Hierarchical Local Binary Patterns for Images and its Application on Image Matching

• Main Authors: Pei-Hsuan Wang, 王珮瑄
• Other Authors: Yao-Hong Tsai
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/79256847779096200598

碩士 === 玄奘大學 === 資訊管理學系碩士班 === 103 === The image matching technology has been developed in these decades. However, due to the variant lighting environment, the system may results in lower accuracy rate. For the technique of image matching, a higher accuracy with a shorter computation time is very important. This paper applys the Hierarchical Local Binary Pattern (HLBP) to describe changes in the characteristics of the texture in images. LBP is a simple and efficient operation for the texture analysis. Image size of the result will be the same size of the original image. LBP is under simple arithmetic mode, and single block operation speed is quite fast. The number of features will affect the amount of analysis time, so we need effective way to reduce them. The reduction ratio of the number of times increases the efficiency. In order to achieve high recognition of the results, we will use the hierarchical way once again to do verification.