015. The first is done along the X and Y directions, and the second along the Z direction.
- the decomposition of the resulting image 5020 is executed in two stages, the first is depicted by dividing line 6020, and the second is depicted by dividing line 6025. The first is done along the X and Y directions, and the second along the Z direction.
- the decomposition of the resulting image 5030 continues until a k th step that is executed in two stages, the first is depicted by dividing line 6030, and the second is depicted by dividmg line 6035. The first is done along the X and Y directions, and the second along the Z direction.
- These first k decompositions are of the HHX-Subbands type.
- the next decompositions are of the XXH-Subbands type.
- the decomposition of image 5040 is depicted by dividing line 6040.
- the decomposition of image 5050 is depicted by dividing line 6050.
- the decomposition of image 5060 is depicted by dividing line 6060.
- Embodiments of the present invention can use a separable wavelet transform.
- Embodiments for lossy applications can use well-known wavelet transforms while embodiments for lossless applications can use wavelet transforms that are described below.
- 2-D Lossless Transforms can be constructed, each based on one of the transforms known in the art. Two embodiments are described herein, the first based on the well-known Haar transform, and the second is based on the lossless version of well-known Cohen-Daubechies-Feauveau (1,3) transform (C.D.F.). Persons skilled in the art can apply the principles described herein below to other transforms.
- any of such well-known lossless transforms can be employed in the X direction.
- a similar but different transform is employed in the Y direction. It is modified in each subband to get the proper scaling of the coefficients. Proper scaling means that coefficients are normalized in a Z 2 -norm as do wavelet coefficients computed in an accurate mathematical transform that is not limited to integer numbers.
- the forward step is the process of coding x(n) to obtain s(n) and d(n).
- the Haar X direction forward step is given by the following equation.
- the inverse step is the process of recovering x(n) from s(n) and d(n).
- the Haar X direction inverse step is given by the following equation.
- the Haar-based low part of the Y direction forward step is given by the following equation.
- Eq 4A shows the following properties of an embodiment of the present invention.
- s(n) comprises scaled LL-subband coefficient.
- s( «) and d (w) are a de-correlated and reversible couple (so they can be transformed back to x(2n) and x(2n + ⁇ )), but d ( «) is not scaled, so it is assigned half the value use in similar transformed known in the art.
- J w (n) is multiplied by 2.
- the LSB of the LH-subband coefficient d(n) is known to be 0 and not encoded.
- the Haar-based low part of the Y direction inverse step is given by the following equation.
- the Haar-based high part of the Y direction forward step is given by the following equation.
- Eq. 4C shows the following properties.
- d ( «) and s(n) comprise a de- correlated and reversible couple, but d ⁇ ( «) is not scaled, so it is assigned twice the value use in similar transformed known in the art. Therefore, d (n) is divided by 2.
- Half-Bit information serves to remind us that its weight is half that of other coefficients and that it is the least significant (from an approximation point of view).
- the Haar-based high part of the Y direction inverse step is given by the following equation.
- the forward step is the process of coding x(n) to obtain s(n) and d(n).
- the CDF (1,3) X direction forward step is given by the following equation.
- the CDF (1,3) X direction inverse step is given by the following equation.
- CDF (1,3) low part Y direction forward step is given by the following equation.
- CDF (1,3) low part Y direction inverse step is given by the following equation.
- CDF (1,3) high part Y direction forward step is given by the following equation.
- CDF (1,3) high part Y direction inverse step is given by the following equation.
- the 2-D transform requires two sequential ID transforms
- 2- D Lossless Transforms to mutually compensate the scale distortion of ID lossless wavelet transform and to keep the proper scaling of wavelet coefficients as described herein above in reference to Eq.'s 4a through 6d and in the above-referenced U.S. Patent Application Serial No. 60/198,017.
- the present invention modifies the lossless ID Haar transform in such a way that allows it to keep the wavelet coefficients as close to their scaled values as possible and to decrease the scatter of scaling factors in different subbands at all the resolution levels.
- the present invention uses a transform in the form of the following equation for HHX-Subbands type decomposition for the coefficients belonging to the HL and HH subbands of the 2-D slices
- Eq. 7B scales s(n) and d(n) by a factor of V2 . Because of the fact that all the coefficients resulting from the 2-D Lossless Transforms are scaled properly, all the wavelet coefficients of our entire 3-D transform are scaled in total effect by 2 in this case.
- the inverse transforms for Eq's 7A and 7B are 7C and 7D respectively.
- the resulting low-resolution 2-D image, that is passed to the next resolution processing is properly scaled. It should be also noted that, as explained herein above, it does not undergo a transform in Z direction.
- a single one-dimensional lossless wavelet transform doesn't preserve the scaling of the resulting coefficients, but the requirement exists, as explained herein above in reference to Fig. 5 and 6 for XXH-Subbands type decomposition to use a one- dimensional transform.
- the present invention can overcome this difficulty in the following way: Two different transforms can be implemented in turn. In one of them, the resulting lowpass coefficients are scaled by a factor of v2 , and in another they are scaled by . This preserves overall scaling of the low pass fraction. In other words, every odd time Eq. 7A is used for the forward step and Eq. 7C for the inverse step, and every even time the following Eq. 7E is used for the forward step, and Eq. 7F for the inverse step.
- Fig. 8 The figure depicts data blocks as defined in Eq. 2.
- Embodiments of the present invention for lossless transformation can utilize two additional "half bit planes", one of which is associated with 2-D wavelet transforms on the XY plane as explained in reference to Eq's 3 through 6. The other is associated with transform in the Z direction as explained in reference to Eq's 7.
- Each of these two "half bit planes" is implemented as a three-dimensional matrix.
- this matrix has size (tileLength/ 2) x (tileLength/ 2) x tileHeight while for the XXH-Subbands type decomposition this matrix has size tileLength x tileLength x (tileHeight/ 2) .
- each data block for lossless implementation contains the following data in encoded format: 1.
- For t_bitPlane ⁇ 2 a. A list of the indices of all subband coefficients whose absolute value is in the range [ 2 ⁇ - bman ⁇ - 2'- b " 'Plam ) .
- b. The sign of all the coefficients of 1.
- t _ bitPlane 1 , which is called herein the "least significant bit plane" : a. A list of the indices of coefficients whose absolute value belongs to the set ⁇ -1,0,1 ⁇ from ⁇ L ⁇ and ⁇ -subbands for HHX-Subbands type of decomposition or from ⁇ L ⁇ and ⁇ -subbands for XXH-Subbands at even resolution type of decomposition or all subbands for XXH- Subbands type of decomposition at odd resolutions. b. A "zero flag" for each coefficient of 2. a, which indicates if the coefficient is equal to zero or not. c. The sign of all the coefficients of 2.a, whose "zero flag" is false. d. The LSB of the coefficients from 2. a that belong to higher bit plane.
- t_bitPlane 0 which is called herein the "half bit plane", a matrix of (tileLength/ 2) x (tileLength/ 2) x tileHeight bits for HHX-Subbands type decomposition, or a matrix of tileLength x tileLength x (tileHeight/ '2) bits for XXH-Subbands type decomposition as their last "half bit”.
- the encoding algorithm employs the forward steps described above with reference to Eq's 3 through 7. It is performed at the server, which is block 120 in Fig. 1.
- the present invention is an imaging system that enables this time consuming task to be swiftly performed for a specific ROI and not necessarily on the full image.
- the following is a description the encoding algorithm for images with a single color component such as grayscale images. The straightforward generalization for an arbitrary number of components is explained later.
- the algorithm receives as input the following parameters listed in Table 1.
- a coefficient belongs to one of three types of groups:
- Type 16 A group of 4x4x1 coefficients (B group) ⁇ coef(4i + x,4j + y,k),x e [0,3],y & [0,3] ⁇
- Type4 A group of 2x2x1 coefficients (A group) ⁇ coef (2/ + x, 2j + y, k) , x e [0, l] , y e [0, l] ⁇
- Typel Single coefficient A Type 16 group is also named herein a 'B' group, and a Type4 group is also named herein as 'A' group.
- the size of the A and B groups of coefficients may be controlled by parameters such as the following:
- the sizes need not necessarily be 4 and 16.
- Step 2710 is an initialization
- step 2720 is an outer loop in which the inner loops nest.
- the first step in the encoding algorithm is its initialization (step 2710), which is conducted as follows.
- the second step in the encoding algorithm can be conducted as follows with reference to Fig. 9.
