A System for Efficient Retrieval of the Images from Large Datasets using Ripplet Transform and Edge Histogram Detector

Content-Based Image Retrieval (CBIR) is image retrieval approach which allows the user to extract an image from a large database depending upon a user specific query. An efficient and effective image retrieval performance is achieved by choosing the best transform and classification techniques. Currently available transform techniques such as Fourier Transform, Cosine Transform, and Wavelet Transform suffer from discontinuities such as edges in images. To overcome this problem, a technique called Ripplet Transform (RT) has been implemented along with the neural network based classifier called multilayered perceptron (MLP) for finding an effective retrieval of image. Classification using multilayered perceptron (MLP) with the Manhattan Distance measure showed varying experimental results for dimensions of Images. The performance of various Transform is compared to find the of particular wavelet function for image retrieval. In proposed system, training and testing is provided on Wang image database. The results of retrieval of images are expressed in terms of Peak Signal to Noise Ratio (PSNR) values compared with various other proposed schemes to show the superiority of our proposed system. Classification using multilayered perceptron (MLP) with the Manhattan Distance measure showed varying experimental results for different sets of Images.