WO2004040503A1 - A computerized method for discovering patterns in unknown multi-dimensional data - Google Patents
A computerized method for discovering patterns in unknown multi-dimensional data Download PDFInfo
- Publication number
- WO2004040503A1 WO2004040503A1 PCT/JP2003/012783 JP0312783W WO2004040503A1 WO 2004040503 A1 WO2004040503 A1 WO 2004040503A1 JP 0312783 W JP0312783 W JP 0312783W WO 2004040503 A1 WO2004040503 A1 WO 2004040503A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- dimensional data
- time series
- minimum cost
- data
- distance matrix
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7847—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content
- G06F16/785—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content using colour or luminescence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/73—Querying
- G06F16/738—Presentation of query results
- G06F16/739—Presentation of query results in form of a video summary, e.g. the video summary being a video sequence, a composite still image or having synthesized frames
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7834—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using audio features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7847—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content
- G06F16/7857—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content using texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/783—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
- G06F16/7847—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content
- G06F16/786—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content using low-level visual features of the video content using motion, e.g. object motion or camera motion
Definitions
- This invention relates in general to data analysis, and more particularly, to a method for discovering patterns in time series data.
- Data mining extracts information from large-scale databases and the Internet. Data mining has been applied to the analysis of market, financial, sensor, and biological data. Data mining should not be confused with pattern matching where data are searched for known patterns.
- Data mining derives rules, trends, regularities and correlations from a large volume of data. Often, data mining is based on artificial intelligence (Al), memory based reasoning (MBR), association rule generation, decision trees (DT), neural analysis, statistical analysis, clustering, and time series analysis.
- Al artificial intelligence
- MLR memory based reasoning
- DT decision trees
- neural analysis statistical analysis
- clustering and time series analysis.
- Clustering identifies homogeneous groups of related information in data. Prior art clustering assumes that relationships among the data are known. Clustermg has been studied extensively in statistics, pattern recognition, and machine learning. Examples of clustering applications include customer segmentation for marketing analysis, and identification of sub-categories of signal databases in sensed data.
- Clustering techniques can be broadly classified into partitional techniques and hierarchical techniques. Partitional clustering separates data into K clusters such that the data in each cluster are more similar to each other than to data in different clusters.
- the value of K can be assigned by a user, or iteratively determined so as to minimize the clustering criterion.
- Hierarchical clustering is a nested sequence of partitions. Agglomerative hierarchical clustering places data in atomic clusters and then merges the atomic clusters into larger and larger clusters until all data are in a single large cluster. Divisive hierarchical clustering reverses the process by starting with all data in one cluster and subdividing the cluster into smaller clusters, see for example, Jain et al., "Algorithms for Clustering Data," Prentice Hall, 1988, Piramuthu et al, “Comparison of SOM neutral network and hierarchical clustering methods " European Journal of Operational Research, 93(2):402-417, September 1996, Michaud, "Four clustering techniques," FGCS Journal, Special Issue on Data Mining, 1997, and Zait et al., “A Comparative study of clustering methods " FGCS Journal, Special Issue on Data Mining, 1997.
- a popular data mining method used for analysis of consumer buying patterns is the identification of non-obvious associations, or association rules.
- An example of an obvious association is that consumers who buy baby formula also buy diapers at the same time. However, it was discovered in 1992 that beer and diapers are often purchased together in the evening hours. Such an association is a good example of a non-obvious association. Normally, one would not associate diapers and beer as strongly related purchase items because beer is usually not considered a baby formula. Such an approach has also been termed market-basket analysis.
- An association is defined as follows. If there is a set of n items I t ..., J n> and a transaction, e.g., a database or operation, that selects a subset of the n items, then an association between two items I ⁇ and I ⁇ - is defined as a rule R for any transaction in both items I ⁇ and / j are selected into the subset.
- a condition of the rale R is defined as the occurrence of item /;.
- a result of the rule R is defined as the occurrence of the item I-
- a support of the rule R is defined as a percentage of the transactions that have both items /; and T
- a combination of the rule R is defined as the occurrence of both items I and I ⁇ - in the same transaction.