- the output of each such bit plane scan is the subband data block defined in Eq. 2. Since the last stage of the encoding algorithm is arithmetic encoding of given symbols, at the beginning of each scan (step 920) the arithmetic encoding output module is redirected to the storage area allocated for the particular data block.
- the output stream is closed (step 950) and the bit plane is decremented (step 960).
- the goal is to efficiently encode the locations of the coefficients that are "exposed" when encoding traverses down to the current bit plane.
- the encoder and decoder have at the time of the scan all the information on what "took place” at the previous higher bit plane scans.
- the heuristics of the algorithm is that coefficients, which are insignificant at the current bit plane, which are coefficients with absolute value ⁇ ⁇ 2 b , are spatially correlated. If it is known that a coefficient is insignificant at the bit plane b then with higher probability so will his neighbors. This explains why the algorithm groups coefficients into groups of 16 and then 4 and tries to efficiently report (with one symbol) their insignificance.
- the significance scan uses four binary probability models as listed in the following table.
- the models listed in Table 2 are simply histograms with two levels. They are initialized to equal probabilities at the beginning of each significance scan and are used by and updated by the arithmetic encoder each time a symbol is encoded. Description now proceeds to details of the significance scan.
- the encoder removes the label Typel6 (step 1050), splits the group to 4 subgroups of 2 x 2 coefficients and labels each of them Type4. If the group of 16 was not labeled Typel 6 (step 1020) then it must have been split at a higher bit plane. In both cases the next step will handle the four subgroups of four coefficients.
- the current group is the fourth subgroup.
- the encoder proceeds to the next subgroup. Else, it removes the label Type4, split the subgroup to the single coefficients and label each of them Typel (step 1100). If the subgroup was not labeled
- Type4 then it must have been split at a higher bit plane. In both cases the next task is to handle each individual coefficient.
- the current coefficient is the fourth coefficient.
- step 1000 in Fig. 10A The encoding accuracy update of step 1000 in Fig. 10A in now explained in detail. This step is performed if one of the following two conditions holds:
- the significant coefficient list contains all the coefficients coef(x,y,z) for which b(x,y,z) ⁇ b and their current approximated value coef(x,y,z) .
- the encoder For each coefficient in the list the encoder arithmetically encodes the value of the test coef(x,y,z) ⁇ coef(x,y,z) .
- a special binary probability model is initialized at the beginning of the step and used by the arithmetic encoder.
- the encoder accordingly simulates the reconstruction done by the decoder and updates the coefficient's approximated value by adding/subtracting from coef(x,y,z) the amount
- Fig. 13 skips all the coefficients that belong to those subbands that have zero least significant bits in all coefficients, as described in detail herein above in reference to Fig. 8.
- step 1230 in Fig. 12A and step 1330 in Fig. 13 returns true if and only if the coefficient is zero.
- the half bit plane scan is now described in reference to the pseudo-code shown in Fig. 12B.
- a scan is made in the three dimensions, labeled x, y and z. In each dimension the scan is made from zero to the size in that dimension. A symbol in encoded arithmetically in each point of the 3-D space.
- the scan could be made in any other particular order such as the reverse order to that embodiment depicted in Fig. 12B, as long as the encoder and the decoder agree on the order of the scan.
- the decoder algorithm is explained herein below, and the relevant decoder scan is described with reference to Fig. 16B.
- the decoding algorithm is a reversed reflection of the encoding algorithm performed in the server and described herein above. Decoding is done at the client computer during the progressive rendering operation as described at step 1705 of Fig. 17.
- Decoding algorithm initialization Assign the value zero to each coefficient Coef (x, y,z) .
- maxBitPlane (t _ resolutior ⁇ ) is available at the client, read the value of maxBitPlane(t ⁇ le) , which in a embodiment of the invention is the first byte.
- the input to each such bit plane scan is encoded data block (t_x,t_y,t_z,t_resolution ,b) .
- the arithmetic decoder's input module is redirected (step 1420) to read from "slot" in the database allocated for the particular data block.
- the tile's coefficients are required at some accuracy level for the rendering and the rule is obviously to use whatever data is present.
- the outer loop is terminated. For example, if the first data block is missing, all of the coefficients retain their initial value of zero. If only the first few data blocks are present then only coefficients with "large" absolute value will be reconstructed and at a low precision.
- the bit plane scan is finished and the data block (t_x,t y,t_z,t_resolution ,b) has been decoded (step 1450)
- Pseudo-code for the bit plane scan is provided in Fig. 15.
- the scan for a given level b, decodes all of the coefficients' data corresponding to the absolute value interval
- the goal of the significance scan is to decode the locations of the coefficients that are "exposed" when the decoder traverses down to the current bit plane. Since the scan is a reversed reflection of the encoding scan as described herein above in reference to Table 2, the description below is brief. Reference is made to Fig. 28A and Fig. 28B, which reflect corresponding Fig. 10A and Fig. 10B.
- Fig. 28 A The first loop is on groups of Typel 6. If there are no more such groups (step 2810) the significance scan is terminated and decoding proceeds to the accuracy update (step 2800). If the current group is still labeled Typel 6 (2820) decoding is performed of the Boolean value b_ Group ⁇ b (2830). If b_ Group ⁇ b (2840), then the group is currently insignificant and decoding proceeds to the next group (2845). Else, the decoder removes the label Typel 6 (step 2850) and splits the Typel 6 group to 4 subgroups of Type4. If the group of 16 was not labeled Typel 6 then it must have been split at a higher bit plane.
- decoding In both cases decoding must now continue to process the four subgroups of four coefficients.
- the decoder uses the same probability models as the encoder for the purpose of arithmetic decoding. Reference is now made to Fig. 28B. Decoding iterates (step 2860) over the subgroups of four coefficients. If a subgroup is labeled 4 Type arithmetic decoding of the value of the test b_ Subgroup ⁇ b is done (step 2880). The arithmetic decoding step is not performed if the following three conditions hold true:
- the current group of Type4 is the fourth subgroup.
- the decoder removes the label Type4, splits the subgroup to the single coefficients and labels each of them Typel. If the subgroup was not labeled Type4 then it must have been split at a higher bit plane. In both cases the decoder must now continue to handle each individual coefficient. Decoding loops over the four coefficients. If a coefficient coef(x,y,z) is still labeled Typel the decoder arithmetically decodes the value of the test b(x,y,z) ⁇ b. It does not perform the arithmetic decoding step if the following three conditions hold:
- the current coefficient is the fourth coefficient.
- the significant coefficient list contains all the coefficients coef(x,y,z) for which b(x,y,z) ⁇ b at an approximated value.
- the encoder provided this list by step 1000 in Fig. 10A, and the decoder handles it in step 2800 in Fig. 28 A.
- the decoder arithmetically decodes the value of test coef (x,y,z) ⁇ coef (x,y,z) , where coef(x,y,z) is the actual value of the coefficient (not known to the decoder). It then accordingly updates the approximated value by adding/subtracting from coef(x,y,z) the amount
- Fig. 16A and Fig. 16B Pseudo-code for the least significant bit plane scan and half bit plane scan are given in Fig. 16A and Fig. 16B.
- the decoder scans all the coefficients from all subbands at step 1610. Decoding of those of the subbands that are skipped by the encoding algorithm is not performed, but their least significant bit is counted as zero instead.
- the order of the scan in Fig. 16B reflects the order of the encoding scan in Fig. 12B. Reference is now made to Fig. 17 for the description of the workflow at the client computer 110.
- Any new ROI generated by the user's action invokes at step 1701 a call from the GUI interface to the client imaging module with the new ROI view parameters.
- the client imaging module then computes in step 1702 which data blocks (described herein above in reference to Fig. 3) are required for the rendering of the ROI and checks if these data blocks are available in the client cache. If not, their coordinate (given by Eq.2) is added to a request list ordered according to some progressive mode.
- the request list is then encoded in step 1703 and sent to the server.
- the server responds to the request by sending back to the client a stream of data blocks, in the order in which they were requested.
- the client inserts them to their appropriate location in the distributed database.
- the rendering of the ROI is invoked. Naturally, the rendering operation is progressive and is able to use only the currently available data in the client's database.
- the imaging module on the client computer 110 receives view parameters listed in Table 4 from the GUI interface module. These parameters are used in step 1702 to generate a request list for the ROI. The same parameters are used to render the ROI in step 1705.
