Stochastic Texture Compensation For Wavelet Compression

Js Zhang,Gx Zhu

2002-01-01

Abstract:Lossy image compression algorithms are most efficient. A key element is the transform coefficients of low- and medium-amplitude is quantized to zero. The unstructured or stochastic texture of image is represented by these coefficients, and is very sensitive for human visual. The conventional encoding of all these coefficients is very bit-rate expensive. Instead, such an unstructured or stochastic texture can be modeled by a noise process and be characterized with very few parameters. Thus, a new scheme can be designed that encodes the structural image information by a conventional wavelet codec and the stochastic texture in a model-based manner is proposed in this article. Its love computational complexity and the parameter coding costs of only 0.013 bpp make it a valuable extension of conventional codecs. A comparison with the most recent image compression standard JPEG2000 showed that the new scheme achieves the same subjective quality while increasing the compression ratio by more than a factor of two.