Content-Based Image Retrieval using colour feature and colour bit planes

Abstract

In this paper, we proposed a novel fusion approach using global and local features. Colour Histogram (CH) is used as global feature and Block Bit Plane (BBP) as a local feature. For similarity measurement, distance d1 is used for CH features and BBPs are compared bit by bit using Hamming distance. Finally the overall similarity is computed as a weighted combination of global and local features. Particle swarm optimisation is used for optimisation of these weights. The proposed method is tested on 1000 and 3000 image database demonstrated significant improvement in precision and average recall rate compared to the current state of art methods.