Multi-view residual image coding technique using adaptive quantization and scanning method

In this paper, an effective coding method for stereo and multi-view images using disparity estimation/compensation technique is proposed. The main contribution of this paper is focused on the residual image coding technique which uses the directional properties of DCT coefficients. Quantization and scanning methods are differently applied according to the directions of blocks. The proposed residual image coding technique shows improvements in compression ratio as well as in PSNR compared with the conventional methods. We apply this algorithm to multi-view images and evaluate the performance of the algorithm with optimal reference view selection. Me basically considered the relative size of occlusion area to select the optimal reference view.