A new feature integration approach and its application to 3D model retrieval

• Main Authors: Yu-Cheng Wang, 王宇晨
• Other Authors: Jau-Ling Shih
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/15345372690358562302

碩士 === 中華大學 === 資訊工程學系(所) === 96 === The rapid generation on digital information has made the development of efficient multimedia retrieval tools become urgently. In recent year, content-based multimedia retrieval system becomes the popular research. Various features are proposed to extract the diverse characteristic of multimedia data. These features can be integrated to improve accuracy of content-based multimedia retrieval system. We propose the automatic relevance selecting method (ARSM) to automatically determine the relevant information for improving retrieval result. The ARSM is used to individually combine with query point movement (QPM), grade evaluation (GEM) and re-weighting feature combination (RFC) methods. Finally, our feature integration method is used to automatically determine feature weighting and modify query feature vector. Experimental results show that proposed method is superior to others.