Fuzzy Fusion of Colour and Shape Features for Efficient Image Retrieval

Content Based Image Retrieval (CBIR) based on colour has been subjecting to research for many years. While partially successful in resolving some important theoretical and practical problems, it is also becoming clear that other sensory elements will also play the vital part as the supplementary signal inputs in enabling better
judgements on digital imagery. For example, geometry shapes will provide important topological information on the image contents. On contrary to colour stimuli, image retrieval based on shapes is still relatively immature. This is due to the complexity and ambiguity on shape definition and the infinite possibility on shape
combinations. In this paper, an innovative approach based on fuzzy fusion of colours and shapes for image retrieval is presented. In this work, the so-called feature vector will play a pivotal role in streamlining the colour and shape features based on the Pseudo Zernike Moments (PZM) for improving the efficiency and accuracy of a CBIR system.