CN104134214A - Digital slice image storing method and digital slice image displaying method based on image pyramid layering - Google Patents

Abstract

The invention relates to a digital information processing technology, in particular to a digital slice image storing method and a digital slice image displaying method based on image pyramid layering, wherein the storing method comprises the following steps that: 1, slice image elements are obtained, and are marked into a zeroth layer, image data in the zeroth layer is divided into data blocks in the dimension of 256*256, and a corresponding coordinate position (X, Y) is recorded; 2, the center of the image elements in a previous layer is used as a base point, the line and row dimension is reduced according to a set proportion, then, sampling is carried out, and the layer number is marked; 3, the second step is repeated until the obtained total data amount reaches the set layer number; and 4, the image elements obtained through sampling are used for building a digital slice pyramid according to the marking sequence. The methods provided by the invention have the advantages that under the condition of the limited network bandwidth, the close progressive of the images can be fast show; and through the image displaying method provided by the invention, the requirement on the browsing terminal equipment performance is greatly reduced, and the digital slice browsing and displaying efficiency is enabled to reach the clinic diagnosis requirements.

Description

A kind of storage of digital slices image and display packing based on image pyramid layering

Technical field

The present invention relates to digital slices image processing techniques, particularly a kind of storage of digital slices image and display packing based on image pyramid layering.

Background technology

Current domestic basic hospital pathologist lacks, pathological diagnosis level is limited, high-quality pathological diagnosis resource mainly concentrates on big city large hospital, the patient of basic unit wants to obtain pathological diagnosis accurately and by patient or kith and kin, pathological section is delivered to higher level's hospital pathology department in person often and hold a consultation, this had both allowed patient expend the plenty of time and had waited for diagnostic result, also will allow patient pay more expenses.Numeral pathology can realize basic unit's pathologist and patient stays indoor enough, can obtain pathological diagnosis suggestion accurately, realizes real " one-stop Pathology Consultation service ", alleviates the present situation of China's pathology uneven geographical distribution of resources.

Pathology slide is by the digital slices in the full visual field of the seamless spliced generation of digital pathological section scanner.Digital slices image is different from conventional medical examination image, and it belongs to large capacity image data.Generally, the data volume that 1.5cm × 1.5cm pathology slide sample 40x multiplying power scanning produces is greatly about 5G left and right, how efficient storage digital slices image realize digital slices fast browsing and show it is one of digital pathology propagation and employment key issue that must solve under common local or wide area network condition.

Summary of the invention

The object of this invention is to provide a kind of efficient storage of digital slices image and display packing based on image pyramid layering fast.

The object of the present invention is achieved like this:

A digital slices image storage method based on image pyramid layering, comprising:

Step is 1.: obtain the pictorial element of section, be labeled as the 0th layer, and these tomographic image data are divided into 256 × 256 dimensional data pieces, and record this respective coordinates position, tomographic image center (X, Y);

Step is 2.: the pictorial element to front one deck is sampled: the pictorial element center of former one deck is basic point, dwindle row, column size according to the ratio of setting, sample again and the mark number of plies, these tomographic image data are divided into 256 × 256 dimensional data pieces simultaneously, and respective coordinates position, records center position (X ', Y ');

Step is 3.: 2. repeating step has reached the number of plies setting to the data total amount that sampling obtains;

Step is 4.: the pictorial element that sampling is obtained is by the sequential build digital slices pyramid of mark, and the pictorial element that first sampling obtains is positioned at the pictorial element below that post-sampling obtains.

Preferably, described step 2. in, the sequence number of mark increases progressively by natural number order.

Preferably, described ratio is 0.5.

Preferably, 7 layers of digital slices pyramid data storages.

Preferably, data sampling method adopts improved Wavelet Transformation Algorithm to realize:

Make f (x, y) ∈ L ²represent a 2D signal, x, y represents respectively its horizontal ordinate and ordinate.Φ (x, y) represents two-dimentional wavelet, and the continuous wavelet transform of f (x, y) is:

$\mathrm{WT}(a;b_1,b_2)=\frac{1}{a}{\∫}_{-\∞}^{+\∞}{\∫}_{-\∞}^{+\∞}f(x,y)\mathrm{\Φ}(\frac{x-b_1}{a},\frac{y-b_2}{a})\mathrm{dxdy}---\left(1\right)$

Wherein, b ₁, b ₂that a is scale factor along the shift factor of image horizontal ordinate and ordinate direction, WT (a; b ₁, b ₂) for by the wavelet coefficient obtaining after converting.

