Query by Multiple Images for Content-Based Image Retrieval Based on Social Network Algorithms

Autor:	Chang, Wei Ling, 張瑋鈴
Druh dokumentu:	學位論文 ; thesis
Popis:	100 In recent years, with the faster and faster development of computer technology, the number of digital images is grown rapidly so that the Content-Based Image Retrieval has become one of important multimedia technologies. Much research has been done on Content-Based Image Retrieval. However, little research has been done on query by multiple images. This thesis investigates the mechanism for query by multiple images. First, MPEG-7 image features and SIFT are extracted from images. Then, we calculate the similarity of images to construct the proximity graph which represents the similarity structure between images. Last, processing of query by multiple images is achieved based on the social network algorithms. Experimental results indicate the proposed method provides high accuracy and precision.
Databáze:	Networked Digital Library of Theses & Dissertations
Externí odkaz:	http://ndltd.ncl.edu.tw/handle/35279325653362859412 Zobrazit plný text záznamu