In this research work HAAR and Gabor wavelet is used for feature extraction process. Feature extraction is the important part in retrieving image from the data set. The features of the media items are extracted and their values and indices are saved in the database. Then the index structure is employed to ideally filter all irrelevant things by checking attributes with the user’s question. Relevance is judged based on the content of media things. With the help of feature extracted from image a histogram is created. For Filtering the photographs median filter is employed. The median filter could be a non –linear digital technique, usually used to remove noise. Jec (joint equal contribution) and l1-l2 (lasso) algorithm helps in similarity comparison of image. It matches the extracted features of image to the dataset images. Lasso is applied to identify those features that are more relevant for capturing image similarity. After getting the similarity images similar to the query image get retrieved. We have used recall and precision as the deciding factor for evaluating the performance of the CBIR.