SYSTEM AND METHOD FOR USING APPARENT SIZE AND ORIENTATION OF AN OBJECT TO IMPROVE VIDEO-BASED TRACKING IN REGULARIZED ENVIRONMENTS

• US 20150131851A1
• Filed: 11/13/2013
• Published: 05/14/2015
• Est. Priority Date: 11/13/2013
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Bernal, Edgar A., Mizes, Howard A., Loce, Robert P.
• Granted Patent: US 9,405,974 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/20036 Morphological image processing
• CPC Class Codes: G06T 2207/20076 Probabilistic image processing
• CPC Class Codes: G06T 2207/30232 Surveillance
• CPC Class Codes: G06T 2207/30236 Traffic on road, railway or crossing
• CPC Class Codes: G06T 7/194 involving foreground-background segmentation
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06T 7/248 involving reference images or patches
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06T 7/269 using gradient-based methods
• CPC Class Codes: G06T 7/73 using feature-based methods
• CPC Class Codes: G06V 10/147 Details of sensors, e.g. sensor lenses fingerprint or palmprint sensors G06V40/13; vascular sensors G06V40/145; eye sensors G06V40/19
• CPC Class Codes: G06V 20/40 in video content extracting overlay text G06V20/62; video retrieval G06F16/70; processing of video elementary streams in video servers H04N21/234; processing of video elementary streams in video clients H04N21/44

Abstract

A system and method for optimizing video-based tracking of an object of interest are provided. A video of a regularized motion environment that comprise multiple video frames is acquired and an initial instance of an object of interest in one of the frames is then detected. An expected size and orientation of the object of interest as a function of the location of the object is then determined. The location of the object of interest is then determined in a next subsequent frame using the expected size and orientation of the object of interest.

Claims

1. A method for optimizing video-based tracking of an object of interest, comprising:

acquiring a video of a regularized motion environment comprising a plurality of video frames;

detecting an initial instance of at least one object of interest in the plurality of video frames including a location thereof;

determining an expected size and an expected orientation of the at least one object of interest as a function of the location; and

localizing the at least one object of interest in at least one subsequent video frame responsive to the determined size and orientation,

wherein at least one of the acquiring, detecting, generating, determining, and localizing is performed by a computer processor.

2. The method according to claim 1, further comprising:

generating a binary mask of the at least one object of interest in a current video frame; and

generating a target representation of the at least one object of interest in accordance with the binary mask and the current video frame.

3. The method according to claim 2, wherein the expected size and the expected orientation of the at least one object of interest is determined in accordance with a geometry of a video camera acquiring the video of the regularized motion environment.

4. The method according to claim 2, further comprising generating a plurality of structuring elements in accordance with the expected size and orientation of the at least one object of interest as a function of the location, wherein morphological operations are performed on the binary mask in accordance with the plurality of structuring elements.

5. The method according to claim 1, wherein the expected size and orientation of the at least one object of interest as a function of the location are determined in accordance with historical information corresponding to size and orientation, wherein the historical information is associated with the regularized motion environment.

6. The method according to claim 1, wherein the expected size and orientation of the at least one object of interest as a function of the location are determined in accordance a remote source of information corresponding to size and orientation, wherein the remote source is associated with the regularized motion environment.

7. The method according to claim 1, wherein localizing the at least one object uses at least one of a template matching tracking, point feature tracking, a mean shift tracking, or a particle filter tracking.

8. A system for optimizing video-based tracking of an object of interest, comprising:

a video acquisition unit configured for acquiring a video of a regularized motion environment in memory, the video comprising a plurality of frames;

an object detection unit configured for detecting an initial instance of an object of interest a frame of the plurality of video frames;

an object characterization unit configured for establishing a target object representation of the detected instance of the object of interest;

an object localization unit configured for determining a location of the object of interest in the frame in accordance with the target representation of the detected instance of the object of interest;

an object size and orientation unit configured for estimating a size and an orientation of the object of interest in a next subsequent frame as a function of the determined location; and

a processor which implements at least one of the video acquisition unit, the object detection unit, the object characterization unit, the object localization unit, and the object size and orientation unit.

9. The system according to claim 8, wherein the object size and orientation unit is further configured for generating a plurality of structuring elements in accordance with the estimated size and orientation of the object of interest as a function of location in the next subsequent frame.

10. The system according to claim 9, wherein the object size and orientation unit is further configured to estimate the size and orientation of the at least one object of interest as a function of the location in accordance with historical information corresponding to size and orientation, wherein the historical information is associated with the regularized motion environment.

11. The system according to claim 9, wherein the object size and orientation unit is further configured to estimate the size and orientation of the at least one object of interest as a function of the location are determined in accordance a remote source of information corresponding to size and orientation, wherein the remote source is associated with the regularized motion environment;

and wherein the object detection unit is further configured for modifying a binary mask of the object of interest in accordance with the plurality of generated structuring elements.

12. The system according to claim 11, wherein the object characterization unit establishes the target object representation of the detected instance of the object of interest in accordance with the generated binary mask and the current video frame.

13. The system according to claim 12, wherein the object localization unit is further configured to determine the location in accordance with at least one of a template matching tracking, a point feature tracking, a mean shift tracking, or a particle filter tracking.

14. A computer-implemented method for optimizing video-based tracking of an object of interest, comprising:

generating a binary mask of a detected instance of an object of interest in one of a plurality of video frames;

establishing a target object representation of the detected instance of the object of interest in accordance with the generated binary mask;

determining a location of the object of interest in the frame in accordance with the target representation of the detected instance of the object of interest;

estimating a size and an orientation of the object of interest as a function of the location; and

localizing the object of interest in a next subsequent frame responsive to the estimated size and orientation.

15. The computer-implemented method according to claim 14, wherein the estimated size and orientation of the at least one object of interest as a function of the location are determined in accordance with historical information or a remote source of information corresponding to size and orientation, wherein the historical information or the remote source of information is associated with the regularized motion environment.

16. The computer-implemented method according to claim 15, further comprising generating a plurality of structuring elements in accordance with the estimated size and orientation of the object of interest as a function of location in the next subsequent frame, wherein the binary mask is modified in accordance with the plurality of structuring elements, and wherein morphological operations are performed on the binary mask in accordance with the plurality of structuring elements.

17. The computer-implemented method according to claim 16, wherein the estimated size and orientation of the object of interest are estimated in accordance with a geometry of a video camera acquiring the video of the regularized motion environment.

18. The computer-implemented method according to claim 17, further comprising:

acquiring a video of a regularized motion environment comprising a plurality of video frames; and

detecting the instance of the object of interest in the one of the plurality of video frames.

19. A method for optimizing video-based tracking of an object of interest, comprising:

acquiring a video of a regularized motion environment comprising a plurality of video frames;

detecting an initial instance of at least one object of interest in an initial video frame of the plurality of video frames including detection of a location thereof;

localizing the at least one object of interest in a plurality of subsequent video frames;

determining an object trajectory of the at least one object of interest localized in the plurality of subsequent video frames;

determining an expected size and an expected orientation of the at least one object of interest as a function of the determined trajectory; and

localizing the at least one object of interest in at least one of the plurality of subsequent video frames based on the determined expected size and expected orientation;

wherein at least one of the acquiring, detecting, localizing, determining, determining, and localizing is performed by a computer processor.

20. The method of claim 20 wherein the determining an object trajectory is responsive to a determined size and orientation of the object of interest within the plurality of subsequent video frames.