1008 results about "Feature descriptor" patented technology

A method for classifying or comparing objects includes detecting points of interest within two objects, computing feature descriptors at said points of interest, forming a multi-resolution histogram over feature descriptors for each object and computing a weighted intersection of multi-resolution histogram for each object. An alternative embodiment includes a method for matching objects by defining a plurality of bins for multi-resolution histograms having various levels and a plurality of cluster groups, each group having a center, for each point of interest, calculating a bin index, a bin count and a maximal distance to the bin center and providing a path vector indicative of the bins chosen at each level. Still another embodiment includes a method for matching objects comprising creating a set of feature vectors for each object of interest, mapping each set of feature vectors to a single high-dimensional vector to create an embedding vector and encoding each embedding vector with a binary hash string.

A system and method for performing object and class recognition that allows for wide changes of viewpoint and distance of objects is disclosed. The invention provides for choosing pose-invariant interest points of a three-dimensional (3D) image, and for computing pose-invariant feature descriptors of the image. The system and method also allows for the construction of three-dimensional (3D) object and class models from the pose-invariant interest points and feature descriptors of previously obtained scenes. Interest points and feature descriptors of a newly acquired scene may be compared to the object and/or class models to identify the presence of an object or member of the class in the new scene.

The invention discloses a human behavior recognition method integrating the space-time dual-network flow and an attention mechanism. The method includes the steps of extracting moving optical flow features and generating an optical flow feature image; constructing independent time flow and spatial flow networks to generate two segments of high-level semantic feature sequences with a significant structural property; decoding the high-level semantic feature sequence of the time flow, outputting a time flow visual feature descriptor, outputting an attention saliency feature sequence, and meanwhile outputting a spatial flow visual feature descriptor and the label probability distribution of each frame of a video window; calculating an attention confidence scoring coefficient per frame time dimension, weighting the label probability distribution of each frame of the video window of the spatial flow, and selecting a key frame of the video window; and using a softmax classifier decision to recognize the human behavior action category of the video window. Compared with the prior art, the method of the invention can effectively focus on the key frame of the appearance image in the originalvideo, and at the same time, can select and obtain the spatial saliency region features of the key frame with high recognition accuracy.

A system and process for identifying corresponding points among multiple images of a scene is presented. This involves a multi-view matching framework based on a new class of invariant features. Features are located at Harris corners in scale-space and oriented using a blurred local gradient. This defines a similarity invariant frame in which to sample a feature descriptor. The descriptor actually formed is a bias/gain normalized patch of intensity values. Matching is achieved using a fast nearest neighbor procedure that uses indexing on low frequency Haar wavelet coefficients. A simple 6 parameter model for patch matching is employed, and the noise statistics are analyzed for correct and incorrect matches. This leads to a simple match verification procedure based on a per feature outlier distance.

An efficient large scale search system for video and multi-media content using a distributed database and search, and tiered search servers is described. Selected content is stored at the distributed local database and tier1 search server(s). Content matching frequent queries, and frequent unidentified queries are cached at various levels in the search system. Content is classified using feature descriptors and geographical aspects, at feature level and in time segments. Queries not identified at clients and tier1 search server(s) are queried against tier2 or lower search server(s). Search servers use classification and geographical partitioning to reduce search cost. Methods for content tracking and local content searching are executed on clients. The client performs local search, monitoring and/or tracking of the query content with the reference content and local search with a database of reference fingerprints. This shifts the content search workload from central servers to the distributed monitoring clients.

In an image retrieving and delivering system and an image retrieving and delivering method, a feature descriptor is retrieved from a data base, in which each of a plurality of images including a moving picture and a static picture is registered with a feature descriptor describing the feature of the image, according to a retrieval condition input by a user, a retrieval result satisfying the retrieval condition is obtained, and the retrieval result is edited and processed according to a delivery condition obtained from a user terminal in which the retrieval result is to be received.

Simultaneous localization and mapping (SLAM) utilizes multiple view feature descriptors to robustly determine location despite appearance changes that would stifle conventional systems. A SLAM algorithm generates a feature descriptor for a scene from different perspectives using kernel principal component analysis (KPCA). When the SLAM module subsequently receives a recognition image after a wide baseline change, it can refer to correspondences from the feature descriptor to continue map building and/or determine location. Appearance variations can result from, for example, a change in illumination, partial occlusion, a change in scale, a change in orientation, change in distance, warping, and the like. After an appearance variation, a structure-from-motion module uses feature descriptors to reorient itself and continue map building using an extended Kalman Filter. Through the use of a database of comprehensive feature descriptors, the SLAM module is also able to refine a position estimation despite appearance variations.

Various methods for tracking and recognition with rotation invariant feature descriptors are provided. One example method includes generating an image pyramid of an image frame, detecting a plurality of interest points within the image pyramid, and extracting feature descriptors for each respective interest point. According to some example embodiments, the feature descriptors are rotation invariant. Further, the example method may also include tracking movement by matching the feature descriptors to feature descriptors of a previous frame and performing recognition of an object within the image frame based on the feature descriptors. Related example methods and example apparatuses are also provided.