- a confidence of the rule R is defined as a ratio of the support of the combination and the support of the condition.
- an improvement of the rule R is defined as a ratio of the support of the rule over the product of the support of the condition /; and the support of the result I
- An association is strong when both the support and the confidence of the association are high. For example, for a grocery store transaction, milk is strongly correlated with every other item sold so the support is high, because most transactions include milk as an item. However, for some items, such as greeting cards, the confidence is low because these are bought infrequently.
- Time series analysis correlates data values as a function of their temporal separation. For example, time series analysis has been used to discover patterns in stock prices, sales volumes, climate data, and EKG analysis. It is generally assumed that there is an underlying deterministic process that generated the time series and that that process is not random. However, time series data of real-world phenomena is often intermixed with non-deterministic data, for example, unavoidable random noise.
- time series are compared using a similarity measure such as the Euclidean distances, or some variation thereof.
- Euclidean distance measurements tend to be unreliable.
- a more robust similarity measure is based on dynamic time warping (DTW), see Berndt et al., "Finding patterns in time series: a dynamic programming approach " Advances in Knowledge Discovery and Data Mining. AAAI/MIT Press, Menlo Park, CA. pp. 229-248, 1966.
- DTW attempts to align time series data by selectively stretching and shrinking the time axis.
- the particular frame selected is usually a good representative of the other frames in the shot according to some criterion.
- Other techniques exploit changes in camera angles or field of view, e.g., zooming.
- Such content- neutral techniques have had moderate success and require supplementary content- specific techniques for semantically satisfactory performance.
- CART classification and regression trees
- Video mining would be particularly useful for discovering interesting patterns in videos where an a priori model of the domain and content, such as editing models, characteristics of a news video, or patterns in a sports video, etc., are not readily available. Disclosure of Invention
- a method discovers patterns in unknown multi-dimensional data.
- a time- series of the multi-dimensional data is generated and a point cross-distance matrix is constructed by self-correlating the time-series.
- the minimum cost paths are then related to sub-sequences in the multidimensional data to discover high-level patterns in the unknown multi-dimensional data.
- Figure 1 is a flow diagram of a video mining method 100 according to the invention.
- Figure 2 is a time series diagram of a golf video
- Figure 3 is a point cross-distance matrix used by the invention.
- Figure 4 is a point cross-distance matrix with repeating patterns
- Figure 5 is a graph of a near-diagonal minimum cost path
- Figure 6 is a graph of a minimum path with threshold angles.
- Figure 1 shows the steps for a video mining method 100 according to our invention.
- Our invention is particularly useful for mining videos where patterns in the content are not known a priori, such as surveillance or home videos.
- the discovered patterns can be used for "blind" summarization of videos.
- blind summarization we mean that we do not need to know the content or genre of the video to be summarized.
- Our method is also useful for carefully prepared videos such as broadcast news, broadcast sports, feature films, etc.
- our method enables blind summarization of videos with a diverse range of variations in content and production styles.
- a video 101 of arbitrary, and perhaps, unknown content and production style The video includes visual and audio data.
- a low-level feature 102 of the content e.g., a group of features can consist of motion activity, color, audio, texture etc. such as MPEG-7 descriptors.
- Mining can be performed with one or more selected features 102, and the results 152 of each mining pass can be combined to infer 160 a higher level of understanding of the input video 101.
- We use the selected feature generate a time series 200 for the video 101.
- Figure 2 is an example of the time series 200 of a video of a golf game based on motion activity.
- Figure 2 plots the level of motion activity in a golf video as a function of time.
- Time is indicated on the x-axis 201 in terms of frame numbers, and motion activity is indicated along the y-axis 202 as the average number of pixels moved in a macro-block since a previous frame.
- the raw motion activity is indicated by the plot 203.
- time series A color histogram of the video frames is another multi-dimensional time series.
- Time-series data in a video can also include symbolic values such as semantic labels, e.g., MPEG-7 descriptors, and audio signals.
