Hybrid color model for image retrieval based on fuzzy histograms

Abstract

A hybrid color model is a color descriptor formed by combining different channels from several other color models. In computer graphics applications such models are rarely used due to redundancy. However, hybrid color models may be of interest for Content-Based Image Retrieval (CBIR). Best features of each color model can be combined to obtain optimum retrieval performance. In this paper, a novel algorithm is proposed for selection of channels for a hybrid color model used in construction of a fuzzy color histogram. This algorithm is elaborated and implemented for use with several common reference datasets consisting of photographs of natural scenes. Result of this experimental procedure is a new hybrid color model named HSY. Using standard datasets and a standard metric for retrieval performance (ANMRR), it is shown that this new model can give an improved retrieval performance. In addition, this model is of interest for use in JPEG compressed domain due to simpler calculation.