A Study on Block Based Vs Object-Based Stereo Image Compression
The first stereo image compression algorithm was to code the sum and difference of the two images. However, the performance of this technique decreases with increased disparity values. A modification to this method is to shift one of the images horizontally to the point where the cross correlation between the images of stereo pair reaches its maximum value. The shifted image is then subtracted from its partner image and the difference is encoded. This method, assumes that objects in the scene have similar disparity values and thus is not particularly efficient. Another approach is to translate the row blocks instead of the whole image.