- time series we mean that the series data has some type of ordering along the x-axis, e.g., temporal.
- Other series ordering, such as spatial are also possible, and our invention is equally applicable to "space" series or other ordered series of multi-dimensional data. Of course in a space series, all paths will be at multiple spatial resolutions.
- Our method 100 measures 130 similar expressions of the low-level features based on the time series 200 of the video, e.g., high, medium, or low in the case of motion, dominant blue, green, or brown colors could be another expression.
- a distance 301 between two windows 302-303 of size w is measured as: w-l
- Figure 3 shows the example point distance matrix 300 where one window 301 is Xj - x i+w and another window 302 is X j - x j+w .
- the line 301 represents equation (1).
- the curves 311 and 312 represent respectively the time series 200. That is, the same time series is along the x - and y-axis, in other words the measurement is self-correlating.
- the point distance matrix 300 can be searched for diagonal lines of short point cross-distances.
- the distances are rendered on a gray-scale where shorter distances are lighter than large distance, hence, the major diagonal where all distances are zero is white, because the measurement is self-correlated.
- the matrix 300 is not limited to any given window size w. Therefore, it is possible to find all possible diagonal paths of short distance values, where short is with respect to some threshold distance value Tp. If time warping is used the path can be "near" or substantially diagonal to the main diagonal 310 in Figure 3.
- Dynamic time warping is used to "stretch” and “compress” time, within certain limits, to allow for a good alignment between similar segments of the video having different lengths of time.
- segment we mean a sequence of frames.
- Dynamic programming can be used to find the time warping of the segments that give a best match. DTW finds an optimum path in the grid under certain constraints
- N and M are the lengths of two segments, and each grid point ( , ) is a distance between the th and h points in the two segments, respectively.
- a possible definition of the minimum cost is an average of the cross- distance values along the path.
- One possible set of constraints is that, at each step along the path can move down one, right one, or diagonally, as shown.
- a threshold angle between the diagonal, and the line drawn between the two end points of the path should be less than a , where 0 ⁇ a ⁇ 45°.
- Such path cost criteria and constraints can be formulated by dynamic programming to find the minimum cost valid path between two points in the point cross-distance matrix.
- a valid path thus obtained even if it is not exactly diagonal, specifies a dynamic time warping of two corresponding sub-sequences in the time series. Hence, we find the appropriate time warping for a good match between two subsequences by using dynamic programming, and find the minimum cost path between two points in the cross-distance matrix.
- Dynamic programming allows us to find the minimum cost path between two given points in the matrix. We are interested in finding the minimum cost paths between all possible pairs of points in the matrix, and select the ones that have a cost below a given threshold.
- the similarity using dynamic time warping (e.g., by finding a minimum cost path in the cross-distance matrix using dynamic programming) is done for querying a given time sequence pattern in a larger time sequence, see Park et al., "Efficient search for similar subsequences of different lengths in sequence databases, International Conference on Data Engineering (ICDE), pp. 23-32, 2000.
- cross-distance matrix is in the form of dot-plots where the matrix elements are either 1 or 0, indicating an exact match or no match. They have been used for qualitative analysis and as visual tools, see Church et al., "Dotplot: A program for exploring self-similarity in millions of lines of text and code " Journal of Computational and Graphical Statistics, 2(2): 153--174, 1993., or for segmenting self-similar regions along the diagonal but not for automatically finding pairs of similar sub-sequences, see Jeffry, "An Automatic Method of Finding Topic Boundaries” Proceedings of the 32" Annual Meeting of the Association for Computational Linguistics, 1994. Our method can be used for any of these applications.
- a Hough transform could be applied to merge multiple short paths into curves.
- the Hough transform isolates features of a particular shape within the matrix 400.
- We could use the Hough transform because a simple analytic description of the features is not possible.
- the Hough transform is tolerant to gaps in feature boundary descriptions and is relatively unaffected by noise in the data.
- the Hough transform is particularly useful for discovering a global description of the features, where the number of solution classes is not known a priori, as is the case here.
- label and "item,” as defined above for associations, equivalently.
