Integrated systems and methods for video-based object modeling, recognition, and tracking
First Claim
1. A method for tracking object images in video files, comprising the steps of:
- receiving a video file, wherein the video file comprises a plurality of frames;
extracting an image of an object from a particular frame in the video file;
extracting a subsequent image of an object from a subsequent frame in the video file;
calculating a similarity value between the extracted object image from the particular frame and the subsequent extracted object image from the subsequent frame;
if the calculated similarity value exceeds a predetermined similarity threshold, collecting the extracted object images from the video file into an object group;
extracting a plurality of object features from the extracted object images collected in the object group;
generating an object model based on the plurality of extracted object features, where the object model is associated with an object identifier; and
updating the object model with additional object features extracted from one or more additional object images collected in the object group but extracted from an additional video file.
10 Assignments
0 Petitions
Accused Products
Abstract
The present disclosure relates to systems and methods for modeling, recognizing, and tracking object images in video files. In one embodiment, a video file, which includes a plurality of frames, is received. An image of an object is extracted from a particular frame in the video file, and a subsequent image is also extracted from a subsequent frame. A similarity value is then calculated between the extracted images from the particular frame and subsequent frame. If the calculated similarity value exceeds a predetermined similarity threshold, the extracted object images are assigned to an object group. The object group is used to generate an object model associated with images in the group, wherein the model is comprised of image features extracted from optimal object images in the object group. Optimal images from the group are also used for comparison to other object models for purposes of identifying images.
-
Citations
31 Claims
-
1. A method for tracking object images in video files, comprising the steps of:
-
receiving a video file, wherein the video file comprises a plurality of frames; extracting an image of an object from a particular frame in the video file; extracting a subsequent image of an object from a subsequent frame in the video file; calculating a similarity value between the extracted object image from the particular frame and the subsequent extracted object image from the subsequent frame; if the calculated similarity value exceeds a predetermined similarity threshold, collecting the extracted object images from the video file into an object group; extracting a plurality of object features from the extracted object images collected in the object group; generating an object model based on the plurality of extracted object features, where the object model is associated with an object identifier; and updating the object model with additional object features extracted from one or more additional object images collected in the object group but extracted from an additional video file. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17)
-
-
18. A method of tracking a specific object through a plurality of frames of a video wherein each frame of the video includes a plurality of objects, comprising the steps of:
-
dividing the video into one or more scenes; identifying, an image of the specific object in a first frame of one of the one or more scenes; assigning the image of the specific object from the first frame of the respective scene to an object group; analyzing a subsequent frame of the respective scene to identify a potential additional image of the specific object; calculating a similarity value between the image of the specific object from the first frame and the potential additional image of the specific object from the subsequent frame; if the calculated similarity value exceeds a predetermined similarity threshold, assigning the potential additional image to the object group; and repeating the steps of analyzing and calculating for additional subsequent frames of the respective scene whereby the object group has assigned thereto all images of the specific object in the analyzed frames of the video. - View Dependent Claims (19, 20, 21, 22, 23)
-
-
24. A method for generating an object model from one or more images in a video, comprising the steps of:
-
receiving a video file, wherein the video file comprises a plurality of video frames; detecting an object image from one of the plurality of video frames; if the detected object image is an optimal object image, extracting a plurality of object features from the optimal object image; determining whether an object model already exists for the object represented by the optimal object image; if an object model does not exist for the object, generating an object model comprising the plurality of object features extracted from the optimal object image; and updating the object model with additional object features extracted from one or more additional optimal object images associated with the object extracted from one or more additional video files. - View Dependent Claims (25, 26, 27, 28)
-
-
29. A system for identifying object images from videos, comprising:
-
a processor; a computer program product that includes a computer-readable medium that is usable by the processor, the medium having stored thereon a sequence of instructions that when executed by the processor causes the execution of the steps of; retrieving a video, wherein the video comprises a plurality of frames; extracting a plurality of object images from the plurality of frames; grouping similar object images together into a group based on a similarity score calculated between features in the plurality of object images; calculating an average similarity score for the plurality of object images in the group as compared to one or more stored object models, where each object model is associated with a respective object identifier; if the calculated average similarity score does not exceed a predetermined threshold, labeling the plurality of object images in the group as unknown; and if the calculated average similarity score exceeds the predefined threshold, identifying the plurality of object images in the group based on the respective object identifier of the most similar stored object model. - View Dependent Claims (30, 31)
-
Specification