Clustering result of image searches by annotations and visual features

• Main Authors: Jeff Chuang, 莊純福
• Other Authors: Fan Wu
• Format: Others
• Language: en_US
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/31811651883440413654

碩士 === 國立中正大學 === 資訊管理所暨醫療資訊管理所 === 98 === Image on web has become one of the most important information for browsers; however, huge number of images retrieved from images search engine makes users hard to find the images they desired. Images search result clustering (ISRC) is proposed to solve this problem. Currently, most ISRC methods do not utilize textual and visual features together to present clustering result, not to mention perform it in a reasonable time; moreover, they only provide one layer clustering result. In this paper, we proposed a new ISRC method called Incremental-Annotations-based image search with clustering (IAISC), which utilizes annotation as textual features and category model as visual features. Our method can provide clustering result based on the semantic meaning and visual trail; furthermore, the result presented by an iteratively structure, which we can expected to make users to find the images they required easily. The simulation result shows our system has high precision that the average precision rate is 73.4%, and the response time performed within two seconds, which shows more efficient than the previous research.