Content-based video compression

Abstract

A video compression method and system including object-oriented compression plus error correction using decoder feedback.

1 . (canceled)  
   
   
       2 . An encoder for motion-compensated video, comprising: 
 (a) first circuitry operable to extract motion data and texture data for a plurality of groups of pixels of input digital video; and    (b) second circuitry coupled to an output of said first circuitry, said second circuitry operable to aggregate said motion data and to aggregate said texture data and to insert a resynchronization word between said aggregated motion data and said aggregated texture data.    
   
   
       3 . The encoder of  claim 2 , wherein: 
 (a) each of said group of pixels are the pixels of a 16 by 16 macroblock of pixels of a frame.    
   
   
       4 . The encoder of  claim 2 , wherein: 
 (a) each of said group of pixels are the pixels of a 16 by 16 macroblock of pixels of a frame and lying within an object within said frame.    
   
   
       5 . The encoder of  claim 2 , wherein: 
 (a) each of said group of pixels are the pixels of an 8 by 8 block of pixels of a frame.    
   
   
       6 . The encoder of  claim 2 , wherein: 
 (a) each of said group of pixels are the pixels of an 8 by 8 block of pixels of a frame and lying within an object within said frame.    
   
   
       7 . The encoder of  claim 2 , further comprising: 
 (a) third circuitry coupled to said second circuitry and operable to encode said motion data with a variable length code.    
   
   
       8 . The encoder of  claim 2 , further comprising: 
 (a) third circuitry coupled to said second circuitry and operable to encode said texture data with a variable length code.    
   
   
       9 . The encoder of  claim 2 , wherein: 
 (a) said first circuitry is also operable to extract shape data for an image object which includes said groups of pixels; and    (b) said second circuitry is also operable to insert a second resynchronization word which separates said shape data from said aggregated motion data and said aggregated texture data.    
   
   
       10 . The encoder of  claim 2 , wherein: 
 (a) said first circuitry is a programmable processor executing a first program; and    (b) said second circuitry is said programmable processor executing a second program.    
   
   
       11 . A decoder for motion-compensated video, comprising: 
 (a) first circuitry operable to interpret a first sequence of symbols as aggregated motion data of groups of pixels and interpret a second sequence of symbols as aggregated texture data of said groups of pixels wherein said first sequence and said second sequence are separated by a resynchronization word.    
   
   
       12 . The decoder of  claim 11 , wherein: 
 (a) each of said group of pixels are the pixels of a 16 by 16 macroblock of pixels of a frame.    
   
   
       13 . The decoder of  claim 11 , wherein: 
 (a) each of said group of pixels are the pixels of a 16 by 16 macroblock of pixels of a frame and lying within an object within said frame.    
   
   
       14 . The decoder of  claim 11 , wherein: 
 (a) each of said group of pixels are the pixels of an 8 by 8 block of pixels of a frame.    
   
   
       15 . The decoder of  claim 11 , wherein: 
 (a) each of said group of pixels are the pixels of an 8 by 8 block of pixels of a frame and lying within an object within said frame.    
   
   
       16 . The decoder of  claim 11 , further comprising: 
 (a) a variable-length-code decoder coupled to an input of said first circuitry and operable to output said motion data.    
   
   
       17 . The decoder of  claim 11 , further comprising: 
 (a) a variable-length-code decoder coupled to an input of said first circuitry and operable to output said texture data.    
   
   
       18 . The decoder of  claim 11 , wherein: 
 (a) said first circuitry is also operable to interpret a third sequence of symbols as shape data for an image object which includes said groups of pixels, wherein a second resynchronization word separates said shape data from said aggregated motion data and said aggregated texture data.    
   
   
       19 . The decoder of  claim 11 , wherein: 
 (a) said first circuitry is a programmable processor executing a first program.