CONTENT-BASED IMAGE RETRIEVAL : A SHORT-TERM AND LONG-TERM LEARNING APPROACH

The ability to search through images based on their content rather than on their low-level features is becoming more important as the number of available images grows. To deal with this problem, we propose a content-based image retrieval system that uses a new short learning technique combined with a novel long term learning approach. The short term learning technique splits an image into several regions and then applies fuzzy support vector machine learning to these regions. The long term learning approach attempts to adaptively learn the semantic concepts represented in the images using relevance feedback and semantic clustering. One important advantage of the proposed system is its ability to scale to large image databases. Experiments comparing the proposed system with several other state-of-the-art techniques show the effectiveness of the proposed system.