Recolored Image Detection

Abstract: Recoloring is a technique that can be transferred image color or them and result in an indispensable transition in human eyes. Image reminder is one if the most important image handling techniques are not special this method is designed to detect frauds. In this sheet, we propose a systematic end-end-end system Pictures from the landscape pictures, The proposed network basically the original image and two received entries light stability and inter-channel contract considering the original input and the probability release it comes to mind. Our algorithm accepts CNN-based profound character the structure consists of three feature extraction modules and a feature fusion module. To train the deep neural network, lets consolidate a database with image that we have remained The fact that that relative quality of using different reuses methods is true. Detailed diagnostic results were created in the created films various methods show that our proposed network is good Generic s=and very robust. We are the first attempt to distinguish recolored images from natural images. We analyze the inte-channel correlation and illumination consistency for natural images which may not hold after the color transfer operation. Based on these two properties, we propose a deep discriminative model for recoloring detection using for Gray scale algorithm.