Image Retrieval Method Based on an Improved Particle Swarm Optimization Algorithm

To retrieval images we required with high effectiveness and accuracy, in this paper, we study on how to implement an effective image retrieval system using particle swarm optimization. Firstly, framework of the Particle Swarm Optimization based image retrieval system is described, which is able to integrate three image visual features (such as Bag of visual words feature, Color feature, Texture feature) together by an improved particle swarm optimization. Secondly, we modify the standard PSO algorithm by describing the position and velocity of each particle in a multiple dimensional binary solution space and a continuous space. Afterwards, several most similar images with highest ranking scores for each testing image are obtained by our algorithm. Finally, experimental results show that using the Corel1K and Corel5K dataset, our proposed algorithm can effectively promote the Precision, Recall and F-measure in the image retrieval system.