Real - Time Dynamic Compression of Video on a Grid - Connected Par - Allel Computer

We consider parallel algorithms for real-time compression of video. Our model of computation is a systolic pipe, where processors are arranged in a linear array (each processor is connected only to its left and right neighbors). The algorithm presented combines quantization techniques with dynamic (adaptive) on-line textual substitution methods for data compression. We first consider an arbitrarily large linear array of processors with a serial data stream (for compression of one-dimensional sources such as digitized speech). We then consider a parallel data stream (where columns of video pixels can be read in unit time) and the implementation of the algorithm on an existing grid-connected machine such as the 16,000 processor MPP developed by NASA.