- clusters we imply that the clusters are labeled, e.g., a cluster of high motion activity frames, or a cluster of loud noise in the sound track of the video.
- associations and fusion of labels 152 obtained using different features allows us to infer 160 high-level patterns 161 in the clusters 151. These patterns can then be used for blind summarization 170 of the video 101 into a summary 171.
- the summarization essentially extracts selected frames of the video according to the high-level patterns 161.
- step 160 we find associations between labels of the clusters.
- One type of association looks for "labels" that correlated in time.
- the correlation can either be simultaneous, or time shifted, e.g., a lot of motion and loud noise signals an "action” event in an entertainment video, or a lot of motion later followed by applause and cheering signals a "scoring" event in a sports video.
- the summary 171 of the video 101 contains a combination of the common and uncommon events depending on the needs of the user.
- the summary 171 contains rare events with strong associations.
- the summary contains common events with strong associations.
- the clustering 150 begins by clustering all similar segments based on the distance metric as described above. Partially overlapping segments are divided so that they either have an identical overlap, or none at all. Pairs of clusters can be merged using an agglomerative procedure to form bigger clusters.
- a set S to contain all segments that are in at least one of the similar pairs of clusters.
- the set Sf U S is the entire video 101, and the set Sf is the complement of the set S.
- the set S° can be considered a "junk" cluster that is not as "meaningful" with respect to the labels A, B, and C, etc.
- the clustering 150 produces the set S of non-overlapping segments of various lengths, and their mutual distances.
- Each cluster 151 has the following associated parameters: cluster compactness, i.e., how similar are the segments in the cluster; number of segments in the cluster; total length of the segments (in frames or time) in the cluster; and average distance between the closest points of the segments which indicates the spreading of the segments in the entire video.
- Each segment in a cluster has the following parameters: average distance to other cluster center, which indicates how representative the segment is of the cluster in which it is contained; and its length (frames or time).
- a summarization cost function C 172 can trade-off an absolute difference between a required summary length, a total length of a set R of selected segment to be included in the summary 171, (i.e., R c S), an average of the distances between each segment in S and a closest segment in a set R. This indicates how well the set R (summary 171) represents the set S, which, of course, is the entire video 101. We can add a te ⁇ n to favor compactness of R in time, or the opposite.
- the summary 171 then is the subset R of set S that minimizes the cost function 172.
- the summary can then be used for new modalities of video control such as "smart" fast-forward and reverse, and instant repositioning in the face of no a priori domain knowledge.
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2004548015A JP2006505038A (en) | 2002-11-01 | 2003-10-06 | Computerized method for finding patterns in unknown multidimensional data |
EP03748726A EP1444644B1 (en) | 2002-11-01 | 2003-10-06 | A computerized method for discovering patterns in unknown multi-dimensional data |
DE60307224T DE60307224T2 (en) | 2002-11-01 | 2003-10-06 | COMPUTER-BASED METHOD FOR DISCOVERING PATTERNS IN UNKNOWN HIGH DIMENSIONAL DATA |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US10/285,928 US7103222B2 (en) | 2002-11-01 | 2002-11-01 | Pattern discovery in multi-dimensional time series using multi-resolution matching |
