Image Compression Scheme Based on Multi-resolution Boundary-Based Shape Descriptors

With the development of the internet, image compression becomes more and more essential. Most multi-media based web applications are far from realizing their full potential largely due to the limitations of the traditional image compression methods. Representing a shape only based on its boundary information, the boundary-based shape descriptors are prospective for image compression. Inspired by how brain recalls objects, we propose a novel image compression scheme. This scheme can transform any boundary-based descriptor to its multi-resolution structure. The multi-resolution descriptor represents an image in a cognitive architecture which is idea for image compression with maintaining the global information and the key detail points of this image. This scheme has high computation accuracy and efficiency, strong flexibility, it is free from parameter settings, robust to errors and not restricted by the image formats.