When 2-d wavelet meets admissibility condition:

$c_{\mathrm{\&Phi;}}=\frac{1}{4{\mathrm{\&pi;}}^2}\&Integral;\&Integral;\frac{{\left|\mathrm{\&Phi;}({\mathrm{\&omega;}}_1,{\mathrm{\&omega;}}_2)\right|}^2}{|{{\mathrm{\&omega;}}_1}^2+{{\mathrm{\&omega;}}_2}^2|}{\mathrm{d\&omega;}}_1{\mathrm{d\&omega;}}_2<\&infin;---\left(2\right)$

There is corresponding wavelet inverse transformation, that is:

$f(x,y)=\frac{1}{c_{\mathrm{\&Phi;}}}{\&Integral;}_0^{+\&infin;}{\&Integral;}_{-\&infin;}^{+\&infin;}{\&Integral;}_{-\&infin;}^{+\&infin;}\frac{1}{a^3}\mathrm{WT}(a;b_1,b_2){\mathrm{\&Phi;}}^{*}(\frac{x-b_1}{a},\frac{y-b_2}{a}){\mathrm{db}}_1{\mathrm{db}}_2\mathrm{da}---\left(3\right)$

In actual applications, conventional wavelet function is the binary function of variable separable:

Φ(x，y)＝Φ ₁(x)Φ ₂(y)，

Derive two-dimentional multiscale analysis by one dimension multiscale analysis, and then derive 2-d wavelet space and wavelet function.F (x, y) is expressed with following breakdown:

$f(x,y)={\mathrm{\&Sigma;}}_{k,m}{c}_{k,m}^j{\mathrm{\&Phi;}1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)+{\mathrm{\&Sigma;}}_{j-1}^j{\mathrm{\&Sigma;}}_{j,m}[a_{k,m}^j{w1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)+b_{k,m}^j{w1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)+r_{k,m}^j{w1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)]---\left(4\right)$

Wherein: $c_{k,m}^j=\&Integral;\&Integral;f(x,y)\stackrel{\&OverBar;}{{\mathrm{\&Phi;}1}_{j,m}\left(x\right)}\stackrel{\&OverBar;}{{\mathrm{\&Phi;}2}_{j,m}\left(y\right)}\mathrm{dxdy}$

$a_{k,m}^j=\&Integral;\&Integral;f(x,y)\stackrel{\&OverBar;}{{w1}_{j,m}\left(x\right)}\stackrel{\&OverBar;}{{\mathrm{\&Phi;}2}_{j,m}\left(y\right)}\mathrm{dxdy}$

$b_{k,m}^j=\&Integral;\&Integral;f(x,y){\mathrm{\&Phi;}1}_{j,m}\left(x\right){w2}_{j,m}\left(x\right)\mathrm{dxdy}$

$r_{k,m}^j=\&Integral;\&Integral;f(x,y){w1}_{j,m}\left(x\right){w2}_{j,m}\left(x\right)\mathrm{dxdy}---\left(5\right)$

In formula (4), Section 1 is the Continuous Approximation of f (x, y) under yardstick j, and j=0, when 1,2,3.... different value, obtains approaching under different scale.Formula (5) is called the discrete details of f (x, y) under yardstick, i.e. data after sampling.

A digital slices method for displaying image based on image pyramid layering, comprises

Step is 1.: user end to server system sends data request service, and described data request service possesses parameters number section identification id, the current data block place number of plies, current data block X coordinate figure, current data block Y coordinate figure;

Step is 2.: server system responds the data request service receiving, and the digital slices of ID under opening sends video data block corresponding to (X, Y) coordinate figure in the corresponding number of plies of digital slices pyramid pictorial element;

Step is 3.: the image display area in client shows the video data block obtaining according to specific dimensions.

