International Journal of Engineering and Advanced Research Technology

In this paper, a robust feature extraction technique for content based image retrieval proposed. Most of the research scholars do with the different feature extraction for feature extraction techniques to enhance image retrieval recognition efficiency. Content based image retrieval algorithms concentrate on the local meannearest neighbor classifier and use its decision rule to supervise the design of a discriminator to verify the retrieval images from the training database. Firstly the input color images can be represented as different sub-blocks to take the features from the train database widely uses COREL database. Then those images are applied for local mean nearest neighbor classifier. The proposed method is a robust feature extraction for content based image retrieval gives better results in terms of recognition rate and less computational rate as compared to existing techniques.