Block-based long-term content-based image retrieval using multiple features

This paper proposes a novel content-based image retrieval technique, which integrates block-based visual features and user´s query concept-based semantic features. It also facilitates short-term and long-term learning processes by integrating users´ historical relevance feedback information. The history is compactly stored in a semantic feature matrix and efficiently represented as semantic features of the images. The short-term relevance feedback technique can benefit from long-term learning. The high-level semantic features are dynamically updated based on users´ query concept and therefore represent the image´s semantic meaning more accurately. Our extensive experimental results demonstrate that the proposed system outperforms its seven state-of-the-art peer systems in terms of retrieval precision and storage space.