International Business Machines Corporation

Daniel Andre Vaquero

Rogerio Schmidt Feris

Arun Hampapur

Lisa Marie Brown

Sharathchandra UmapathiRao Pankanti

Ankur Datta

Lisa Maire Brown

Techniques for facilitating a video surveillance search of a person are provided. The techniques include maintaining a database of one or more attributes of one or more people captured on one or more video cameras, indexing the one or more attributes in the database extracted from the one or more video cameras, and pruning one or more images captured from the one or more video cameras using the one or more attributes and one or more items of qualifying information to facilitate a video surveillance search of a person.

Facilitating people search in video surveillance

Methods, systems, and computer program products for parsing objects are provided herein. A method includes producing a plurality of versions of an image of an object derived from an input, wherein each version comprises one of one multiple resolutions of said image of said object; computing an appearance probability at each of a plurality of regions on the one or more lowest resolution versions of said plurality of versions of said image for at least one attribute for said object; determining a configuration of the at least one attribute in the one or more lowest resolution versions based on at least the appearance probability in each of the plurality of regions; and outputting said configuration.

Semantic parsing of objects in video

In response to a query of discernable facial attributes, the locations of distinct and different facial regions are estimated from face image data, each relevant to different attributes. Different features are extracted from the estimated facial regions from database facial images, which are ranked in base layer rankings as a function of relevance of extracted features to attributes relevant to the estimated regions, and in second-layer rankings as a function of combinations of the base layer rankings and relevance of the extracted features to common ones of the attributes relevant to the estimated regions. The images are ranked in relevance to the query as a function of the second-layer rankings.

Hierarchical ranking of facial attributes

Techniques, systems, and articles of manufacture for multispectral detection of attributes for video surveillance. A method includes generating one or more training sets of one or more multispectral images, generating a group of one or more multispectral box features, using the one or more training sets to select one or more of the one or more multispectral box features to generate a multispectral attribute detector, and using the multispectral attribute detector to identify a location of an attribute in video surveillance, wherein using the multispectral attribute detector comprises, for one or more locations on each spectral band level of the multispectral image, applying the multispectral attribute detector and producing an output indicating attribute detection or an output indicating no attribute detection, and wherein the attribute corresponds to the multispectral attribute detector.

Multispectral detection of personal attributes for video surveillance

Methods, systems, and computer program products for parsing objects in a video are provided herein. A method includes producing a plurality of versions of an image of an object derived from a video input, wherein each version has a different resolution of said image of said object; computing an appearance score at each of a plurality of regions on the lowest resolution version of said plurality of versions of said image for at least one attribute for said object, wherein said appearance score denotes a probability of the at least one attribute appearing in the region; determining a configuration of the at least one attribute in the lowest resolution version based on at least the appearance score in each of the plurality of regions in the lowest resolution version; and displaying said configuration.

Techniques for tracking an individual across two or more cameras are provided. The techniques include detecting an image of one or more individuals in each of two or more cameras, tracking each of the one or more individuals in a field of view in each of the two or more cameras, applying a set of one or more attribute detectors to each of the one or more individuals being tracked by the two or more cameras, and using the set of one or more attribute detectors to match an individual tracked in one of the two or more cameras with an individual tracked in one or more other cameras of the two or more cameras.

Attribute-based person tracking across multiple cameras

Methods, systems, and computer program products for parsing objects in a video are provided herein. A method includes producing a plurality of versions of an image of an object, wherein each version has a different resolution of said image of said object, and computing an appearance score at each of a plurality of regions on the lowest resolution version for at least one attribute for said object. Such a method also includes analyzing one or more other versions to compute a resolution context score for each of the plurality of regions in the lowest resolution version, and determining a configuration of the at least one semantic attribute in the lowest resolution version based on the appearance score and the resolution context score.

Methods, systems, and computer program products for parsing objects in a video are provided herein. A method includes producing a plurality of versions of an image of an object, wherein each version has a different resolution of said image of said object, and computing an appearance score at each of a plurality of regions on the lowest resolution version for at least one semantic attribute for said object. Such a method also includes analyzing one or more other versions to compute a resolution context score for each of the plurality of regions in the lowest resolution version, wherein said resolution context score denotes an extent to which finer spatial structure exists in the one or more others versions than in the lowest resolution version, and determining a configuration of the at least one semantic attribute in the lowest resolution version based on the appearance score and the resolution context score.

In response to a query of discernable facial attributes, the locations of distinct and different facial regions are estimated from face image data, each relevant to different attributes. Different features are extracted from the estimated facial regions from database facial images, which are ranked in base layer rankings by matching feature vectors to a base layer ranking sequence as a function of edge weights. Second-layer rankings define second-layer attribute vectors as combinations of the base-layer feature vectors and associated base layer parameter vectors for common attributes, which are matched to a second-layer ranking sequence as a function of edge weights. The images are thus ranked for relevance to the query as a function of the second-layer rankings.

Techniques, systems, and computer program products for parsing objects in a video are provided herein. A method includes producing and storing a plurality of versions of an image of an object derived from a video input, wherein each version of said image has a different resolution of said image; computing an appearance score at each of a plurality of regions on the lowest resolution version of said image for a plurality of semantic attributes with associated parts for said object, said appearance score denoting a probability of each semantic attribute appearing in the region; analyzing increasingly higher resolution versions than the lowest resolution version to compute a resolution context score for each region in the lowest resolution version; and ascertaining an optimized configuration of body parts and associated semantic attributes in the lowest resolution version, said ascertaining utilizing the appearance scores and the resolution context scores.

Techniques for detecting an attribute in video surveillance include generating training sets of multispectral images, generating a group of multispectral box features comprising receiving input of a detector size of a width and height, a number of spectral bands in the multispectral images, and integer values representing a minimum and maximum width and height of multispectral box features, fixing a feature width and height, generating feature building blocks with the fixed width and height, placing a feature building block at a same location for each spectral band level, and enumerating combinations of the feature building blocks through each spectral level until all sizes within the integer values have been covered, and wherein each combination determines a multispectral box feature, using the training sets to select multispectral box features to generate a multispectral attribute detector, and using the multispectral attribute detector to identify a location of an attribute in video surveillance.

In response to a query of discernible facial attributes, the locations of distinct and different facial regions are estimated from face image data, each relevant to different attributes. Different features are extracted from the estimated facial regions from database facial images, which are ranked in base layer rankings by matching feature vectors in a bipartite graph to a base layer ranking sequence as a function of edge weights parameterized by an associated base layer parameter vector. Second-layer rankings define second-layer attribute vectors as bilinear combinations of the base-layer feature vectors and associated base layer parameter vectors for common attributes, which are matched in a bipartite graph to a second-layer ranking sequence as a function of edge weights parameterized by associated second-layer parameter vectors. The images are thus ranked for relevance to the query as a function of fusing the second-layer rankings.

Growing community of inventors

Santa Barbara, CA, United States of America

Daniel Andre Vaquero