Preferably, the image of described client is shown as the data block of 256 × 256 pixels.

Preferably, described client is provided with caching mechanism, and it comprises not browsing information prestrain mechanism and browsing information memory mechanism, in the time of the new view data of client-requested, first detects client buffer district and whether loads desired data; If not, 1. 2. 3. execution step, if load, transfers view data from buffer zone.

The present invention compared to existing technology outstanding and useful technique effect is:

The present invention adopts the mode that builds image pyramid layering to present digital slices image, can be under limited network bandwidth condition, can present fast closely going forward one by one of image, and method for displaying image of the present invention utilizes not browsing information prestrain mechanism greatly to alleviate the demand to browsing terminal equipment performance, make the present invention in actual application, ensure that digital slices is browsed with display efficiency and reach clinical diagnosis requirement.

Brief description of the drawings

Fig. 1 is the digital slices image schematic diagram of the image pyramid layering that builds of the present invention.

Fig. 2 is caching mechanism schematic diagram of the present invention.

Embodiment

Below in conjunction with accompanying drawing, with specific embodiment, the invention will be further described:

A digital slices image storage method based on image pyramid layering, comprising:

Step is 1.: obtain the pictorial element of section, be labeled as the 0th layer, and these tomographic image data are divided into 256 × 256 dimensional data pieces, and record this respective coordinates position, tomographic image center (X, Y);

Step is 2.: the pictorial element to front one deck is sampled: the pictorial element center of former one deck is basic point, dwindle row, column size according to the ratio of setting, sample again and the mark number of plies, these tomographic image data are divided into 256 × 256 dimensional data pieces simultaneously, and record respective coordinates position (X ', Y ');

Step is 3.: the data total amount that 2. repeating step obtains to sampling has reached the number of plies setting or set the row, column size of sampling;

Step is 4.: the pictorial element that sampling is obtained is by the sequential build digital slices pyramid of mark, the pictorial element that first sampling obtains is positioned at pictorial element below (the 0th layer of storage raw data of the present invention that post-sampling obtains, on the basis of the 0th layer, generate i (i=1,2,) layer pyramid data, data volume in last layer pyramid is 1/4 of next layer data, by that analogy, until data save as the 7th layer).

Preferably, described step 2. in, the sequence number of mark increases progressively by natural number order.

Preferably, described ratio is 0.5, that is, the row, column size of last layer pictorial element is the half of the row, column size of next tomographic image element.

Its concrete building process is as follows:

(1) at the bottom of the tower of regulation image pyramid, be 0 grade, at the bottom of original image is image pyramid tower, and be divided into 256 × 256 Pixel Dimensions size data pieces;

(2) from the bottom of tower, whenever the number of plies adds 1, last tomographic image is carried out to resampling according to position corresponding relation, obtain i tomographic image, the computing formula of i tomographic image size is as follows:

r(i)＝|r(0)/2 ⁱ|

c(i)＝|c(0)/2 ⁱ|

In formula: the row, column size of raw video at the bottom of r (0) and c (0) expression tower, the i.e. height and width of raw video, r (i) and c (i)---the row, column size of i level image in tower.

As shown in Figure 1, in figure, (a) is the 0th grade of image, and in figure, (b) is the 1st grade of image, and in figure, (c) is 3rd level image.

(3) sampling of concrete data adopts improved wavelet algorithm to realize:

Make f (x, y) ∈ L ²represent a 2D signal, x, y represents respectively its horizontal ordinate and ordinate.Φ (x, y) represents two-dimentional wavelet, and the continuous wavelet transform of f (x, y) is:

$\mathrm{WT}(a;b_1,b_2)=\frac{1}{a}{\&Integral;}_{-\&infin;}^{+\&infin;}{\&Integral;}_{-\&infin;}^{+\&infin;}f(x,y)\mathrm{\&Phi;}(\frac{x-b_1}{a},\frac{y-b_2}{a})\mathrm{dxdy}---\left(1\right)$

Wherein, b ₁, b ₂that a is scale factor along the shift factor of image horizontal ordinate and ordinate direction, WT (a; b ₁, b ₂) for by the wavelet coefficient obtaining after converting.

