An Efficient MDC based Set Partitioned Embedded Block Image Coding

Abstract

In this paper, fast, efficient, simple and widely used Set Partitioned Embedded bloCK based coding is done on
Multiple Descriptions of transformed image. The maximum potential of this type of coding can be exploited with discrete
wavelet transform (DWT) of images. Two correlated descriptions are generated from a wavelet transformed image
to ensure meaningful transmission of the image over noise prone wireless channels. These correlated descriptions are
encoded by set partitioning technique through SPECK coders and transmitted over wireless channels. Quality of
reconstructed image at the decoder side depends upon the number of descriptions received. More the number of
descriptions received at output side, more enhance the quality of reconstructed image. However, if any of the multiple
description is lost, the receive can estimate it exploiting the correlation between the descriptions. The simulations
performed on an image on MATLAB gives decent performance and results even after half of the descriptions is
lost in transmission