A Study of Image Retrieval System Based on Feature Extraction, Selection, Classification and Similarity Measurements

The image contents are considerably increasing in the present digital world. The retrieval of images plays a major role in various domains such as geographical information satellite systems, medical diagnosis, industry inspection, web searching, biometrics and so on. This yields a demand to satisfy human needs for developing a highly effective retrieval systems. A major research efforts have been made in the field of Content-Based Medical Image Retrieval (CBMIR) system. This paper provides a comprehensive review in the field of CBMIR with respect to optimized classification of texture features based similarity framework. Various combinations of feature extraction algorithms are analyzed to extract the texture features. To reduce the high dimension of texture features, a Bio-inspired Meta Heuristic Algorithms (BMHA) are studied for selecting the best features. Machine learning algorithms are analyzed for improving the classification accuracy. Finally similarity measurement is taken to identify the similarity between the query image and database of images. In addition, the importance of various procedures and performance is analyzed, besides limitations of each technique. The Precision and Recall are used as a performance metrics to evaluate the CBMIR systems. From the study, it is suggested to use a hybrid algorithms for developing an effective CBMIR systems.