- the basic parameters of a ROI are worldPolygon, sliceRange and scale, which determine uniquely the ROI view. If the ROI is to be rendered onto a viewing device with limited resolution, then a sliceRange, containing a large number of slices of the 3- D image, will be coupled by a small scale.
- the other view parameters determine the way in which the view will be rendered.
- the parameters deviceDepth and viewQuality determine the quality of the rendering operation. In cases the viewing device is of low resolution or the user sets the quality parameter to a lower quality, the transfer size can be reduced significantly, as described herein below in reference to step 1702.
- the parameter luminanceMap is typically used in medical imaging for grayscale images that are of higher resolution than the viewing device.
- screens display grayscale images using 8 bits, while medical images sometimes represent each pixel using 16 bits.
- the parameter progressiveMode determines the order in which data blocks should be transmitted from the server computer 120.
- the "Progressive By Accuracy” mode is the preferred mode for viewing in low bandwidth environments.
- “Progressive By Resolution” mode is easier to implement since it does not require the more sophisticated accuracy (bit plane) management and therefore is commonly found in previous solutions.
- the superiority of the "progressive by accuracy” mode can be mathematically proved by showing the superiority of non-linear approximation over linear approximation for the class of real-life images.
- the "Progressive by Spatial Order” mode is preferred for a cine view of the sequence of 2-D slices one by one. In this mode the image data is ordered and received from the first slice to the last one, such that rendering can be done in parallel to the transmission.
- the curve influences not only the mapping, i.e. the drawing to the screen, but also the request from the server.
- the request is created such that the image approximation in the client side is close enough to the original image, i.e., the RMS (Root Mean Square error) is visually negligible.
- the RMS Root Mean Square error
- the RMS can be increased or reduced.
- the maximal RMS increasing factor depends on the maximal derivative of the curve as in Eq. 9.
- / is the approximated image
- / is the mapping function
- max( ') is the maximal derivative of the curve.
- the request list should be increased (more bit-planes should be requested from the server) in order to improve the approximation accuracy. Conversely, if the RMS increasing factor is smaller than 1, the request list can be reduced. The exact specification of this is given in the follows description.
- the client imaging module calculates data blocks request list ordered according to the progressiveMode. Given the parameters worldPolygon, sliceRange and scale it is possible to determine which subband tiles (such as defined in Eq. 1) in the frequency domain participate in the reconstruction of the ROI in the image domain. These tiles contain all the coefficients that are required for an Inverse Subband/Wavelet Transform (IWT) step that produces the ROI.
- IWT Inverse Subband/Wavelet Transform
- the parameter dyadicResolution(ROI) which is the lowest possible dyadic resolution higher than the resolution of the ROI, is first calculated. Any subband tiles of a higher resolution than dyadicResolution(ROT) do not participate in the rendering operation.
- the ROI is a low resolution rendering of the full image at the resolution n.
- all the subband coefficients at the subband resolutions l ⁇ j ⁇ n participate in the IWT of the ROI.
- these large coefficients are visually significant for a rendering of this ROI, even at very high qualities.
- these large coefficients are only required at a very low precision.
- the ROI is at the highest resolution of the image.
- the parameter deviceDepth listed in Table 4 is set at low resolution (indicating that the viewing device is of low resolution) or the parameter viewQuality is set at low quality (indicating that the user is currently not interested in high quality rendering)
- the last data blocks of each subband tile corresponding to high precision need not be ordered.
- minRequestPlane min (maxBUPlane (t _ resolution) , minRequestPlane + 1 J .
- the client imaging module encodes the request list into a request stream which is sent to the server computer 120 via the media 130 (see Fig. 1).
- the request list encoding algorithm is a "cube based" procedure.
- the heuristics of the algorithm is that the requested data block usually can be grouped into data block "cubes”. From the request list of data blocks indexed as in Eq.2 the encoding algorithm computes structures of the type presented in Eq.l 1.
- the encoding algorithm attempts to create the shortest possible list of structures of the type defined by Eq.l 1, which can be done by collecting the data blocks to the largest possible cubes. It is important to note that the algorithm ensures the order of data blocks in the request list is not changed, since the server will respond to the request stream by transmitting data blocks in the order in which they were requested. This is particularly preferable when a user switches to another subsequent group of 3-D image slices that has not been viewed before. In such a case the request list might be composed of hundreds of requested data blocks, but they will be collected to one (x,y,z) cube as defined by Eq.l 1 for each pair ( t_ resolution, t_ bitPlane). In step 1704 in Fig.
- the client computer 110 upon receiving from the server an encoded stream containing data blocks, decodes the stream and inserts the data blocks into their appropriate location in the distributed database using their ID as a key.
- the decoding algorithm performed in this event is a reversed step of the encoding process described in reference to step 1703. Since the client computer 110 is aware of the order of the data blocks in the encoded stream, only the size of each data block need be reported along with the actual data. In case that the server informs of an empty data block, the receiving module marks the appropriate slot in the database as existing but empty.
- each subband tile is of size tileLength x tileLength x tileHeight and that the wavelet basis used has a maximal filter size maxFilterSizeXY in X and Y directions and maxFilterSizeZ in Z direction, then defining hFilterSize - numbers of HHX-
- step 1705 the step of progressive rendering.
- the client performs rendering operations of the ROI.
- the client runs two program threads: communications and rendering.
- the rendering thread runs in the background and draws into a pre-allocated "off-screen” buffer. Only then does the client use device and system dependant tools to output the visual information from the "off-screen" to the rendering device such as the screen or printer.
- the rendering algorithm performs reconstruction of the ROI at the highest possible quality based on the available data at the client. That is data that were previously cached or data that recently arrived from the server. For efficiency, the progressive rendering is performed only for the portion of the ROI that is affected by newly arrived data. Specifically, data that arrived after the previous rendering task began. This updated region is obtained using the method of step 1704 described herein above.
- the parameters of the rendering algorithm are composed of two sets as follows.
- the first set comprises ROI parameters as listed in Table 4 herein above.
- the second set comprises parameters transmitted from the server as listed in Table 6 herein below, with the exception of one parameter, the jumpSize parameter, which is a parameter particular to the server.
- the rendering algorithm computes pixels at the dyadic resolution dyadicResolution(ROT). Recall that this is the lowest possible dyadic resolution, which is higher than the resolution of the ROI.
- the obtained image is then resized to the correct resolution and mapped to the depth of output device. Discussion of the rendering rate now follows. As ROI data is transmitted to the client, the rendering algorithm is performed at certain time intervals of a few seconds. At each point in time, only one rendering task is performed for any given displayed image. To ensure that progressive rendering does not become a bottleneck, two rates are measured: the data block transfer rate and the ROI rendering speed. If it is predicted that the transfer will be finished before a rendering task, a small delay is inserted, such that rendering will be performed after all the data arrives.
- This data-structure is required to efficiently store subband coefficients, in memory, during the rendering algorithm. This is required since the coefficients are represented in long integer (lossless coding mode) or floating-point (lossy coding mode) precision, which typically require more memory than a pixel representation (1 or 2 bytes). In a lossy mode the coefficients at the client side are represented using floating-point representation, even if they were computed at the server side using an integer implementation to minimize round-off errors.
- the rendering algorithm begins with initializing coefficient and pixel memory 3-D matrices.
- dyadicWidih(RO ⁇ ) the width of the T projection of the ROI onto the resolution dyadicResolution(RO ⁇ ) and height(RO ⁇ ) the number of slices in the ROI.
- dyadicWidih(RO ⁇ ) the width of the T projection of the ROI onto the resolution dyadicResolution(RO ⁇ )
- height(RO ⁇ ) the number of slices in the ROI.
- For each component two 3-D matrices of coefficients are allocated with dimensions dyadicWidth (ROl) x dyadicWidth (ROI) x height (ROI)
- the algorithm executes a multi- resolution march from the first to the last tile of the ROI in Z direction.
- the pseudocode for this recursive march is given in Fig. 20.
- First the allocated matrix is filled with sub-tiles of coefficients calculated on the lower resolution, giving low pass subbands of the current resolution.
- the matrix is complemented by the high pass subbands (wavelet) coefficients, decoded from the database or read from the memory cache as described herein below in reference to Fig. 21.
- An inverse subband transform described herein below in reference to Fig. 3 produces multiresolution pixels.
- Each time the highest resolution DyadicResolution(ROl) is reconstructed it is supplied as input to the resizing and luminance mapping step.
- the subband coefficients data structure described in herein above in reference to Fig. 20 is filled on a tile basis.
- Each such subband tile is obtained either by decoding the corresponding data blocks stored in the database or by reading from the memory cache.
