Sign coding and estimation of zero-quantized coefficients in the reformed EZW algorithm

Wavelet transforms have proven to be very powerful tools for image compression. Wavelet transform coefficients are comprised by both a magnitude and a sign. While efficient algorithms exist for coding the transform coefficient magnitudes, most current wavelet image coding algorithms are not as efficient at coding the sign of the transform coefficients. In this paper, sign coding and estimation of zero-quantized coefficients are applied in a reformed embedded zerotree wavelet image encoder. In addition to using intraband wavelet coefficients in a sign coding context model, a projection technique capitalizes on the non-orthogonality of biorthogonal wavelet transform basis vectors to improve the prediction of high-frequency coefficients, that allows nonintraband wavelet coefficients to be incorporated into the context model. At the decoder improved reconstruction estimation of zero-quantized coefficients are derived. It is shown that these techniques yield higher PSNR improvements.