System and method for visually tracking occluded objects in real time
First Claim
1. A method for tracking an object having plural parts, wherein the object is tracked as an articulated structure, the method comprising:
- obtaining digital representations of the object from multiple viewpoints;
generating disparity maps between the digital representations; and
digitally segmenting the disparity maps and modeling selected parts of the object using a predefined 3-D Gaussian statistical model of the object with fixed eigenvalues.
2 Assignments
0 Petitions
Accused Products
Abstract
The present invention is embodied in a system and method for digitally tracking objects in real time. The present invention visually tracks three-dimensional (3-D) objects in dense disparity maps in real time. Tracking of the human body is achieved by digitally segmenting and modeling different body parts using statistical models defined by multiple size parameters, position and orientation. In addition, the present invention is embodied in a system and method for recognizing mutual occlusions of body parts and filling in data for the occluded parts while tracking a human body. The body parts are preferably tracked from frame to frame in image sequences as an articulated structure in which the body parts are connected at the joints instead of as individual objects moving and changing shape and orientation freely.
662 Citations
55 Claims
-
1. A method for tracking an object having plural parts, wherein the object is tracked as an articulated structure, the method comprising:
-
obtaining digital representations of the object from multiple viewpoints;
generating disparity maps between the digital representations; and
digitally segmenting the disparity maps and modeling selected parts of the object using a predefined 3-D Gaussian statistical model of the object with fixed eigenvalues. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12)
-
-
13. A method for tracking a human body having multiple body parts comprising:
-
obtaining a digital representation of the human body;
predefining generative statistical models of image formation for each body part and using the statistical models to recognize mutual body part occlusions;
filling in data for the occluded body parts with the statistically recognized mutual occlusions. - View Dependent Claims (14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27)
-
-
28. An object tracking system for dynamically tracking articulated structures within a sequence of images, wherein the structures are made of segmented parts, comprising:
-
a predefined statistical model of each articulated structure;
a depth map processor that computes range image data and segments a disparity map into different parts of the structure;
a pixel classification processor that classifies pixels of each image within defined part segments; and
an occlusion detector and handler that estimates missing data and resolves the missing data as occluded parts. - View Dependent Claims (29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55)
-
Specification