Image Compression using an Efficient Edge Cartoon + Texture Model

Wavelet-based image coders optimally represent smooth regions and isolated point singularities. However, w av elet coders are less adept at representing perceptually important edge singularities, and coding performance su ers signi cantly as a result. In this paper, w epropose a no veltw o-stageimage coder framework based on modeling images as edge cartoons + textures. In stage 1, w einfer and eÆciently code the edge information from the image using a multiscale wedgelet decomposition. In stage 2, we code the residual, \edgeless" texture image using a standard w av elet coder. Our preliminary coder improves signi cantly o ver standard wavelet coding techniques in terms