Robust , Variable Bit - rate Coding Using

We demonstrate the use of a Diierential Vector Quantization (DVQ) architecture for the coding of digital images. An Artiicial Neural Network (ANN) is used to develop entropy-biased codebooks which yield substantial data compression without entropy coding and are very robust with respect to transmission channel errors. We discuss how these codebooks can be used to realize variable bit-rate coders. Two methods are presented for variable bit-rate coding using the described DVQ algorithm. In the rst method, both the encoder and the decoder have multiple codebooks of diierent sizes. In the second, variable bit-rates are achieved by using subsets of one xed codebook. We compare the performance of these approaches under conditions of error-free and error-prone channels. Our results show that this coding technique is very resistant to channel errors, and yields pictures of excellent visual quality at moderate compression rate.