- the memory cache is used to store coefficients in a simple encoded format. The motivation for its use is as follows.
- the decoding algorithm described herein above in reference to Table 3 is computationally intensive and thus should be avoided whenever possible.
- the rendering module uses a memory cache where subband coefficient are stored in a relatively simple encoded format, which decodes relatively fast. For each required subband tile the following extraction procedure is performed, as depicted in Fig. 21.
- step 2130 If no data blocks are available in the database for the subband tile (step 2130), its coefficients are set to zero (step 2140). If the tile's memory cache storage is updated (step 2150), in other words if it stores coefficients in the same precision as in the database, then the coefficients can be efficiently read from there (step 2160). The last possibility is that the database holds the subband tile in higher precision. Then, the tile is decoded (step 2170) down to the lowest available bit plane using the decoding algorithm described herein above in reference to Table 3, and the cached representation is replaced (step 2180) with a representation of this higher precision information.
- the image obtained by the previous step must be re-sized to this resolution.
- the operating system's implementation is sub-sampling which in many cases produces an aliasing effect, which is not visually pleasing.
- the imaging system uses the method of linear interpolation.
- the output of the interpolation is written to an "off-screen" buffer. From there it is displayed on the screen using system device dependant methods.
- luminanceMap is active, mapping to 8-bit screen is performed using the mapping function described in reference to Fig. 18.
- an uncompressed digital 3-D image is stored in, for example, storage 122 of the server computer 120.
- This uncompressed digital 3-D image may be a sequence of two-dimensional images, stored with a selected resolution and depth.
- each of these the uncompressed digital image may be contained in a DICOM file.
- step 2201 the server computer 120 performs preprocessing. This step is a computation done on data read from the original digital 3-D image. The results of the computation are stored in the server cache device 121. After this fast computation a "ready to serve" message is sent from the server to the client containing basic information on the image.
- the server receives an encoded stream of requests for data blocks associated with a ROI that needs to be rendered at the client.
- the server then decodes the request stream and extracts the request list.
- step 2203 the server reads from cache or encodes data block associated with low-resolution portions of the ROI, using the cached result of the preprocessing Step 2201.
- the server (step 2204) either reads from cache or performs a local and efficient version of the preprocessing step 2201. Specifically, a local portion of the uncompressed 3-D image volume, associated with the ROI, is read from the storage 122, processed and encoded. In step 2205, the data that was encoded in step 2203 and step 2204 is progressively sent to the client in the order it was requested.
- the preprocessing step 2201 is now described in further detail with reference to
- the preprocessing algorithm is to provide the fastest response to the user's request to interact with the 3-D image. Once this fast computational step is performed, the server is able to provide efficient transmission of any client ROI requests that will follow.
- the first ROI is a view of several 2-D slices of the 3-D image volume at low resolution (thumbnails).
- the preprocessing algorithm begins with a request for an uncompressed image that has not been processed before or that has been processed but the result of this previous computation has been deleted from the cache. As explained, this unique algorithm is preferable over the simpler procedure of encoding the full image into some progressive format. This latter technique provides a significantly slower response to the user's initial request then the technique described below provided by the present invention.
- a "ready to serve ROI” message is sent to the client containing basic information on the 3-D image. While some of this information, 3-D image dimensions, original color space, resolution etc., is available to the user of the client computer, most of this information is "internal” and required by the client to formulate ROI request lists (step 1702) and progressively render (step 1705).
- ROI request lists step 1702
- progressively render step 1705
- Table 6 Preprocessing Parameters Given an input image, the parameters described in Table 6 above are either computed or chosen. These parameters are also written into a header sent to the client and are used during the progressive rendering step 1705. The most important parameters to select are the following:
- the losslessMode is a mode in which progressive transmission of images takes place until lossless quality is obtained. Choosing this mode requires the preprocessing algorithm to use certain reversible wavelet transforms and may slow down the algorithm.
- the subbandTransformType(j) parameters are used for the (dynamic) selection of wavelet basis that is crucial to the performance of the imaging system.
- the selection can be non-stationary, meaning a different transform for each resolution/.
- the selection is derived from the following considerations.
- the first consideration is of coding performance (in a rate/distortion sense). This is obviously required from any subband/wavelet transform.
- the second consideration is of an approximation of ideal low pass. It is preferred to choose a transform such that low resolutions of the image will be of high visual quality. Some filters produce poor quality low resolutions even before any compression takes place.
- the third consideration is of fast transform implementation.
- the associated fast transform must be implemented using fast operations such as lifting steps, or integer shifts and additions, etc.
- Some good examples are the Haar and CDF transforms (1,3), (2,2).
- the fourth consideration is of fast low pass implementation. This is a very important parameter, since together with the parameter named jumpSize herein, it determines almost all of the complexity of the algorithm.
- the CDF (1,3) is in this respect the "optimal" transform with three vanishing moments.
- the dual scaling function is the simple B-spline of order 1
- its low pass is simple averaging.
- the sequence of CDF transforms, using the B-spline of order 1 as the dual scaling function, but with wavelets with increasing number of vanishing moments are in some sense "optimal" in our system. They provide a framework for both real time response and good coding efficiency.
- the fifth consideration is the existence of a lossless mode. If the losslessMode parameter is true, filters must belong to a subclass of reversible transforms.
- the sixth consideration is of low system I/O. If the network 157 depicted in Fig. 1 connecting between the image residing on the storage 122 and the imaging server 120 is slow, the bottleneck of the preprocessing stage (and the whole imaging system for that fact) might be simply the reading of the original image. In such a case we may choose a transform with a "lazy" sub-sampling low pass filter that corresponds to efficient selective reading of the input image. Many interpolating subband transforms with increasing number of vanishing moments can be selected to suit this requirement. However, this choice should be avoided whenever possible, since it conflicts with the first and second considerations.
- the decompositionTypefj parameters control the type of multi-resolution decomposition performed on each resolution /.
- These parameters together with the parameter named jumpSize give us the opportunity to achieve quick response to the first view of the 3-D image.
- the common first view is the group of several subsequent 2-D slices of the 3-D image volume at thumbnail resolution.
- the parameter named jumpSize controls the tradeoff between fast response to the user's initial request for interaction with the image and response times to subsequent ROI requests. WhenjumpSize is large, the initial response is faster, but each computation of a region of interest with higher resolution than the jump might require processing of a large portion of the original image.
- the threshold ⁇ ) parameters control the visual quality in the case of lossy compression.
- the tradeoff is quality for bit-rate.
- the choice is also derived from the mathematical properties of the pre-chosen subband transform(s). In the case of lossless compression these parameters are not used since in lossless mode there is no threshold operation performed.
- each such memory frame buffer stores the low-pass coefficients at various resolutions.
- the memory complexity is limited by several single frames and is independent of the number of frames in the original sequence.
- the resolutions are scanned simultaneously from start to end in the z direction. For each resolution the corresponding frame buffer stores low- pass coefficients or "pixels" at that resolution.
- the core of the preprocessing algorithm are steps 2302 to 2305 where tiles of pixels of size are read from the memory frame buffers and handled one at a time.
- Step 2302 the tile is transformed into a tile of size [ tileLength , tileLength, tileHeight ] containing two types of coefficient data: subband coefficients and "pixels" at a lower resolution.
- the subband coefficients are processed in step 2303 and step 2304 and are stored in the cache.
- the lower resolution pixels are inserted in step 2305 to a lower resolution memory frame buffer.
- a memory management module performs the following check. If the part of the sub-tile exceeds the current virtual boundaries of the memory frame buffer, the corresponding first frames of the.memory frame buffer are considered unnecessary anymore. Their memory is (virtually) re-allocated for the purpose of storing the sub-tile's new data.
- the pseudo-code of the memory constraint scan algorithm is detailed in Fig. 24.
- Step 2301 the low pass filters of the transforms subbandTransformTypeQ), numberO ⁇ esolutions - jumpSize ⁇ j ⁇ numberO ⁇ esolutions , are used to obtain a low resolution 3-D image at the resolution numberO ⁇ esolutions -jumpSize (as can be seen in Fig. 5).
- the low pass calculation is initiated by an input operation of tiles of pixels from the memory frame buffers performed in step 2302. Whenever there is an attempt to read missing low resolution tiles, they are computed by low-pass filtering the original image and inserted into the memory frame buffer. As explained herein above, the insertion over-writes frames which are no longer required, such that the algorithm is memory constraint.
