Multilayered contourlet based image compression

• Authors: Fang Liu;Yanli Liu
• Venue: MMM'07 Proceedings of the 13th international conference on Multimedia Modeling - Volume Part I
• Year: 2007

After studying contourlet transform and multilayered image representation, we present a Multilayered Contourlet Based Image Compression algorithm (MCBIC) in this paper. It decomposes image into a superposition of coherent layers: piecewise smooth regions layer, directional information layer, e.g., textures and edges. MCBIC uses the best basis to deal with different layers of multilayered representation, and retains the most significant structures of the image. In MCBIC, the first layer of the image is coded in wavelet, which acquires information of piecewise smooth regions; because the contourlet transform is an efficient directional multiresolution image representation, so the second layer of the image is coded in contourlet, which captures directional structure information of the image. Furthermore, each layer is encoded independently with a different transform and the combination of the compressed layers can always be perfect reconstructed. Our experiments demonstrate that MCBIC is efficient in coding images that possess mostly textures and contours. Our experimental results also show that MCBIC is competitive to the contourlet algorithm, SPIHT algorithm and the multilayered image compression approach in terms of the PSNR-rate curves, and is visually superior to these algorithms for the mentioned images.