A Fuzzy-based Intelligent Image Retrieval System

• Main Authors: Ming-Cheng Cheng, 鄭明政
• Other Authors: Been-Chian Chien
• Format: Others
• Language: en_US
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/22294526104065827385

碩士 === 義守大學 === 資訊工程學系 === 90 === In the applications of image retrieval, the results of retrieval often can not satisfy the requirements of users effectively. The reason is that these systems can’t match properly between user high-level semantics and low-level image features. In order to narrow down the gap between high-level semantics and low-level image features, we develop a multi-layer image retrieval data model in this thesis. We also propose an architecture for the multi-layer image data model. The architecture consists of image segmentation, features extraction, high-level concepts, high/low level mapping module, query engine and relevance feedback. We develop a fuzzy-based image segmentation approach to segment the original images into meaningful regions. Then the linguistic fuzzy term are used to measure and process the similarity of image features. Through the descriptions of linguistic terms on low-level features, the image can be retrieval by high-level semantic concepts. We implement the proposed system and make same experiments. The experimental results show that the image segmentation algorithm can extract meaningful objects from images effectively and the retrieval results of proposed system can satisfy the requirements of users effectively