Method of extracting contour of image, method of extracting object from image, and video transmission system using the same method

Abstract

An object extraction method comprising the steps of comparing each portion of first image data obtained by capturing an extraction object with another portion of the first image data or second image data, thereby generating shape data as an image representing a provisional region of the extraction object in the first image data, and matching a contour of the shape data to a contour of the extraction object by using the first image data and the provisionally generated shape data.

1 - 7 . (canceled)  
   
   
       8 . An object extraction method comprising: 
 inputting image data representing an original image and shape data representing each of provisional regions of objects in the original image;    arranging a plurality of search reference blocks, each having a first size, along a contour portion of each of the provisional regions;    searching same image as the search reference blocks for a similar block, for each of the search reference blocks, which has a texture similar to that of a corresponding search reference block of the search reference blocks and a block size larger than that of the search reference block;    size-correcting shape data obtained from a corresponding one of the similar blocks by reduction processing using a pixel value of one of pixels around a sampling point of the shape data as a sampling value; and    replacing the shape data in each of the search reference blocks with the size-corrected shape data obtained from a corresponding one of the similar blocks, the shape data having different pixel values in an object region and a background region which differ from each other.    
   
   
       9 - 17 . (canceled)  
   
   
       18 . The method according to  claim 8 , further including raster-scanning the shape data to sequentially set a block around a pixel which has a pixel value different from that of an adjacent pixel and fails to be included in any of previously set blocks.  
   
   
       19 . The method according to  claim 8 , wherein the blocks are linked together in an overlapping state.  
   
   
       20 . A computer system comprising: 
 means for inputting image data representing an original image and shape data representing each of provisional regions of objects in the original image;    means for arranging a plurality of search reference blocks, each having a first size, along a contour portion of each of the provisional regions;    means for searching the same image as the search reference blocks for a similar block, for each of the search reference blocks, which has a texture similar to that of a corresponding search reference block of the search reference blocks and a block size larger than that of the search reference block;    size-correcting shape data obtained from a corresponding one of the similar blocks by reduction processing using a pixel value of one of pixels around a sampling point of the shape data as a sampling value; and    replacing the shape data in each of the search reference blocks with the size-corrected shape data, the shape data having different pixel values in an object region and a background region which differ from each other.