US10/285,928 | 2002-11-01 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2004040503A1 true WO2004040503A1 (en) | 2004-05-13 |
Family
ID=32175300
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2003/012783 WO2004040503A1 (en) | 2002-11-01 | 2003-10-06 | A computerized method for discovering patterns in unknown multi-dimensional data |
Country Status (6)
Country | Link |
---|---|
US (1) | US7103222B2 (en) |
EP (1) | EP1444644B1 (en) |
JP (1) | JP2006505038A (en) |
CN (1) | CN100365648C (en) |
DE (1) | DE60307224T2 (en) |
WO (1) | WO2004040503A1 (en) |
Families Citing this family (57)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20030200134A1 (en) * | 2002-03-29 | 2003-10-23 | Leonard Michael James | System and method for large-scale automatic forecasting |
TW569148B (en) * | 2002-04-09 | 2004-01-01 | Ind Tech Res Inst | Method for locating facial features in an image |
WO2004019224A2 (en) * | 2002-08-26 | 2004-03-04 | Koninklijke Philips Electronics N.V. | Unit for and method of detection a content property in a sequence of video images |
US7375731B2 (en) * | 2002-11-01 | 2008-05-20 | Mitsubishi Electric Research Laboratories, Inc. | Video mining using unsupervised clustering of video content |
US7143352B2 (en) * | 2002-11-01 | 2006-11-28 | Mitsubishi Electric Research Laboratories, Inc | Blind summarization of video content |
JP4662779B2 (en) * | 2003-01-06 | 2011-03-30 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | Device for switching to similar video content |
WO2005083890A1 (en) * | 2004-02-26 | 2005-09-09 | Mitsubishi Denki Kabushiki Kaisha | Time series data dimensional compressor |
US7483572B2 (en) * | 2004-08-25 | 2009-01-27 | Mitsubishi Electric Research Laboratories, Inc. | Recovering a non-linear warping function from images |
US8010324B1 (en) | 2005-05-09 | 2011-08-30 | Sas Institute Inc. | Computer-implemented system and method for storing data analysis models |
US7188032B2 (en) | 2005-06-30 | 2007-03-06 | International Business Machines Corporation | Incremental determination of Teiresias patterns |
US7636698B2 (en) * | 2006-03-16 | 2009-12-22 | Microsoft Corporation | Analyzing mining pattern evolutions by comparing labels, algorithms, or data patterns chosen by a reasoning component |
US7711734B2 (en) * | 2006-04-06 | 2010-05-04 | Sas Institute Inc. | Systems and methods for mining transactional and time series data |
US8019162B2 (en) * | 2006-06-20 | 2011-09-13 | The Nielsen Company (Us), Llc | Methods and apparatus for detecting on-screen media sources |
US8112302B1 (en) | 2006-11-03 | 2012-02-07 | Sas Institute Inc. | Computer-implemented systems and methods for forecast reconciliation |
US8542899B2 (en) * | 2006-11-30 | 2013-09-24 | Definiens Ag | Automatic image analysis and quantification for fluorescence in situ hybridization |
US8527324B2 (en) * | 2006-12-28 | 2013-09-03 | Oracle Otc Subsidiary Llc | Predictive and profile learning salesperson performance system and method |
US20090216611A1 (en) * | 2008-02-25 | 2009-08-27 | Leonard Michael J | Computer-Implemented Systems And Methods Of Product Forecasting For New Products |
WO2009154597A1 (en) * | 2008-06-19 | 2009-12-23 | Thomson Licensing | Adaptive video key frame selection |
US9245000B2 (en) * | 2008-08-05 | 2016-01-26 | Vmware, Inc. | Methods for the cyclical pattern determination of time-series data using a clustering approach |
US9547710B2 (en) * | 2008-08-05 | 2017-01-17 | Vmware, Inc. | Methods for the cyclical pattern determination of time-series data using a clustering approach |
US10346617B1 (en) * | 2010-01-25 | 2019-07-09 | Hrl Laboratories, Llc | Protocol for securely searching streaming data with constant bandwidth |
US9613292B1 (en) * | 2012-01-26 | 2017-04-04 | Hrl Laboratories, Llc | Secure multi-dimensional pattern matching for secure search and recognition |
