Title

Adaptive Quantization of Image Subbands with Efficient Overhead Rate Selection

Authors

Y. Yoo, A. Ortega and B. Yu

Where

3rd. Int'l. Conf. on Image Proc. ICIP '96, Lausanne, Switzerland, Sept. 1996.

Abstract

Much of the recent on subband image coding has relied on adaptive quantization to achieve substantial gains over other traditional techniques, for example those based on the DCT. In a typical different quantizers are used for each subband, or for blocks within a given subbands, and the quantizers are explicitly sent to the decoder as overhead. Then, the quantized coefficients are transmitted using adaptive entropy coding, typically through arithmetic coding. In this example forward adaptation is used for the quantizers and backward adaptation for the entropy coding. In this paper we show that a combination of forward and backward adaptation methods can be used to update the quantizers as well. Specifically, we present an algorithm where classification can be done based on the quantized past data and where the quantizer to be used within one class can itself be adapted on the fly.

Download Postscript File