A detection method is based on a statistical analysis of the appearance of model patches from all possible viewpoints in the scene, and incorporates 3D geometry during both matching and pose estimation processes. By analyzing the computed probability distribution of the visibility of each patch from different viewpoints, a reliability measure for each patch is estimated. That reliability measure is useful for developing industrial augmented reality applications. Using the method, the pose of complex objects can be estimated efficiently given a single test image. In a fast method to detect objects in a given image or video sequence, a series of hierarchical feature descriptors permit balancing between the complexity of feature extraction and cost of combinatorial matching. The feature descriptors are derived from the 3D model of the object of interest along with the available calibrated real-life images or videos. The variability of the real-life images defines the granularity of the feature descriptors.

A method of providing a set of feature descriptors configured to be used in matching an object in an image of a camera is provided. The method includes: a) providing at least two images of a first object; b) extracting in at least two of the images at least one feature from the respective image, c) providing at least one descriptor for an extracted feature, and storing the descriptors; d) matching descriptors in the first set of descriptors; e) computing a score parameter based on the result of the matching process; f) selecting at least one descriptor based on its score parameter; g) adding the selected descriptor(s) to a second set of descriptors; and h) updating the score parameter of descriptors in the first set based on a selection process and to the result of the matching process.

Systems and methods for describing image content establish image description records which include an object set (24), an object hierarchy (26) and entity relation graphs (28). For image content, image objects can include global objects (O0 8) and local objects (O1 2 and O2 6). The image objects are further defined by a number of features of different classes (36, 38 and 40), which in turn are further defined by a number of feature descriptors. The relationships between and among the objects in the object set are defined by the object hierarchy (26) and entity relation graphs (28). The image description records provide a standard vehicle for describing the content and context of image information for subsequent access and processing by computer applications such as search engines, filters, and archive systems.

Face verification is performed using video data. The two main modules are face image capturing and face verification. In face image capturing, good frontal face images are captured from input video data. A frontal face quality score discriminates between frontal and profile faces. In face verification, a local binary pattern histogram is selected as the facial feature descriptor for its high discriminative power and computational efficiency. Chi-Square (χ²) distance between LBP histograms from two face images are then calculated as a face dissimilarity measure. The decision whether or not two images belong to the same person is then made by comparing the corresponding distance with a pre-defined threshold. Given the fact that more than one face images can be captured per person from video data, several feature based and decision based aggregators are applied to combine pair-wise distances to further improve the verification performance.

Method and system for finding targets within visual data, comprising receiving target object information. Generating a set of target object semantic attributes from the target object information. Identifying a plurality of candidate objects within visual data. Generating a set of low-level feature descriptors from the visual data for each candidate object. Generating from the set of low-level feature descriptors a set of candidate semantic attributes for each candidate object within the visual data. Identifying one or more portions of the visual data containing a candidate object, from the plurality of candidate objects, having a set of candidate object semantic attributes that match the set of target object semantic attributes. Providing an output indicating the identified one or more portions of the visual data.

A method, apparatus and computer program product are provided for extracting spatio-temporal features with the aid of sensor information. An exemplary method comprises receiving video data and auxiliary sensor data and associating the two with timestamp information. The method may also include segmenting an input data stream into stable segments and extracting temporal features from the associated video data. The method may further include extracting temporal features either form the whole video or only from the video data where little or no stable segments are detected and performing camera view motion compensation by using information provided by the auxiliary sensors for modifying the feature-descriptors.

The invention relates to a multi-scale normal feature point cloud registering method. The multi-scale normal feature point cloud registering method is characterized by including the steps that two-visual-angle point clouds, including the target point clouds and the source point clouds, collected by a point cloud obtaining device are read in; the curvature of radius neighborhoods of three scales of points is calculated, and key points are extracted from the target point clouds and the source point clouds according to a target function; the normal vector angular deviation and the curvature of the key points in the radius neighborhoods of the different scales are calculated and serve as feature components, feature descriptors of the key points are formed, and a target point cloud key point feature vector set and a source point cloud key point feature vector set are accordingly obtained; according to the similarity level of the feature descriptors of the key points, the corresponding relations between the target point cloud key points and the source point cloud key points are preliminarily determined; the wrong corresponding relations are eliminated, and the accurate corresponding relations are obtained; the obtained accurate corresponding relations are simplified with the clustering method, and the evenly-distributed corresponding relations are obtained; singular value decomposition is carried out on the final corresponding relations to obtain a rigid body transformation matrix.

A method for detecting a vehicle running a stop signal includes acquiring at least two evidentiary images of a candidate violating vehicle captured from at least one camera monitoring an intersection. The method includes extracting feature points in each of the at least two evidentiary images. The method includes computing feature descriptors for each of the extracted feature points. The method includes determining a correspondence between feature points having matching feature descriptors at different locations in the at least two evidentiary images. The method includes extracting at least one attribute for each correspondence. The method includes determining if the candidate violating vehicle is in violation of running the stop signal using the extracted attribute.