Video segmentation using statistical pixel modeling

Abstract

A method for segmenting video data into foreground and background portions utilizes statistical modeling of the pixels. A statistical model of the background is built for each pixel, and each pixel in an incoming video frame is compared with the background statistical model for that pixel. Pixels are determined to be foreground or background based on the comparisons. The method for segmenting video data may be further incorporated into a method for implementing an intelligent video surveillance system.

1 . A method of implementing an intelligent video surveillance system, comprising: 
 segmenting video into foreground and background components, the segmenting comprising: 
 obtaining a sequence of video frames;  
 building and updating at least one background statistical model for each region of the video frames, based on the video frames; and  
 assigning labels to the regions, based on the at least one background statistical model;  
   identifying and classifying objects based on the labeled regions; and    analyzing behaviors of at least one object.    
   
   
       2 . A computer-readable medium comprising software implementing the method of  claim 1 .  
   
   
       3 . An intelligent video surveillance system comprising a computer system comprising: 
 a computer; and    a computer-readable medium according to  claim 2 .    
   
   
       4 . The method of  claim 1 , wherein said analyzing behaviors of at least one of the objects comprises: 
 tracking at least one of the objects.    
   
   
       5 . The method of  claim 1 , further comprising: 
 creating at least one rule to detect at least one specific activity;    wherein said analyzing behaviors of at least one of the objects includes applying the at least one rule.    
   
   
       6 . The method of  claim 5 , wherein said at least one rule includes at least one virtual tripwire and determining when the at least one virtual tripwire is crossed.  
   
   
       7 . The method of  claim 5 , wherein said at least one rule includes a definition of at least one area and the determining at least one of when an object enters, when an object leaves, and when an object loiters in the at least one area.  
   
   
       8 . The method of  claim 5 , wherein said at least one rule includes at least one of determining when an object is added to a scene and determining when an object is removed from a scene.  
   
   
       9 . A method of implementing an automated closed-circuit television (CCTV) surveillance system, comprising: 
 providing CCTV equipment generating an input video stream; and    implementing the method of  claim 1 .    
   
   
       10 . A method of implementing an automated security system, comprising the method of  claim 1 .  
   
   
       11 . A method of implementing an automated anti-terrorism system, comprising the method of  claim 1 .  
   
   
       12 . A method of implementing an automated market research system, comprising the method of  claim 1 .  
   
   
       13 . The method of  claim 12 , wherein said analyzing behaviors of at least one of the objects comprises: 
 tracking behaviors of at least one of the objects in at least one retail location.    
   
   
       14 . A method of implementing an automated traffic monitoring system, comprising the method of  claim 1 .  
   
   
       15 . The method of  claim 14 , wherein said analyzing behaviors of at least one of the objects comprises at least one of: 
 detecting wrong-way traffic;    detecting a broken-down vehicle;    detecting an accident; and    detecting a road blockage.    
   
   
       16 . An apparatus for intelligent video surveillance adapted to perform the method comprising: 
 segmenting video into foreground and background components, the segmenting comprising: 
 obtaining a sequence of video frames;  
 building and updating at least one background statistical model for each region of the video frames, based on the video frames; and  
 assigning labels to the regions, based on the at least one background statistical model;  
   identifying and classifying objects based on the labeled regions; and    analyzing behaviors of at least one object.    
   
   
       17 . The apparatus of  claim 16 , wherein the apparatus comprises application-specific hardware to emulate a computer and/or software adapted to perform said segmenting, said obtaining, said building, said assigning, said identifying, and said analyzing.  
   
   
       18 . A method of implementing an intelligent video surveillance system, comprising: 
 building and updating at least one background statistical model for an input video sequence;    labeling regions of frames of the input video sequence based on the at least one background statistical model;    identifying and classifying at least one object in the input video sequence based on the labeled regions; and    analyzing behavior of said at least one object.    
   
   
       19 . A computer-readable medium comprising software implementing the method of  claim 18 .  
   
   
       20 . An apparatus for intelligent video surveillance adapted to perform the method comprising: 
 building and updating at least one background statistical model for an input video sequence;    labeling regions of frames of the input video sequence based on the at least one background statistical model;    identifying and classifying at least one object in the input video sequence based on the labeled regions; and    analyzing behavior of said at least one object.    
   
   
       21 . The apparatus of  claim 20  wherein the apparatus comprises application-specific hardware to emulate a computer and/or software adapted to perform said segmenting, said obtaining, said building, said assigning, said identifying, and said analyzing.