US8631040B2 (en) | 2010-02-23 | 2014-01-14 | Sas Institute Inc. | Computer-implemented systems and methods for flexible definition of time intervals |
US20120209642A1 (en) * | 2011-02-15 | 2012-08-16 | Chicago Mercantile Exchange Inc. | Identification of trading activities of entities acting in concert |
US9336493B2 (en) | 2011-06-06 | 2016-05-10 | Sas Institute Inc. | Systems and methods for clustering time series data based on forecast distributions |
US9047559B2 (en) | 2011-07-22 | 2015-06-02 | Sas Institute Inc. | Computer-implemented systems and methods for testing large scale automatic forecast combinations |
US20130103638A1 (en) * | 2011-10-25 | 2013-04-25 | Chetan Kumar Gupta | Computing a hierarchical pattern query from another hierarchical pattern query |
US8650113B2 (en) | 2011-10-28 | 2014-02-11 | Chicago Mercantile Exchange Inc. | Identification of accounts that are too profitable or too lossy |
US8750613B2 (en) | 2011-12-13 | 2014-06-10 | The Nielsen Company (Us), Llc | Detecting objects in images using color histograms |
US8897553B2 (en) | 2011-12-13 | 2014-11-25 | The Nielsen Company (Us), Llc | Image comparison using color histograms |
US8897554B2 (en) | 2011-12-13 | 2014-11-25 | The Nielsen Company (Us), Llc | Video comparison using color histograms |
US10621364B1 (en) * | 2012-01-26 | 2020-04-14 | Hrl Laboratories, Llc | Generic pattern matching system |
US9075713B2 (en) * | 2012-05-24 | 2015-07-07 | Mitsubishi Electric Research Laboratories, Inc. | Method for detecting anomalies in multivariate time series data |
US9244887B2 (en) | 2012-07-13 | 2016-01-26 | Sas Institute Inc. | Computer-implemented systems and methods for efficient structuring of time series data |
US9087306B2 (en) | 2012-07-13 | 2015-07-21 | Sas Institute Inc. | Computer-implemented systems and methods for time series exploration |
US9147218B2 (en) | 2013-03-06 | 2015-09-29 | Sas Institute Inc. | Devices for forecasting ratios in hierarchies |
KR102223419B1 (en) | 2013-06-14 | 2021-03-04 | 마이크로소프트 테크놀로지 라이센싱, 엘엘씨 | Fast grouping of time series |
KR101554216B1 (en) * | 2013-06-18 | 2015-09-18 | 삼성에스디에스 주식회사 | Method and apparatus thereof for verifying bad patterns in sensor-measured time series data |
US9934259B2 (en) | 2013-08-15 | 2018-04-03 | Sas Institute Inc. | In-memory time series database and processing in a distributed environment |
JP2015099505A (en) * | 2013-11-19 | 2015-05-28 | キヤノン株式会社 | Data processing apparatus and data processing method |
US10169720B2 (en) | 2014-04-17 | 2019-01-01 | Sas Institute Inc. | Systems and methods for machine learning using classifying, clustering, and grouping time series data |
US9892370B2 (en) | 2014-06-12 | 2018-02-13 | Sas Institute Inc. | Systems and methods for resolving over multiple hierarchies |
US9208209B1 (en) | 2014-10-02 | 2015-12-08 | Sas Institute Inc. | Techniques for monitoring transformation techniques using control charts |
US20160110478A1 (en) * | 2014-10-17 | 2016-04-21 | General Electric Company | System and methods for quantization and featurization of time-series data |
EP3035274A1 (en) * | 2014-12-17 | 2016-06-22 | Tata Consultancy Services Limited | Interpretation of a dataset |
US9418339B1 (en) | 2015-01-26 | 2016-08-16 | Sas Institute, Inc. | Systems and methods for time series analysis techniques utilizing count data sets |
CN107209770B (en) | 2015-03-17 | 2020-10-30 | 惠普发展公司,有限责任合伙企业 | System and method for analyzing events and machine-readable storage medium |
US10983682B2 (en) | 2015-08-27 | 2021-04-20 | Sas Institute Inc. | Interactive graphical user-interface for analyzing and manipulating time-series projections |
WO2018009733A1 (en) * | 2016-07-07 | 2018-01-11 | Aspen Technology, Inc. | Computer system and method for monitoring key performance indicators (kpis) online using time series pattern model |
US10331490B2 (en) | 2017-11-16 | 2019-06-25 | Sas Institute Inc. | Scalable cloud-based time series analysis |
