Image reconstruction from very low bitrate multiscale edge model

摘要

In this paper, a progressive image reconstruction scheme based on a very low bit-rate multiscale edge model is proposed. This mutliscale edge model decomposes and represents the image with its multiscale primal sketch and the background. Multiscale primal sketch is the union of the pulse edges and the ramp edges extracted at consecutive scales. Background is the small image survived after a hierarchical scale transform of edge-removed image. Edge smoothing is performed to remove edges at each scale. For the reconstruction, a bottom-up approach is used by synthesizing edges on the reconstructed image obtained from previous scale. As edge synthesis is conducted at consecutive scales, the visual quality of the reconstructed image is progressively increasing. Experiment shows that the proposed scheme is advantageous at both low bitrate representation and progressiveness.