Due to the advancement in technologies there is an increase in the number of digitized images, hence there is a requirement of large image dataset. to manage such a large dataset the most common method used is CBIR.In this system we obtain images which are most likely to be the same as the input image given by the user. Here retrieval is done based on different properties of the image, which will be represented as feature vectors. In our paper we resize the image, then extract features of the images and then the query feature vector is matched using different similarity measurements. At the output end images will be classified.
Keywords : Content-based image retrieval, image database, PCA, FLD, Manhattan distance, indexing, Euclidean distance, subspace mixture model, Minkowski distance.