- the server needs to read (again) the associated local area of the original image and perform (using step 2204) more computations.
- the low pass filtering can be implemented efficiently in integer numbers without paying much attention to round-off errors.
- the low pass filter is implemented using a special lossless "integer to integer" transform technique. Unfortunately, this means that this step is slower in lossless mode, due to this special implementation. Nevertheless, it is the embodiment as it provides fast response to the user's first request to view the image in a true lossless efficient imaging system.
- jumpSize parameter defines the number of lossless wavelet low pass steps should be done.
- a single low pass step is the same for Haar and CDF (1,3) and for the HHX-Subbands type of decomposition that is the only one used in the "jump". This step is defined by the following Eq.12.
- the server performs these steps efficiently (almost like the lossy algorithm) by a single operation that simulates exactly jumpSize low pass steps defined by Eq.12.
- the simplicity of the formula makes filters such as Haar and CDF (1,3) optimal in the sense of low pass efficiency.
- Step 2302 is one step of an efficient local subband transform (FWT) performed on a tile of pixels at the resolution l ⁇ j ⁇ numberOf Resolutions- jumpSize.
- the type of transform is determined by the parameters decompositionTypeQ) and subbandTransformTypeQ). The transform is performed on an extended tile of pixels of size
- the output of the step is a subband tile as defined by Eq.2 of size [tileLength,tileLength,tileHeight] containing two types of coefficient data: subband/wavelet coefficients and low resolution coefficients/pixels.
- the transform step can be efficiently implemented in integer numbers. This saves converting the original pixels to floating-point representation and facilitates the next quantization step (used for lossy transform).
- the subband transform of step 2302 outputs two types of coefficients: scaling function (low pass) and wavelet (high pass). For the lossless case their wavelet coefficients also include additional Half bit plane data.
- the wavelet coefficients are treated in step 2303 and step 2304 while the scaling function coefficients are treated in step 2305. Tiles of pixels which are located on the boundaries sometimes need to be padded by extra frames, rows and/or columns of pixels, such that they will formulate a complete tile of size [tileLength,tileLength,tileHeight] .
- step 2303 unless losslessMode is true, the subband coefficients calculated in step 2302 are quantized. This procedure is performed at this time for the following reason. It is required that the coefficients computed in the previous step will be stored in the cache 121. To avoid writing huge amounts of data to the cache 121, some compression is required. Thus, the quantization step serves as a preparation step for the next variable length encoding step. It is important to point out that the quantization step has no effect on compression results. Namely, the quantization step is synchronized with the encoding algorithm ensuring that the results of the encoding algorithm of quantized and non-quantized coefficients are identical.
- a tile of an image at the resolution/ is quantized using the given threshold, thresholdQ) : for each coefficients x , the quantized value . It is advantageous to choose the parameters thresholdQ) to be dyadic such that the quantization can be implemented using integer shifts.
- the quantization procedure performed on a subband tile is as follows:
- a solution sets to zero subband coefficients that are not associated with the actual image, but only with a padded portion. To do this the amount of padding is taken into account together with the parameters maxFilterSizeXY and maxFilterSizeZ. The motivation for this procedure of the removal of these coefficients is coding efficiency.
- variable length length (i,j,k) is first written using log 2 (maxBitPlane(tile)j bits. Then for each coefficient in the group length(i,j,k) + l bits are written representing the coefficient's value. The least significant bit represents the coefficient's sign. For example, if it is 1 then the variable length encoded coefficient is assumed to be negative. The HalfBit subband coefficients are written last using one bit per coefficient.
- step 2305 the low pass coefficients are inserted into the pyramidal frame buffer data structure at the appropriate location. If the size of the subband tile is [tileLength,tileLength,tileHeight], then the size of the low pass sub-tile is [tileLength/2,tileLength/2,tileHeight] for the HHX Subbands type decomposition and [tileLength,tileLength,tileHeight/2] for XXHjSubbands type, as explained herein above in reference to Fig. 4.
- Step 2202 is the decoding of the request stream. This is the inverse step 1703. Once the request stream arrives at the server, it is decoded back to a data block request list. Each data structure defined by Eq.ll and representing a group of requested data blocks is converted to the sub-list of these data blocks.
- step 2203 is only performed whenever the data blocks associated with low resolution subband tile are not available in the server cache 121.
- Step 2501 is the inverse step of step 2304.
- subband tiles of low resolution that is resolutions lower than numberO ⁇ esolutions - jumpSize , were stored in the cache using a variable length type algorithm.
- the variable length representation is first decoded, and then the algorithm uses the stored value maxBitPlane(tile) .
- HalfBit subband for lossless case, else and in the lossy case following decoding algorithm if performed.
- log 2 bits are read representing the variable length of the group.
- HalfBit subband coefficient are read bit by bit from cache, and then the decoding algorithm is performed as in the lossy case.
- step 2502 the encoding algorithm described in reference to Table 1 encodes the requested data blocks associated with the extracted subband tile.
- step 2201 In case that a jumpSize greater than zero is used in step 2201 and the resolution of the ROI is greater than numberOfResolutions - jumpSize the following is the calculation. It is a variation on the preprocessing step described in reference to Fig. 23 .
- the server receives a request list of data blocks its checks the following. If a data block has been previously computed (present in the cache 121), or if it is associated with a low resolution subband tile data block, then it is either read from the cache or handled in step 2203. Else, the coordinates of the data block are used in the same manner as in step 1704 described herein above to find the minimal portion of the ROI that needs to be processed. A local version of the preprocessing algorithm is then performed for this local portion.
- the encoded data tiles are sent to the client, in the order they were requested.
- data blocks will be empty.
- cubes of empty data blocks are collectively reported to the client using a structure of the type defined by Eq.l 1 under the restriction that they are reported in the order in which they were requested.
- Eq.l 1 the restriction that they are reported in the order in which they were requested.
Abstract
Description
Claims
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
AU2002229090A AU2002229090A1 (en) | 2000-12-14 | 2001-12-14 | Three-dimensional image streaming system and method for medical images |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US25631800P | 2000-12-14 | 2000-12-14 | |
US60/256,318 | 2000-12-14 |
Publications (2)
Publication Number | Publication Date |
---|---|
WO2002048966A1 WO2002048966A1 (en) | 2002-06-20 |
WO2002048966A9 true WO2002048966A9 (en) | 2002-12-27 |
Family
ID=22971791
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US2001/048933 WO2002048966A1 (en) | 2000-12-14 | 2001-12-14 | Three-dimensional image streaming system and method for medical images |
Country Status (3)
Country | Link |
---|---|
US (1) | US7376279B2 (en) |
AU (1) | AU2002229090A1 (en) |
WO (1) | WO2002048966A1 (en) |
Families Citing this family (58)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7236637B2 (en) * | 1999-11-24 | 2007-06-26 | Ge Medical Systems Information Technologies, Inc. | Method and apparatus for transmission and display of a compressed digitized image |
