An adaptive approach for content based image retrieval using Gaussian firefly algorithm

Abstract

An adaptive content based image retrieval (CBIR) approach based on relevance feedback and Gaussian Firefly algorithm is proposed in this paper. Feature extraction has been done with the Euclidean distance estimation between the pixels; relevance feedback (RF) based approach but all concerns with the extraction of image accuracy. This research work has a focused approach to increase the performance by optimizing image feature by adopting with the firefly algorithm (FA). Further, to improve the retrieval accuracy, random walk concepts based on Gaussian distribution is used to move all the fireflies to global best at the end of each iteration. Experiments demonstrate that the proposed method shows more accuracy and better performance compared to particle swarm optimization and genetic algorithm and the use of Gaussian distribution further improve the retrieval accuracy. © Springer-Verlag Berlin Heidelberg 2013.