Ridgelet Analysis And Its Application In Image Compression

Ridgelet and Curvelet, a new multiscale analysis tool developed by Donoho and his collaborators for representation of higher-dimensional intermittency, are presented in the paper. Ridgelet transform is a combination of radon transform and wavelet transform, with multi-orientations, multiscale and localization. which is very effective at representing objects with singularities along straight lines just as wavelet for objects with isolated point singularities. Curvelet99 transform is based on subband filtering and ridgelet analysis, which is much suit for approximating objects smoothing away from smooth curves.Good non-linear approximation is always followed by sparse presentation and desirable compression. we propose a losssy compression algorithm based on ridgelet transform for images with straight edges . Simulation results in higher compression ratio and better reconstructive image quality comparing with JPEG-2000 which employ wavelet transform...