Content Based Image retrieval is a process of retrieving the images similar to the query image. The most tedious job is retrieving an accurate image for the query image. The collection of images retrieved must be accurate at the same time efficient. The retrieved images are accurate or not can be known only by the direct feedback of the users. The aim of this paper , is to retrieve images similar to the query image by using Faster Region-based Convolutional Neural Networks.Computation of a correlation-based feature matrix is made that comprises of contextual information from the retrieval list which fed into a faster regionbased convolutional neural network regression model for evaluating the quality of the retrieved images.

Keywords : CBIR, Neural networks, Faster RCNN, image retrieval, image processing