Grey Level Distribution Analysis for Block-Based Texture Synthesis

• Main Authors: Chun-Yi Wu, 吳俊億
• Other Authors: Chung-Ming Kuo
• Format: Others
• Language: zh-TW
• Published: 2014
• Online Access: http://ndltd.ncl.edu.tw/handle/04826488406908643488

碩士 === 義守大學 === 資訊工程學系 === 102 === Texture synthesis plays an important role in computer graphics. In most practical applications, texture synthesis can be used for 3D object texture mapping, image recovery, virtual reality, etc. The texture synthesis algorithms can be divided into two categories: pixel-based and block-based method. In recent research patch-based method is getting attention increasingly because it can preserve the structures of the source image with low computational cost. In order to improve the texture quality and keep structure continuity effectively, we consider the entire size of the source texture as the size of synthesis patch. In this thesis, we propose a method based on grey level distribution analysis to identify what type the source texture is. According to the texture structures, texture images can be divided into two different types: structure texture and statistics structure texture. Then the appropriate block size can be determined based on the categorization. Experimental results indicate that the proposed identification process achieves more adaptive texture synthesis.