High-Performance and Low-Power, Wavelet-Transform Visible Image Encoder

The VLSI implementation of a novel image encoder/decoder chip, the Wavelet-Vector Quantization (WVQ), is proposed for use in high resolution, real-time, and low-power image compression applications. This single chip custom hybrid VLSI CMOS processor is based architecturally on the wavelet transform - vector quantization encoding algorithm in image compression and is designed to be used in a focal-plane configuration at the heart of the electronic still camera. The WVQ processor is designed to receive as input a digitized image from the CCD front-end and output a compressed version of the image for subsequent on-board storage or electronic transmission. The unique hardware implementation of the wavelet transform encoder (which incorporates a novel pipelined architecture) and the vector quantization post- processor (which is based on a fast search architecture) will make it possible to generate visually lossless encoded images with compression ratios better than 30:1 on time scales less than a few milliseconds.