A Hybrid PSO and SVM Algorithm for Content Based Image Retrieval

• Published in: Computational Science and Its Applications - ICCSA 2016 Vol. 9786; pp. 583 - 591
• Main Authors: Wang, Xinjian, Luo, Guangchun, Qin, Ke, Chen, Aiguo
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2016; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Classification distance; Image retrieval; Neighbor density; Network hardware; PSO; SVM
• ISBN: 9783319420844; 3319420844
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-319-42085-1_48

In order to improve the speed and accuracy of image retrieval, This paper presents a hybrid optimization algorithm which originates from Particle Swarm Optimization (PSO) and SVM (Support Vector Machine). Firstly, it use PSO algorithm, The image in the database image as a particle in PSO algorithm, After operation, return to the optimum position of the image. Secondly, use SVM to feedback the related images, Use the classification distance and nearest neighbor density to measure the most valuable image, After update classifier, choose the furthest point from the classification hyperplane as target image. Finally, the proposed method is verified by experiment, the experimental results show that this algorithm can effectively improve the image retrieval speed and accuracy.