Novel Image coding and Inverse Halftoning

Texture is an important feature in images and has§been widely used in many applications. Based on the§classified textures, this book presents a novel§learning- and texture-based approach to design more§efficient image processing algorithms. For§context-based arithmetic coding, the block- and§texture-based training process is first applied to§train the multiple-template (MT) from the most§representative texture features. Based on the MT, we§next present a texture- and MT-based arithmetic§coding algorithm to compress error-diffused images.§For predictive coding, to improve the least§square approach, we present a texture-based training§process to construct the multiple-window (MW) for§various image contents. Based on the MW, the texture-§and MW-based prediction scheme is presented to§compress gray images. For inverse halftoning, based§on the proposed variance gain-based decision tree, a§texture-based training process is presented to §construct a lookup tree-table which will be used in §the reconstructing process. In the reconstructing §process, we propose an edge-based refinement scheme §to enhance the quality of the the§reconstructed gray image.