Hybrid Evolutionary Algorithm Based Relevance Feedback Approach for Image Retrieval

Searching images from the large image databases is one of potential research areas multimedia research. The most challenging task for nay CBIR system to capture high level semantic user. researchers domain are trying fix this issue with help Relevance Feedback (RF). However existing RF based approaches needs a number iteration fulfill user's requirements. This paper proposed novel methodology achieve better results in early reduce user interaction system. In previous work it reported that SVM approach generating CBIR. Therefore, focused on approach. To enhance performance applied Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) before applying feedback. main objective using these meta-heuristic was increase positive sample size SVM. Firstly steps PSO by incorporating feedback secondly GA result generated through PSO, finally GA. technique named as Algorithm- Support Vector Machine (PSO-G A-SVM-RF). Precisions, recall F-score used metrics assessment validation PSO-GA-SVM-RF experiments conducted coral dataset having 10908 images. From experimental proved outperformed then various well known approaches.