Multiplexing stereo signals using digital watermarking

Abstract

3D-images' depth and aspects are reconstructed by the human eye-brain system by appropriately combining the Left and Right components of such images.. The method for multiplexing Left and Right images in stereoscopic images requires double quantity of information and more bandwidth: with Right (Q amount of information) and Left (Q amount of information), 3D image will require 2*Q amount of information. Our objective is to reduce this amount of information, and send the 3D image over one communication channel instead of two which will save bandwidth. In this paper, we tried different methods for reducing the amount of information from reducing the size of the right image, to compressing one image but they did not give 100% accuracy, and resulted in blurred images. Our proposed solution is to encode Right image based on the Left. Hence, the amount of information will become Q * (1 + α) instead of 2*Q, with α very small. This encoding will be done using one of the disparity methods in order to prepare the image to be sent over one channel using one of the watermarking techniques.