Foreground/Background Separation Using Reference Images

Abstract

A technique involves distinguishing between foreground and background regions of a digital image of a scene. First and second images are captured of nominally a same scene. The first image is a relatively high resolution image taken with the foreground more in focus than the background, while the second image is a relatively low resolution reference image taken with the background more in focus than the foreground. Regions of the captured images are assigned as foreground or background. In accordance with the assigning, one or more processed images are rendered based on the first image or the second image, or both.

What is claimed is: 
     
         1 . A method of distinguishing between foreground and background regions of a digital image of a scene, the method comprising:
 (a) capturing first and second images of nominally the same scene, the first image being a relatively high resolution image taken with the foreground more in focus than the background, and the second image being a relatively low resolution reference image taken with the background more in focus than the foreground;   (b) storing the captured images;   (c) assigning regions of the first image as foreground or background; and   (d) rendering in accordance with the assigning one or more processed images based on the first image or the second image, or both, at a digital rendering device, display or printer, or combinations thereof.   
     
     
         2 . The method of  claim 1 , wherein said storing comprising storing said images in DCT-coded format. 
     
     
         3 . The method of  claim 2 , wherein said assigning comprising assigning regions of said first image as foreground or background according to whether one or more sums of selected higher order DCT coefficients decreases or increases for equivalent regions of the second image. 
     
     
         4 . The method of  claim 3 , wherein regions are assigned as foreground when the one or more sums decreases and as background when the one or more sums increases. 
     
     
         5 . The method of  claim 3 , wherein the selected high order coefficients are the top 25% of coefficients. 
     
     
         6 . The method of  claim 3 , wherein the assigning further comprises:
 (i) calculating the sum of selected high order coefficients for a plurality of DCT blocks of the first image,   (ii) calculating the sum of the same high order coefficients for a plurality of DCT blocks of the second image,   (iii) comparing the sum calculated in step (i) for each DCT block of the first image with the sum calculated in step (ii) for the DCT block of the equivalent region of the second image,   (iv) if the sum calculated for a given DCT block of the first image is greater than the sum calculated for the DCT block of the equivalent region of the second image, assigning the given block as corresponding to a foreground region of the image, and   (v) if the sum calculated for a given DCT block of the first image is less than the sum calculated for the DCT block of the equivalent region of the second image, assigning that block as corresponding to a background region of the image.   
     
     
         7 . The method of  claim 6 , wherein said calculating and comparing is performed as a post processing stage in an external device to said digital camera. 
     
     
         8 . The method of  claim 6 , wherein said calculating and comparing is performed as a post processing stage in a digital camera. 
     
     
         9 . The method of  claim 1 , further comprising aligning said first and second images. 
     
     
         10 . The method of  claim 1 , further comprising creating a digital map from said assigning said blocks as corresponding to background or foreground. 
     
     
         11 . The method of  claim 1 , further comprising equating the image resolution of said first and second images. 
     
     
         12 . The method of  claim 1 , wherein said second image comprises a pre- or post-view version of the first image. 
     
     
         13 . A hand-held or otherwise portable or spatial or temporal performance-based image capture device, comprising:
 (a) one or more lenses, an aperture and a photodetector for capturing an original main image and one or more reference images having a temporal or spatial overlap or proximity with the original image, or combinations thereof;   (b) a processor;   (c) one or more processor-readable media having embedded therein programming code for the processor to perform a digital image processing method of distinguishing between foreground and background regions of a digital image of a scene, the method comprising:   (i) capturing first and second images of nominally the same scene, the first image being a relatively high resolution main image taken with the foreground more in focus than the background, and the second image being a relatively low resolution reference image taken with the background more in focus than the foreground;   (ii) storing the captured images;   (iii) assigning regions of the first image as foreground or background; and   (iv) rendering in accordance with the assigning one or more processed images based on the first image or the second image, or both, at a digital rendering device, display or printer, or combinations thereof.   
     
     
         14 . The image capture device of  claim 13 , wherein said storing comprising storing said images in DCT-coded format. 
     
     
         15 . The image capture device of  claim 14 , wherein said assigning comprising assigning regions of said first image as foreground or background according to whether one or more sums of selected higher order DCT coefficients decreases or increases for equivalent regions of the second image. 
     
     
         16 . The image capture device of  claim 15 , wherein regions are assigned as foreground when the one or more sums decreases and as background when the one or more sums increases. 
     
     
         17 . The image capture device of  claim 15 , wherein the selected high order coefficients are the top 25% of coefficients. 
     
     
         18 . The image capture device of  claim 15 , wherein the assigning further comprises:
 (i) calculating the sum of selected high order coefficients for a plurality of DCT blocks of the first image,   (ii) calculating the sum of the same high order coefficients for a plurality of DCT blocks of the second image,   (iii) comparing the sum calculated in step (i) for each DCT block of the first image with the sum calculated in step (ii) for the DCT block of the equivalent region of the second image,   (iv) if the sum calculated for a given DCT block of the first image is greater than the sum calculated for the DCT block of the equivalent region of the second image, assigning the given block as corresponding to a foreground region of the image, and   (v) if the sum calculated for a given DCT block of the first image is less than the sum calculated for the DCT block of the equivalent region of the second image, assigning that block as corresponding to a background region of the image.   
     
     
         19 . The image capture device of  claim 18 , wherein said calculating and comparing is performed as a post processing stage in an external device to said digital camera. 
     
     
         20 . The image capture device of  claim 19 , wherein said calculating and comparing is performed as a post processing stage in a digital camera. 
     
     
         21 . The image capture device of  claim 13 , wherein the method further comprises aligning said first and second images. 
     
     
         22 . The image capture device of  claim 13 , wherein the method further comprises creating a digital map from said assigning said blocks as corresponding to background or foreground. 
     
     
         23 . The image capture device of  claim 13 , wherein the method further comprises equating the image resolution of said first and second images. 
     
     
         24 . The image capture device of  claim 13 , wherein said second image comprises a pre- or post-view version of the first image.