DISTRIBUTED SOURCE CODING OF STILL IMAGES (WedAmOR3)

Author(s) : Cagatay Dikici (INSA de Lyon, France) Radhouane Guermazi (INSA de Lyon, France) Khalid Idrissi (INSA de Lyon, France) Atilla Baskurt (INSA de Lyon, France)

Abstract : We propose a compression scheme for still images, by exploiting the theory of Distributed Coding of correlated multisources. Two corrupted versions of an image are encoded separately but decoding jointly. Our approach results in twofold. i) use of decomposition of lw-pass wavelet coefficients for creating the Side Information, and ii) variable-length coset creation by estimating the bit-rate of the cosets on the encoder using the joint distribution statistics of the original image and the side info. In the case of coding for mobile terminals, the proposed codec exploits the channel coding principles in order to have a simple encoder with a low transmission rate and high PSNR. Experimental results are given for lossy encoding case.

Menu