Article Preview
Top1. Introduction
The-state-of-art escalation in the digital data production steadily due to the dominant use of internet and digital equipment in various fields such as medical (Mitra, Murthy, & Pal, 2004; Zhang, Brady, & Smith, 2001), entertainment, education, media, online business etc. becoming by keywords. This makes the system cumbersome to manage the abundant data is and human annotations as if in text-based systems. Therefore, here is a tremendous call for an efficient system to retrieve the precise images from the massive database rather than labels. Today Content based image retrieval (CBIR) system is the most well-known system for some applications, CBIR consists the important and essential steps such as feature extraction, relevance feedback, similarity measurements etc., and here feature extraction is the most prominent step in preprocessing that depends on the technique make use of extract the features from the only image like local data as color, texture, shape, human faces etc., further features are categorized into local features such as color features, layout, shape features and texture (Deng, Manjunath, Kenney, Moore, & Shin, 2001; Manjunath, Ohm, Vasudevan, & Yamada, 2001) and high-level features such as faces, biometric, neural networks etc. To retrieve an accurate image with the help of one and only feature is arduous due to the probability of user taking photographs in any direction like several regions which include random direction of capturing of image, optical device, uneven illumination and that of posing expressions, relevance feedback etc. (Jing, Li, Zhang, & Zhang, 2004; Su, Zhang, Li & Ma, 2003) so the system demands the combination of two or multiple features and filtering process. A general and upgraded survey is typified in the further sections.