US20020044696A1 (en) * | 1999-11-24 | 2002-04-18 | Sirohey Saad A. | Region of interest high resolution reconstruction for display purposes and a novel bookmarking capability |
US7421136B2 (en) * | 1999-11-24 | 2008-09-02 | Ge Medical Systems Information Technologies Inc. | Image tessellation for region-specific coefficient access |
AUPR432701A0 (en) * | 2001-04-10 | 2001-05-17 | Lions Eye Institute Of Western Australia Incorporated, The | Virtual service system for client and service provider users and method therefor |
JP5388393B2 (en) * | 2001-04-27 | 2014-01-15 | キヤノン株式会社 | Image processing apparatus, image processing method, and control program |
US7290011B2 (en) * | 2003-11-26 | 2007-10-30 | Idx Investment Corporation | Image publishing system using progressive image streaming |
US20050111746A1 (en) * | 2003-11-26 | 2005-05-26 | Kumar Bharath S. | Progressive medical image volume navigation |
SE0401021D0 (en) * | 2004-04-21 | 2004-04-21 | Sectra Imtec Ab | Data reduction for the production of computer generated graphics and analysis |
US8868772B2 (en) * | 2004-04-30 | 2014-10-21 | Echostar Technologies L.L.C. | Apparatus, system, and method for adaptive-rate shifting of streaming content |
JP4644449B2 (en) * | 2004-07-12 | 2011-03-02 | 富士通株式会社 | Image display device and image display program |
US7738710B2 (en) * | 2004-08-02 | 2010-06-15 | Electronics For Imaging, Inc. | Methods and apparatus for communicating and displaying compressed image data |
JP4408794B2 (en) * | 2004-11-18 | 2010-02-03 | ザイオソフト株式会社 | Image processing program |
US8423673B2 (en) | 2005-03-14 | 2013-04-16 | Citrix Systems, Inc. | Method and apparatus for updating a graphical display in a distributed processing environment using compression |
US7489825B2 (en) * | 2005-07-13 | 2009-02-10 | Ge Medical Systems | Method and apparatus for creating a multi-resolution framework for improving medical imaging workflow |
CN100438406C (en) * | 2006-01-23 | 2008-11-26 | 北京航空航天大学 | Remote rendering based three-dimensional model network distribution method |
US8633927B2 (en) | 2006-07-25 | 2014-01-21 | Nvidia Corporation | Re-render acceleration of frame with lighting change |
US8175115B2 (en) * | 2006-11-17 | 2012-05-08 | General Electric Company | Method and system for iterative reconstruction |
US8725801B2 (en) * | 2006-11-21 | 2014-05-13 | General Electric Company | Systems and methods for image sharing in a healthcare setting while maintaining diagnostic image quality |
JP2008259612A (en) * | 2007-04-11 | 2008-10-30 | Fujifilm Corp | Projection image generator and its program |
JP4895386B2 (en) * | 2007-05-14 | 2012-03-14 | キヤノン株式会社 | Workflow generation device, workflow generation method and program |
GB0718607D0 (en) * | 2007-05-16 | 2007-10-31 | Seereal Technologies Sa | Holograms |
US8391353B2 (en) * | 2007-10-16 | 2013-03-05 | Thomson Licensing | Methods and apparatus for artifact removal for bit depth scalability |
US8078749B2 (en) * | 2008-01-30 | 2011-12-13 | Microsoft Corporation | Synchronization of multidimensional data in a multimaster synchronization environment with prediction |
US20100050110A1 (en) * | 2008-08-19 | 2010-02-25 | General Electric Company | Integration viewer systems and methods of use |
US9474500B2 (en) * | 2009-02-05 | 2016-10-25 | The Research Foundation Of State University Of New York | Method and system for transfer of cardiac medical image data files |
US8300965B2 (en) * | 2009-03-24 | 2012-10-30 | General Electric Company | Methods and apparatus to perform multi-focal plane image acquisition and compression |
US8559733B2 (en) * | 2009-03-31 | 2013-10-15 | Citrix Systems, Inc. | Methods and systems for approximating progressive image encoding using image partitioning |
TWI401614B (en) * | 2010-02-05 | 2013-07-11 | Pixart Imaging Inc | Data arrangement method and data arrangement device |
US8571280B2 (en) * | 2010-02-22 | 2013-10-29 | Canon Kabushiki Kaisha | Transmission of medical image data |
EP2635983A4 (en) * | 2010-11-03 | 2017-03-08 | Mobile Imaging in Sweden AB | Progressive multimedia synchronization |
US10395762B1 (en) | 2011-06-14 | 2019-08-27 | Merge Healthcare Solutions Inc. | Customized presentation of data |
US8867807B1 (en) | 2011-09-23 | 2014-10-21 | Dr Systems, Inc. | Intelligent dynamic preloading and processing |
WO2014084271A1 (en) * | 2012-11-27 | 2014-06-05 | 株式会社東芝 | Medical image processing device and medical image processing program |
US9521384B2 (en) | 2013-02-14 | 2016-12-13 | Red.Com, Inc. | Green average subtraction in image data |
US9530226B2 (en) * | 2014-02-18 | 2016-12-27 | Par Technology Corporation | Systems and methods for optimizing N dimensional volume data for transmission |
JP2015197329A (en) * | 2014-03-31 | 2015-11-09 | 三菱重工業株式会社 | Data transmission system, data transmission apparatus, data transmission method and data transmission program |
WO2016000093A1 (en) * | 2014-06-05 | 2016-01-07 | Empire Technology Development Llc | Data interpolation |
US9674554B2 (en) | 2014-10-28 | 2017-06-06 | Sony Corporation | Image processing system with coding mode and method of operation thereof |
US9591330B2 (en) | 2014-10-28 | 2017-03-07 | Sony Corporation | Image processing system with binary adaptive Golomb coding and method of operation thereof |
US10356410B2 (en) | 2014-10-28 | 2019-07-16 | Sony Corporation | Image processing system with joint encoding and method of operation thereof |
US9294782B1 (en) | 2014-10-28 | 2016-03-22 | Sony Corporation | Image processing system with artifact reduction mechanism and method of operation thereof |
US10063889B2 (en) | 2014-10-28 | 2018-08-28 | Sony Corporation | Image processing system with conditional coding and method of operation thereof |
US9357232B2 (en) | 2014-10-28 | 2016-05-31 | Sony Corporation | Image processing system with binary decomposition and method of operation thereof |
US10074178B2 (en) * | 2015-01-30 | 2018-09-11 | Dental Imaging Technologies Corporation | Intra-oral image acquisition alignment |
US10425350B1 (en) | 2015-04-06 | 2019-09-24 | EMC IP Holding Company LLC | Distributed catalog service for data processing platform |
US10791063B1 (en) | 2015-04-06 | 2020-09-29 | EMC IP Holding Company LLC | Scalable edge computing using devices with limited resources |
US10706970B1 (en) | 2015-04-06 | 2020-07-07 | EMC IP Holding Company LLC | Distributed data analytics |
US10776404B2 (en) | 2015-04-06 | 2020-09-15 | EMC IP Holding Company LLC | Scalable distributed computations utilizing multiple distinct computational frameworks |
US10860622B1 (en) * | 2015-04-06 | 2020-12-08 | EMC IP Holding Company LLC | Scalable recursive computation for pattern identification across distributed data processing nodes |
US10015106B1 (en) | 2015-04-06 | 2018-07-03 | EMC IP Holding Company LLC | Multi-cluster distributed data processing platform |
KR20230091200A (en) * | 2015-12-14 | 2023-06-22 | 파나소닉 인텔렉츄얼 프로퍼티 코포레이션 오브 아메리카 | Three-dimensional data coding method, three-dimensional data decoding method, three-dimensional data coding device, and three-dimensional data decoding device |
US10656861B1 (en) | 2015-12-29 | 2020-05-19 | EMC IP Holding Company LLC | Scalable distributed in-memory computation |
WO2018123801A1 (en) * | 2016-12-28 | 2018-07-05 | パナソニック インテレクチュアル プロパティ コーポレーション オブ アメリカ | Three-dimensional model distribution method, three-dimensional model receiving method, three-dimensional model distribution device, and three-dimensional model receiving device |
EP3649783A1 (en) | 2017-07-05 | 2020-05-13 | Red.Com, Llc | Video image data processing in electronic devices |
US10462495B2 (en) | 2017-08-09 | 2019-10-29 | Vital Images, Inc. | Progressive lossless compression of image data |
US11606569B2 (en) * | 2018-09-25 | 2023-03-14 | Apple Inc. | Extending supported components for encoding image data |
JP2022511489A (en) * | 2018-12-04 | 2022-01-31 | ヴィド スケール インコーポレイテッド | Tile group compartmentalization |
US20210314621A1 (en) * | 2020-04-07 | 2021-10-07 | Teradici Corporation | Method and apparatus for tiled wavelet image encoding and decoding |
Family Cites Families (103)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US499705A (en) * | 1893-06-20 | Shears | ||