US10338994B1 (en) | 2018-02-22 | 2019-07-02 | Sas Institute Inc. | Predicting and adjusting computer functionality to avoid failures |
US10255085B1 (en) | 2018-03-13 | 2019-04-09 | Sas Institute Inc. | Interactive graphical user interface with override guidance |
CN109783051B (en) * | 2019-01-28 | 2020-05-29 | 中科驭数(北京)科技有限公司 | Time series similarity calculation device and method |
CN110059735A (en) | 2019-04-04 | 2019-07-26 | 东北大学 | A kind of visualization time series pattern matching process based on Hough transformation |
US11500932B2 (en) | 2020-05-19 | 2022-11-15 | International Business Machines Corporation | Change-point driven feature selection for multi-variate time series clustering |
CN113237951A (en) * | 2021-05-11 | 2021-08-10 | 重庆大学 | Metal plate fatigue damage ultrasonic guided wave detection method based on shape context dynamic time warping |
WO2023177664A1 (en) * | 2022-03-15 | 2023-09-21 | HawkEye 360, Inc. | Temporal rectification of radio frequency signals |
Family Cites Families (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4718093A (en) * | 1984-03-27 | 1988-01-05 | Exxon Research And Engineering Company | Speech recognition method including biased principal components |
US4982438A (en) * | 1987-06-02 | 1991-01-01 | Hitachi, Ltd. | Apparatus and method for recognizing three-dimensional shape of object |
JP2843079B2 (en) * | 1989-12-22 | 1999-01-06 | 本田技研工業株式会社 | Driving path determination method |
EP0435660B1 (en) * | 1989-12-29 | 1997-06-04 | Canon Kabushiki Kaisha | Method of evaluating objects based upon image processing, and inspection apparatus using said method |
US6014626A (en) * | 1994-09-13 | 2000-01-11 | Cohen; Kopel H. | Patient monitoring system including speech recognition capability |
JP2815045B2 (en) * | 1996-12-16 | 1998-10-27 | 日本電気株式会社 | Image feature extraction device, image feature analysis device, and image matching system |
US6369835B1 (en) * | 1999-05-18 | 2002-04-09 | Microsoft Corporation | Method and system for generating a movie file from a slide show presentation |
US6477515B1 (en) * | 1999-08-11 | 2002-11-05 | The United States Of America As Represented By The Secretary Of The Navy | Efficient computation of least cost paths with hard constraints |
US6834122B2 (en) * | 2000-01-22 | 2004-12-21 | Kairos Scientific, Inc. | Visualization and processing of multidimensional data using prefiltering and sorting criteria |
US6665852B2 (en) * | 2000-12-01 | 2003-12-16 | Sun Microsystems, Inc. | Piecewise linear cost propagation for path searching |
US20030001862A1 (en) * | 2001-06-29 | 2003-01-02 | Chu Michael H. | Method for the minimization of artifacts in full frame animations transferred to NTSC interlaced video |
US7225183B2 (en) * | 2002-01-28 | 2007-05-29 | Ipxl, Inc. | Ontology-based information management system and method |
TW569148B (en) * | 2002-04-09 | 2004-01-01 | Ind Tech Res Inst | Method for locating facial features in an image |
-
2002
- 2002-11-01 US US10/285,928 patent/US7103222B2/en not_active Expired - Fee Related
-
2003
- 2003-10-06 DE DE60307224T patent/DE60307224T2/en not_active Expired - Lifetime
- 2003-10-06 JP JP2004548015A patent/JP2006505038A/en active Pending
- 2003-10-06 EP EP03748726A patent/EP1444644B1/en not_active Expired - Fee Related
- 2003-10-06 WO PCT/JP2003/012783 patent/WO2004040503A1/en active IP Right Grant
- 2003-10-06 CN CNB2003801000687A patent/CN100365648C/en not_active Expired - Fee Related
Non-Patent Citations (4)
Title |
---|
COOPER M ET AL: "Scene boundary detection via video self-similarity analysis", PROCEEDINGS 2001 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (CAT. NO.01CH37205), PROCEEDINGS 2001 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, THESSALONIKI, GREECE, 7-10 OCT. 2001, 2001, Piscataway, NJ, USA, IEEE, USA, pages 378 - 381 vol.3, XP002265506, ISBN: 0-7803-6725-1 * |
