Using visual features to design a content-based image retrieval method optimized by particle swarm optimization algorithm

This paper presents a content-based image retrieval method using three kinds of visual features and 12 distance measurements, which is optimized by particle swarm optimization (PSO) algorithm. For convenience, it is called the CBIRVP method hereafter. First, the CBIRVP method extracts three kinds of features: color, texture, and shape features of images. Subsequently, it employs appropriate distance measurements for each kind of features to calculate the similarities between a query image and others in the database D. Also, the PSO algorithm is utilized to optimize the CBIRVP method via searching for nearly optimal combinations between the features and their corresponding similarity measurements, as well as finding out the approximately optimal weights for three similarities with respect to three kinds of features. Finally, experimental results demonstrate that the CBIRVP method outperforms other existing methods under consideration here.