US3506327A (en) | 1964-04-23 | 1970-04-14 | Battelle Development Corp | Wavefront reconstruction using a coherent reference beam |
US3950103A (en) | 1972-10-27 | 1976-04-13 | Canadian Patents And Development Limited | Method and apparatus to determine spatial distribution of magnitude and phase of electro-magnetic fields especially optical fields |
DE2640140C2 (en) | 1976-09-07 | 1982-10-07 | Philips Patentverwaltung Gmbh, 2000 Hamburg | Method and arrangement for redundancy-reducing image coding |
DE2640157C2 (en) | 1976-09-07 | 1982-10-07 | Philips Patentverwaltung Gmbh, 2000 Hamburg | Method and arrangement for redundancy-reducing coding of pictures |
US4136954A (en) | 1976-12-29 | 1979-01-30 | Jamieson John A | Imaging apparatus including spatial-spectral interferometer |
US4223354A (en) | 1978-08-30 | 1980-09-16 | General Electric Company | Phase corrected raster scanned light modulator and a variable frequency oscillator for effecting phase correction |
US4393456A (en) | 1981-03-19 | 1983-07-12 | Bell Telephone Laboratories, Incorporated | Digital filter bank |
DE3171311D1 (en) | 1981-07-28 | 1985-08-14 | Ibm | Voice coding method and arrangment for carrying out said method |
US4674125A (en) | 1983-06-27 | 1987-06-16 | Rca Corporation | Real-time hierarchal pyramid signal processing apparatus |
US4599567A (en) | 1983-07-29 | 1986-07-08 | Enelf Inc. | Signal representation generator |
US4652881A (en) | 1984-01-10 | 1987-03-24 | Lewis Bernard L | Efficient adaptive filter bank |
FR2577084B1 (en) | 1985-02-01 | 1987-03-20 | Trt Telecom Radio Electr | BENCH SYSTEM OF SIGNAL ANALYSIS AND SYNTHESIS FILTERS |
US4701006A (en) | 1985-02-20 | 1987-10-20 | Stanford University | Optical-digital hologram recording |
GB2181318B (en) | 1985-10-04 | 1989-12-28 | Sony Corp | Two-dimensional finite impulse response filters |
US4760563A (en) | 1986-01-09 | 1988-07-26 | Schlumberger Technology Corporation | Seismic exploration using exactly invertible discrete transformation into tau-p space |
US4929223A (en) | 1986-02-18 | 1990-05-29 | Adolph Coors Company | Packaging alignment system |
DE3732085A1 (en) | 1986-03-26 | 1989-04-06 | Ant Nachrichtentech | DIGITAL FILTER TREE |
US4663660A (en) | 1986-06-20 | 1987-05-05 | Rca Corporation | Compressed quantized image-data transmission technique suitable for use in teleconferencing |
GB8621994D0 (en) | 1986-09-12 | 1986-10-22 | Crosfield Electronics Ltd | Image processing |
US4868868A (en) | 1986-09-30 | 1989-09-19 | Oki Electric Industry Co., Ltd. | Sub-band speech analyzing and synthesizing device |
FR2606576B1 (en) | 1986-11-07 | 1989-02-03 | Labo Electronique Physique | DEVICE FOR TRANSMITTING HIGH DEFINITION TELEVISION IMAGES IN NARROW BAND CHANNELS |
GB2197766B (en) | 1986-11-17 | 1990-07-25 | Sony Corp | Two-dimensional finite impulse response filter arrangements |
US4815023A (en) | 1987-05-04 | 1989-03-21 | General Electric Company | Quadrature mirror filters with staggered-phase subsampling |
US4817182A (en) | 1987-05-04 | 1989-03-28 | General Electric Company | Truncated subband coding of images |
BE1000643A5 (en) * | 1987-06-05 | 1989-02-28 | Belge Etat | METHOD FOR CODING IMAGE SIGNALS. |
DE3853555T2 (en) | 1987-06-09 | 1995-08-17 | Sony Corp | Processing of the motion vector in digital television pictures. |
US4837517A (en) | 1987-07-16 | 1989-06-06 | Schlumberger Technology Corporation | Spatial frequency method and apparatus for investigating earth conductivity with high vertical resolution by induction techniques |
US4785349A (en) | 1987-10-05 | 1988-11-15 | Technology Inc. 64 | Digital video decompression system |
US4936665A (en) | 1987-10-25 | 1990-06-26 | Whitney Theodore R | High resolution imagery systems and methods |
US5156943A (en) | 1987-10-25 | 1992-10-20 | Whitney Theodore R | High resolution imagery systems and methods |
US4827336A (en) | 1987-12-18 | 1989-05-02 | General Electric Company | Symbol code generation processing from interframe DPCM of TDM'd spatial-frequency analyses of video signals |
US5095447A (en) | 1988-03-25 | 1992-03-10 | Texas Instruments Incorporated | Color overlay of scanned and reference images for display |
US5018210A (en) | 1988-03-25 | 1991-05-21 | Texas Instruments Incorporated | Pattern comparator with substage illumination and polygonal data representation |
US5001764A (en) | 1988-03-25 | 1991-03-19 | Texas Instruments Incorporated | Guardbands for pattern inspector |
US4985927A (en) | 1988-03-25 | 1991-01-15 | Texas Instruments Incorporated | Method of detecting and reviewing pattern defects |
US4897717A (en) | 1988-03-30 | 1990-01-30 | Starsignal, Inc. | Computer-based video compression system |
EP0339589A3 (en) | 1988-04-28 | 1992-01-02 | Sharp Kabushiki Kaisha | Orthogonal transform coding system for image data |
US4982283A (en) | 1988-05-06 | 1991-01-01 | General Electric Company | Line-sequential pyramid processing of a plurality of raster-scanned image variables |
US4829378A (en) | 1988-06-09 | 1989-05-09 | Bell Communications Research, Inc. | Sub-band coding of images with low computational complexity |
US4904073A (en) | 1988-08-10 | 1990-02-27 | Aware, Inc. | Fractal tiling for multiple mirror optical devices |
FR2637400B1 (en) | 1988-09-30 | 1990-11-09 | Labo Electronique Physique | DEVICE FOR IMPROVED PROCESSING OF AN ECHOGRAPHIC SIGNAL |
FR2643986B1 (en) | 1989-03-03 | 1991-05-17 | Thomson Csf | WAVES SIGNAL ANALYSIS METHOD |
JPH02305182A (en) | 1989-05-19 | 1990-12-18 | Fuji Photo Film Co Ltd | Picture signal compressing and encoding device |
US5072308A (en) | 1989-06-21 | 1991-12-10 | International Mobile Machines Corporation | Communication signal compression system and method |
US4987480A (en) | 1989-07-11 | 1991-01-22 | Massachusetts Institute Of Technology | Multiscale coding of images |
US4974187A (en) | 1989-08-02 | 1990-11-27 | Aware, Inc. | Modular digital signal processing system |
US5073964A (en) | 1989-08-04 | 1991-12-17 | Aware, Inc. | Signal processing device and method |
US5241395A (en) | 1989-08-07 | 1993-08-31 | Bell Communications Research, Inc. | Adaptive transform coding using variable block size |
US5014134A (en) | 1989-09-11 | 1991-05-07 | Aware, Inc. | Image compression method and apparatus |
US5173880A (en) | 1989-12-26 | 1992-12-22 | Exxon Production Research Company | Method of generating seismic wavelets using seismic range equation |
US5068911A (en) | 1990-02-09 | 1991-11-26 | Aware, Inc. | Method and apparatus for representing an image |
US5103306A (en) | 1990-03-28 | 1992-04-07 | Transitions Research Corporation | Digital image compression employing a resolution gradient |
US4999705A (en) | 1990-05-03 | 1991-03-12 | At&T Bell Laboratories | Three dimensional motion compensated video coding |
DE4016172C1 (en) | 1990-05-19 | 1991-03-28 | Werner 5900 Siegen De Ackermann | |
US5101446A (en) | 1990-05-31 | 1992-03-31 | Aware, Inc. | Method and apparatus for coding an image |
IT1248679B (en) * | 1990-06-01 | 1995-01-26 | Enichem Anic Spa | PHENOL PURIFICATION PROCEDURE |
US5128757A (en) | 1990-06-18 | 1992-07-07 | Zenith Electronics Corporation | Video transmission system using adaptive sub-band coding |
EP0465732B1 (en) | 1990-07-11 | 1996-10-02 | Koninklijke Philips Electronics N.V. | Apparatus for deriving a compatible low-definition interlaced television signal and other components from an interlaced high-definition television signal and apparatus for reconstructing the original signal |
US5148498A (en) | 1990-08-01 | 1992-09-15 | Aware, Inc. | Image coding apparatus and method utilizing separable transformations |
US5081645A (en) | 1990-08-06 | 1992-01-14 | Aware, Inc. | Novel spread spectrum codec apparatus and method |
US5128791A (en) | 1990-08-13 | 1992-07-07 | Bell Communications Research, Inc. | Multi-channel HDTV system |
US5097331A (en) | 1990-08-24 | 1992-03-17 | Bell Communications Research, Inc. | Multiple block-size transform video coding using an asymmetric sub-band structure |
US5049992A (en) | 1990-08-27 | 1991-09-17 | Zenith Electronics Corporation | HDTV system with receivers operable at different levels of resolution |
US5049993A (en) | 1990-10-03 | 1991-09-17 | Bell Communications Research, Inc. | Format conversion preprocessing method and circuit |