When 2-d wavelet meets admissibility condition:

$c_{\mathrm{\&Phi;}}=\frac{1}{4{\mathrm{\&pi;}}^2}\&Integral;\&Integral;\frac{{\left|\mathrm{\&Phi;}({\mathrm{\&omega;}}_1,{\mathrm{\&omega;}}_2)\right|}^2}{|{{\mathrm{\&omega;}}_1}^2+{{\mathrm{\&omega;}}_2}^2|}{\mathrm{d\&omega;}}_1{\mathrm{d\&omega;}}_2<\&infin;---\left(2\right)$

There is corresponding wavelet inverse transformation, that is:

$f(x,y)=\frac{1}{c_{\mathrm{\&Phi;}}}{\&Integral;}_0^{+\&infin;}{\&Integral;}_{-\&infin;}^{+\&infin;}{\&Integral;}_{-\&infin;}^{+\&infin;}\frac{1}{a^3}\mathrm{WT}(a;b_1,b_2){\mathrm{\&Phi;}}^{*}(\frac{x-b_1}{a},\frac{y-b_2}{a}){\mathrm{db}}_1{\mathrm{db}}_2\mathrm{da}---\left(3\right)$

In actual applications, conventional wavelet function is the binary function of variable separable:

Φ(x，y)＝Φ ₁(x)Φ ₂(y)，

Derive two-dimentional multiscale analysis by one dimension multiscale analysis, and then derive 2-d wavelet space and wavelet function.F (x, y) is expressed with following breakdown:

$f(x,y)={\mathrm{\&Sigma;}}_{k,m}{c}_{k,m}^j{\mathrm{\&Phi;}1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)+{\mathrm{\&Sigma;}}_{j-1}^j{\mathrm{\&Sigma;}}_{j,m}[a_{k,m}^j{w1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)+b_{k,m}^j{w1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)+r_{k,m}^j{w1}_{j,m}\left(x\right){\mathrm{\&Phi;}2}_{j,m}\left(x\right)]---\left(4\right)$

Wherein: $c_{k,m}^j=\&Integral;\&Integral;f(x,y)\stackrel{\&OverBar;}{{\mathrm{\&Phi;}1}_{j,m}\left(x\right)}\stackrel{\&OverBar;}{{\mathrm{\&Phi;}2}_{j,m}\left(y\right)}\mathrm{dxdy}$

$a_{k,m}^j=\&Integral;\&Integral;f(x,y)\stackrel{\&OverBar;}{{w1}_{j,m}\left(x\right)}\stackrel{\&OverBar;}{{\mathrm{\&Phi;}2}_{j,m}\left(y\right)}\mathrm{dxdy}$

$b_{k,m}^j=\&Integral;\&Integral;f(x,y){\mathrm{\&Phi;}1}_{j,m}\left(x\right){w2}_{j,m}\left(x\right)\mathrm{dxdy}$

$r_{k,m}^j=\&Integral;\&Integral;f(x,y){w1}_{j,m}\left(x\right){w2}_{j,m}\left(x\right)\mathrm{dxdy}---\left(5\right)$

In formula (4), Section 1 is the Continuous Approximation of f (x, y) under yardstick j, and j=0, when 1,2,3.... different value, obtains approaching under different scale.Formula (5) is called the discrete details of f (x, y) under yardstick, i.e. data after sampling.

(4) data after sampling are divided into 256 × 256 Pixel Dimensions size data pieces successively.

(5), until be building up to pinnacle of a pagoda, when 7 layers of digital slices data storages, i.e. progression N=6.

On basis based on image pyramid layering, in digital slices procedure for displaying, digital slice of data is carried out to piecemeal processing.In the time carrying out dynamic load, as required the data block of the data of layering and piecemeal is combined, data request service is made to response, under different displaying ratios and resolution, carry out roaming switch, realize dynamic on-demand and load map datum, greatly improve system real time.

A digital slices method for displaying image based on image pyramid layering, comprises

Step is 1.: user end to server system sends data request service, and described data request service possesses parameters number section identification id, the current data block place number of plies, current data block X coordinate figure, current data block Y coordinate figure;

Step is 2.: server system responds the data request service receiving, the digital slices of ID under opening, send video data block corresponding to (X, Y) coordinate figure in the corresponding number of plies of the digital slices pyramid pictorial element building with described image storage method;

Step is 3.: the image display area in client shows the video data block obtaining.

