基于矢量图象压缩的信源信道联合Turbo译码.pdf

_______________________________________________________________________________
Joint souce- the
image quality based on the chain algorithm is better than that based on the LBG algorithm.
Furthermore, by using this chain algorithm for vector quantimized images, we add turbo code to
protect some significiant bits of the source data, and better image qualities can be obtained.
This paper is organized as follows. Section II presents the chain algorithm for vector
quantimized images. The JSCC method with turbo code for vector quantimized images is
proposed in Section III. Section IV concludes this paper.

II. CHAIN ALGORITHM

Traditionally the codebook is produced for vector quantimized image by using the LBG
algorithm which hownever does not considered the channel conditions. If the vector quantimized
image is transmitted over noisy channel, some indexes will deviate from the right ones, which will
often cause unpleasant error distion for image. But in this paper, we propose the chain algorithm
for index assignment. Smulations show that this algorithm is robust to channel errors.

Introduction of the chain algorithm

As we know, for vector quantimized images, it is the