JP3012698B2 (en) | 1991-01-29 | 2000-02-28 | オリンパス光学工業株式会社 | Image data encoding apparatus and encoding method |
US5121191A (en) | 1991-03-15 | 1992-06-09 | Aware, Inc. | Method and apparatus for coding motion pictures |
US5262958A (en) | 1991-04-05 | 1993-11-16 | Texas Instruments Incorporated | Spline-wavelet signal analyzers and methods for processing signals |
JPH04323963A (en) | 1991-04-23 | 1992-11-13 | Canon Inc | Picture processing method and device |
US5235434A (en) | 1991-06-27 | 1993-08-10 | Polaroid Corporation | Method and apparatus for selectively adjusting the brightness of large regions of an image |
US5347479A (en) | 1991-12-27 | 1994-09-13 | Nec Corporation | Small-size wavelet transform apparatus |
JP3015187B2 (en) | 1992-01-16 | 2000-03-06 | キヤノン株式会社 | Developing sleeve |
JPH06125543A (en) | 1992-10-12 | 1994-05-06 | Toshiba Corp | Encoding device |
US5412741A (en) | 1993-01-22 | 1995-05-02 | David Sarnoff Research Center, Inc. | Apparatus and method for compressing information |
JP3293920B2 (en) | 1993-01-22 | 2002-06-17 | キヤノン株式会社 | Image processing apparatus and method |
IL104636A (en) | 1993-02-07 | 1997-06-10 | Oli V R Corp Ltd | Apparatus and method for encoding and decoding digital signals |
US5381145A (en) | 1993-02-10 | 1995-01-10 | Ricoh Corporation | Method and apparatus for parallel decoding and encoding of data |
US5420891A (en) | 1993-03-18 | 1995-05-30 | New Jersey Institute Of Technology | Multiplierless 2-band perfect reconstruction quadrature mirror filter (PR-QMF) banks |
US5546477A (en) | 1993-03-30 | 1996-08-13 | Klics, Inc. | Data compression and decompression |
EP0622741A3 (en) | 1993-03-30 | 1998-12-30 | KLICS, Ltd. | Device and method for data compression/decompression |
GB2281465B (en) | 1993-08-27 | 1997-06-04 | Sony Uk Ltd | Image data compression |
US5495292A (en) | 1993-09-03 | 1996-02-27 | Gte Laboratories Incorporated | Inter-frame wavelet transform coder for color video compression |
JP3077866B2 (en) * | 1993-11-18 | 2000-08-21 | 日本電気株式会社 | Memory module |
US5453945A (en) | 1994-01-13 | 1995-09-26 | Tucker; Michael R. | Method for decomposing signals into efficient time-frequency representations for data compression and recognition |
US5534925A (en) | 1994-05-02 | 1996-07-09 | Cognitech Inc. | Image compression by optimal reconstruction |
US5606359A (en) | 1994-06-30 | 1997-02-25 | Hewlett-Packard Company | Video on demand system with multiple data sources configured to provide vcr-like services |
US5602589A (en) | 1994-08-19 | 1997-02-11 | Xerox Corporation | Video image compression using weighted wavelet hierarchical vector quantization |
US5699458A (en) | 1995-06-29 | 1997-12-16 | Intel Corporation | Efficient browsing of encoded images |
US5710835A (en) * | 1995-11-14 | 1998-01-20 | The Regents Of The University Of California, Office Of Technology Transfer | Storage and retrieval of large digital images |
JPH09237354A (en) * | 1996-02-29 | 1997-09-09 | Chokosoku Network Computer Gijutsu Kenkyusho:Kk | Method for transferring and displaying three-dimensional shape data |
US5886733A (en) | 1996-05-17 | 1999-03-23 | Sun Microsystems, Inc. | Method and apparatus for successive refinement of broadcasted video frames |
US5982362A (en) | 1996-05-30 | 1999-11-09 | Control Technology Corporation | Video interface architecture for programmable industrial control systems |
US5832300A (en) | 1996-06-20 | 1998-11-03 | Intel Corporation | System for maintaining a minimum level of digitized data signal quality while allowing bandwidth dependent quality enhancement with additional enhancement data packets |
IL119082A (en) | 1996-08-16 | 2001-04-30 | Virtue Ltd | Method for creating graphic images |
US5861920A (en) | 1996-11-08 | 1999-01-19 | Hughes Electronics Corporation | Hierarchical low latency video compression |
US5838377A (en) | 1996-12-20 | 1998-11-17 | Analog Devices, Inc. | Video compressed circuit using recursive wavelet filtering |
US6049821A (en) | 1997-01-24 | 2000-04-11 | Motorola, Inc. | Proxy host computer and method for accessing and retrieving information between a browser and a proxy |
US6038257A (en) | 1997-03-12 | 2000-03-14 | Telefonaktiebolaget L M Ericsson | Motion and still video picture transmission and display |
US5872965A (en) | 1997-06-30 | 1999-02-16 | Sun Microsystems, Inc. | System and method for performing multiway branches using a visual instruction set |
US6049342A (en) | 1997-10-27 | 2000-04-11 | International Business Machines Corporation | System and method for controlling the downloading and display of an animated sequence from a network |
US6009435A (en) | 1997-11-21 | 1999-12-28 | International Business Machines Corporation | Progressive compression of clustered multi-resolution polygonal models |
US6711297B1 (en) * | 1998-07-03 | 2004-03-23 | University Of Pittsburgh - Of The Commonwealth System Of Higher Education | Methods and apparatus for dynamic transfer of image data |
US6314452B1 (en) * | 1999-08-31 | 2001-11-06 | Rtimage, Ltd. | System and method for transmitting a digital image over a communication network |
-
2001
- 2001-12-14 US US10/022,122 patent/US7376279B2/en not_active Expired - Fee Related
- 2001-12-14 AU AU2002229090A patent/AU2002229090A1/en not_active Abandoned
- 2001-12-14 WO PCT/US2001/048933 patent/WO2002048966A1/en not_active Application Discontinuation
Also Published As
Publication number | Publication date |
---|---|
WO2002048966A1 (en) | 2002-06-20 |
US20030005140A1 (en) | 2003-01-02 |
US7376279B2 (en) | 2008-05-20 |
AU2002229090A1 (en) | 2002-06-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US7376279B2 (en) | Three-dimensional image streaming system and method for medical images | |
US7454074B2 (en) | System and method for the lossless progressive streaming of images over a communication network | |
US6314452B1 (en) | System and method for transmitting a digital image over a communication network | |
US6807308B2 (en) | Multi-resolution image data management system and method based on tiled wavelet-like transform and sparse data coding | |
US6873343B2 (en) | Scalable graphics image drawings on multiresolution image with/without image data re-usage | |
EP1072015B1 (en) | Multiresolution compressed image management system and method | |
US8948496B2 (en) | Dynamic transfer of three-dimensional image data | |
US6904176B1 (en) | System and method for tiled multiresolution encoding/decoding and communication with lossless selective regions of interest via data reuse | |
US6347157B2 (en) | System and method for encoding a video sequence using spatial and temporal transforms | |
US20020021758A1 (en) | System and method for efficient transmission and display of image details by re-usage of compressed data | |
WO1996029818A1 (en) | Progressive transmission of images | |
US20020099853A1 (en) | Information processing apparatus, method of controlling the same, information processing system, and computer-readable memory | |
WO2001069585A1 (en) | System and method for efficient transmission and display of image details by re-usage of compressed data | |
Anju et al. | An approach to medical image compression using filters based on lifting scheme | |
AU4522000A (en) | Digital image coding |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ OM PH PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG UZ VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
AK | Designated states |
Kind code of ref document: C2 Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CO CR CU CZ DE DK DM DZ EC EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ OM PH PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG UZ VN YU ZA ZM ZW |
|
AL | Designated countries for regional patents |
Kind code of ref document: C2 Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZM ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GQ GW ML MR NE SN TD TG |
|
COP | Corrected version of pamphlet |
Free format text: PAGES 1/33-33/33, DRAWINGS, REPLACED BY NEW PAGES 1/33-33/33; DUE TO LATE TRANSMITTAL BY THE RECEIVING OFFICE |
|
REG | Reference to national code |
Ref country code: DE Ref legal event code: 8642 |
|
122 | Ep: pct application non-entry in european phase | ||
NENP | Non-entry into the national phase |
Ref country code: JP |
|
WWW | Wipo information: withdrawn in national office |
Country of ref document: JP |