Image analysis – Applications – Target tracking or detecting
Reexamination Certificate
1999-01-05
2001-05-01
Johns, Andrew W. (Department: 2621)
Image analysis
Applications
Target tracking or detecting
Reexamination Certificate
active
06226388
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to object tracking in video devices, more particularly for an automated process and apparatus to track objects in video images.
2. Background of the Invention
Video devices that capture or display video images use object tracking for several purposes. In one example, a video device automatically focuses the picture on a specific object. In another example, the device automatically zooms in on that object. Other purposes include correction of brightness or contrast of the object or highlighting the object in a display.
Current methods typically use one of three types of algorithms: 1) correlation between adjacent frames; 2) geometrical object models; or, 3) illumination models. Correlation techniques typically use either a dot product or sum of squared differences (SSD) analysis to compare images. A portion of the initial frame is compared against several candidate regions in the next frame. Either technique produces a score for each candidate region of the next frame compared to the initial frame. The highest score indicates the location of the object in the next frame.
The second type of algorithm, geometric techniques, uses geometric modeling of objects in the image. A three-dimensional model of an object, such as a face, is created and projected onto a number of two-dimensional planes corresponding to every possible angle of orientation of the object. An input image is then compared against each possible two-dimensional projection until a match is found. This technique is computationally expensive and does not contemplate other objects that can partially occlude the tracked object, such as if it moved behind something in the next frame.
A third current method of tracking objects uses illumination modeling. All kinds of illumination and all possible condition changes are developed, then used in analyzing the next image. This method can be exhaustive, although techniques are available to reduce the load. Even with these techniques, the illumination approach has a disadvantage if the illumination changes in a way that is not predicted, making the tracking inaccurate. Also, problems occur if the object being projected alters its orientation, which will change the illumination model of the object, and partial occlusion remains a problem for this technique as well.
In summary, these current techniques have one or more disadvantages. These include sensitivity to object orientation and/or scale changes, sensitivity to partial occlusions and shadows, sensitivity to camera automatic exposure and lighting condition changes, or they are computationally intensive. Higher computation intensity results in slow response time, making real-time processing problematic.
Therefore, a method and apparatus are needed that track objects quickly with a reduced sensitivity to these problems.
SUMMARY OF THE INVENTION
One embodiment of the invention is a method and apparatus for tracking objects in video images. The method involves developing a color histogram of region in a first image, which includes the object, establishing a reference histogram. The next frame of the video input is received and candidate regions are identified. Color histograms for the candidate regions are compared to the reference histogram and used to locate and identify the object being tracked. Motion analysis provides correction of object location for the current frame and provides a prediction for the next frame. The predication can be used to more accurately identify candidate regions.
The method can be improved or its performance altered with addition of option modules, such as one that smoothes the tracking trajectory.
Alternative embodiments of the invention include an apparatus for tracking objects in a video capture system, or a video display system.
REFERENCES:
patent: 5031049 (1991-07-01), Toyama et al.
patent: 5187585 (1993-02-01), Kaneda et al.
patent: 5412487 (1995-05-01), Nishimura et al.
patent: 5422828 (1995-06-01), Choate et al.
patent: 5434617 (1995-07-01), Bianchi
patent: 5473369 (1995-12-01), Abe
patent: 5546125 (1996-08-01), Tomitaka et al.
patent: 5552823 (1996-09-01), Kageyama
patent: 5610653 (1997-03-01), Abecassis
patent: 5631697 (1997-05-01), Nishimura et al.
patent: 5696503 (1997-12-01), Nasberg
patent: 5714999 (1998-02-01), Jeong et al.
patent: 5745126 (1998-04-01), Jain et al.
patent: 5798787 (1998-08-01), Yamaguchi et al.
patent: 5912994 (1999-06-01), Norton et al.
patent: 5982909 (1999-11-01), Erdem et al.
Lucas, et al. An iterative Image Registration Technique with an Application to Stereo Vision,Proc. Int. Joint Conf. Artificial Intelligence, pp. 674-679, 1981.
Papanikolopoulos, et al. Visual Tracking of a Moving Target by a Camera Mounted on a Robot: A Combination of Control and Vision,IEEE Transactions of Robotics and Automation, vol. 9, No 1, Feb. 1993, pp. 14-35.
Rehg, et al. Visual Tracking with Deformation Models,Proceedings of IEEE Int'l Conference on Robotics and Automation, vol. 9, No. 1, 1993, pp. 844-850.
Hager, et al. Real-Time Tracking of Image Regions with Changes in Geometry and Illumination,Proceedings of IEEE Conference on Computer Vision and Pattern Recognition, pp. 403-410.
Black, et al. Eigen Tracking: Robust Matching and Tracking of Articulated Objects Using a View-Based Representation,Proceedings of European Conference on Computer Vision, 1996, pp.329-341.
Swain, et al. Color Indexing,International Journal of Computer Vision, vol. 7, No. 1, 1991, pp. 11-32.
Qian Richard J.
Sampsell Jeffrey B.
Johns Andrew W.
Marger Johnson & McCollom
Nakhjavan Shervin
Sharp Labs of America, Inc.
LandOfFree
Method and apparatus for object tracking for automatic... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for object tracking for automatic..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for object tracking for automatic... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2516155