FOOTE J.: "Visualizing music and audio using self-similarity", PROC. INT. ACM CONF. MULTIMEDIA, October 1999 (1999-10-01), Orlando, Florida, USA, pages 77 - 84, XP002265508 * |
HELFMAN J: "Dotplot patterns: a literal look at pattern languages", THEORY AND PRACTICE OF OBJECT SYSTEMS, 1996, WILEY, USA, vol. 2, no. 1, pages 31 - 41, XP002265507, ISSN: 1074-3227 * |
KEOGH E J ET AL: "Scaling up dynamic time warping for data mining applications", PROCEEDINGS. KDD-2000. SIXTH ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, PROCEEDINGS OF KDD-2000. SIXTH ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, BOSTON, MA, USA, 20-23 AUG. 2000, 2000, New York, NY, USA, ACM, USA, pages 285 - 289, XP002265509, ISBN: 1-58113-233-6 * |
Also Published As
Publication number | Publication date |
---|---|
EP1444644B1 (en) | 2006-08-02 |
DE60307224D1 (en) | 2006-09-14 |
CN1685359A (en) | 2005-10-19 |
CN100365648C (en) | 2008-01-30 |
JP2006505038A (en) | 2006-02-09 |
US7103222B2 (en) | 2006-09-05 |
EP1444644A1 (en) | 2004-08-11 |
US20040088722A1 (en) | 2004-05-06 |
DE60307224T2 (en) | 2007-06-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
EP1444644B1 (en) | A computerized method for discovering patterns in unknown multi-dimensional data | |
EP1456779B1 (en) | Method for summarizing unknown content of video | |
US7375731B2 (en) | Video mining using unsupervised clustering of video content | |
Bhatt et al. | Multimedia data mining: state of the art and challenges | |
Patel et al. | Content based video retrieval systems | |
US8311394B2 (en) | Video abstraction | |
Asha Paul et al. | Key-frame extraction techniques: A review | |
Ioannidis et al. | Weighted multi-view key-frame extraction | |
US7305133B2 (en) | Pattern discovery in video content using association rules on multiple sets of labels | |
Papadopoulos et al. | Automatic summarization and annotation of videos with lack of metadata information | |
Gao et al. | Key-frame selection for video summarization: An approach of multidimensional time series analysis | |
Kotsiantis et al. | Multimedia mining. | |
Bhute et al. | Multimedia indexing and retrieval techniques: A review | |
Mohanta et al. | A novel technique for size constrained video storyboard generation using statistical run test and spanning tree | |
Saravanan et al. | A review on content based video retrieval, classification and summarization | |
Wang et al. | Video indexing and retrieval based on key frame extraction | |
Patel et al. | Multimedia data mining: An overview | |
Damnjanovic et al. | Video summarisation for surveillance and news domain | |
Crucianu et al. | first name, surname Michel CRUCIANU | |
Haidar et al. | Detection algorithm of audiovisual production invariant | |
Sravya et al. | Object Recognition for Video Retrieval | |
Viitaniemi et al. | Concept-based video search with the PicSOM multimedia retrieval system | |
Marchand-Maillet et al. | Multimodal Fusion for Content-based Retrieval and Management of Multimedia Data | |
Yang et al. | Cross-Modal Correlation Mining using Graph Algorithms Jia-Yu Pan Carnegie Mellon University Pittsburgh, USA Phone: 412-268-1845 | |
Wang et al. | SHOT CLASSIFICATION FOR CONTENT-BASED INDEXING |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): CN JP |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HU IE IT LU MC NL PT RO SE SI SK TR |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2003748726 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2004548015 Country of ref document: JP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 20038A00687 Country of ref document: CN |
|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
WWP | Wipo information: published in national office |
Ref document number: 2003748726 Country of ref document: EP |
|
WWG | Wipo information: grant in national office |
Ref document number: 2003748726 Country of ref document: EP |