Preferably, the image of described client is shown as the data block of 256 × 256 pixels.

Preferably, described client is provided with caching mechanism, and it comprises not browsing information prestrain mechanism and browsing information memory mechanism, in the time of the new view data of client-requested, first detects client buffer district and whether loads desired data; If not, 1. 2. 3. execution step, if load, transfers view data from buffer zone.

At present, our digital slices image adopts pyramid layering storage, supposes that the image number of plies is 3 layers, and every layer of dimension is 2, entire image 2G, as shown in Figure 2:

In image loading procedure, can first load " low frequency range of the low frequency range of low frequency range " to buffer zone, when user initially uses, call in this parts of images datarams and show image, load " low frequency range of the low frequency range of low frequency range " other 3/4ths high-frequency informations simultaneously; In the time that user browses more specific location information, call other 3/4ths high-frequency informations of demonstration " low frequency range of the low frequency range of low frequency range ", make it further see image detail clearly, simultaneously other 3/4ths high-frequency informations of backstage loaded in parallel " low frequency range of low frequency range "; By that analogy.

In actual use, buffer zone is left in the region that user browsed in, and time limit is 24 hours, and buffer size can be set to 5GB, before image shows, needs to detect this locality and whether possesses space.

Above-described embodiment is only preferred embodiment of the present invention, not limits the scope of the invention according to this, therefore: all equivalences of doing according to structure of the present invention, shape, principle change, within all should being covered by protection scope of the present invention.

Claims (7)

1. the digital slices image storage method based on image pyramid layering, it is characterized in that, comprise: step 1.: the pictorial element that obtains section, be labeled as the 0th layer, and these tomographic image data are divided into 256 × 256 dimensional data pieces, and record this respective coordinates position, tomographic image center (X, Y);

Step is 2.: the pictorial element to front one deck is sampled: the pictorial element center of former one deck is basic point, dwindle row, column size according to the ratio of setting, sample again and the mark number of plies, these tomographic image data are divided into 256 × 256 dimensional data pieces simultaneously, and respective coordinates position, records center position (X ', Y ');

Step is 3.: 2. repeating step has reached the number of plies setting to the data total amount that sampling obtains;

Step is 4.: the pictorial element that sampling is obtained is by the sequential build digital slices pyramid of mark, and the pictorial element that first sampling obtains is positioned at the pictorial element below that post-sampling obtains.

2. a kind of storage means of simulating pyramidal digital slices image according to claim 1, is characterized in that, described step 2. in, the sequence number of mark increases progressively by natural number order.

3. a kind of storage means of simulating pyramidal digital slices image according to claim 1, is characterized in that, described ratio is 0.5.

4. a kind of storage means of simulating pyramidal digital slices image according to claim 1, is characterized in that, 7 layers of described digital slices pyramid data storages.

5. the digital slices method for displaying image based on image pyramid layering, it is characterized in that, comprise that step 1.: user end to server system sends data request service, described data request service possesses parameters number section identification id, the current data block place number of plies, current data block X coordinate figure, current data block Y coordinate figure;

Step is 2.: server system responds the data request service receiving, the digital slices of ID under opening, send in the corresponding number of plies (X, Y) coordinate figure corresponding, and the video data block of the digital slices pyramid pictorial element building with image storage method described in claim 1;

Step is 3.: the image display area in client shows the video data block obtaining according to specific dimensions.

6. a kind of digital slices method for displaying image based on image pyramid layering according to claim 5, it is characterized in that, described client is provided with caching mechanism, it comprises not browsing information prestrain mechanism and browsing information memory mechanism, in the time of the new view data of client-requested, first detect client buffer district and whether load desired data; If not, 1. 2. 3. execution step, if load, transfers view data from buffer zone.

7. a kind of digital slices method for displaying image based on image pyramid layering according to claim 5, is characterized in that, the image of described client is shown as the data block of 256 × 256 pixels.