Design and implementation of a large scale content based image retrieval system

• Main Author: Yee, Sau Wen.
• Other Authors: Hoi Chu Hong
• Format: Final Year Project
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Computer science and engineering::Information systems::Information storage and retrieval
• Online Access: http://hdl.handle.net/10356/17017
• Institution: Nanyang Technological University

The purpose of this report is to describe the research and solution to the problem of designing a web-based large scale Content Based Image Retrieval (CBIR) system, named LSCBIR. The final LSCBIR system has indexed 1 million images collected from flickr. In order to narrow down the semantic gap between high-level concepts and low-level features, the multi-modal image retrieval which uses both text and content-based searching will be investigated. To allow user interact with the system, relevance feedback (RF is implemented using Support Vector Machines (SVM) active learning. Next, the description of the primate features of an image and the algorithms used to calculate the similarity between extracted features, are explained. To enhance system’s completeness, a user management system and a system administration application are included. Finally, experiments are conducted to evaluate the performance of the proposed algorithm. The experiment results show combination of effective text and content-based searching results has a better retrieval performance than the individual content-based searching results.