Color-Based Image Retrieval Using Compact Binary Signatures

Significant research has focused on determining efficient methodologies for retrieving images in large image databases. This thesis addresses the design and implementation of a new image abstraction technique based on compact signature bit-strings and an appropriate similarity metric. Performance evaluation of image retrieval on a large heterogeneous database of up to 50,000 images demonstrated that the proposed technique is able to outperform the use of Global Color Histograms by up to 55%, and that of Color Coherence Vectors by up to 20% in terms of retrieval effectiveness – this relative advantage was also observed when using classical Precision vs. Recall curves. Perhaps more important is the fact that the proposed approach saves 75% (87.5%) of storage space when compared to Global Color Histograms (Color Coherence Vectors). This makes it possible to store and search an image database of reasonable size, without the aid of (complex) disk-based access structures.