An Efficient Coding Method for Teleconferencing and Medical Image Sequences

Abstract

In this paper, an efficient video coding, based on three-dimensional vector quantization system has been proposed. There are a few contributions of this paper, the proposed technique uses a pyramid code book based model instead of conventional single level code book. According to this model, if a specific motion cannot be matched accurately in higher level code book according to coarse partitioning of the frame, a more refined code book description is to be used. This technique also uses an adaptive vector quantization algorithm to train the code book for optimal performance with varying time. The performance of the proposed technique is further improved by using a potential based codebook post processing technique which enables the vector quantizer to possess higher correlation preservation property. The proposed post-processing technique leads to a much higher inter index correlation. The algorithm not only accelerates the vector quantization processing, but also results in better rate-distortion performance. The above discussed approach can be used both for teleconferencing videos and to compress images obtained from confocal laser scanning microscopy. The experimental results show that the proposed method gives a higher subjective and objective quality of the reconstructed frames at a better compression ratio. When applied to edge detection filter on reconstructed frames it produces more acceptable results than a standard codec. It was observed that the proposed method outperforms the teleconferencing compression standard H.261 and LBG based vector quantization compression techniques