Performance analysis of feature extraction and classification techniques in CBIR

Abstract

Content Based Image Retrieval (CBIR) plays an important role in multimedia search engine optimization. The most useful feature extraction techniques are Principal Component Analysis (PCA), Linear discriminant analysis (LDA), Independent Component Analysis (ICA). These techniques are used to extract the important features from a query image. Support Vector Machine (SVM) and Nearest Neighbour (NN) are two most renowned classification techniques. In this paper we analyse the performance of feature extraction techniques (PCA, LDA, and ICA) and classification techniques (SVM, NN) used in CBIR. The performance metrics are Recognition Rate, F-Score. Based on this performance evaluation models, it is observed that Principal Component Analysis with Support Vector Machine provide more recognition accuracy than others.