Rate-distortion model and analytical bit allocation for wavelet-based region of interest coding

We introduce a novel rate-distortion (RD) model for images coded with a progressive wavelet coder, such as SPIHT (set partitioning in hierarchical trees) or JPEG2000, and especially designed to capture RD behavior when different parts of an image are refined at different speeds. Our model is an extension of Mallat´s model (Mallat and Falzon 1998), which takes into account that the rates used are not necessarily the same throughout the image. Because of the different rates, certain modeling approximations (e.g., those for coarse quantization) can not be used uniformly throughout the image. Allocating different numbers of bits to different regions in an image is a problem that appears in applications such as region of interest (ROI) coding or multiple description coding (MDC), where the wavelet coefficients are divided by different factors before coding to enable different bit allocations to different regions. We show that the proposed RD model provides a more accurate estimate than Mallat´s model.