A combination of Super-resolution and Deep Learning Approaches applied to Image Forgery Detection

There are a number of techniques used to create forgery images and with the case of copy-move or splicing techniques normally compatible resolutions of original and faked images are assumed in the detection. This would make the process of forgery images becoming monotony and not much challenges. In this paper, we propose an appropriate method to detect tampered images with the change in resolutions of the splicing areas that have been inserted within the images then a combination method of the superresolution approach and the deep learning technique would make an efficient method for image forgery detection. Specifically, we implement the CNN model namely VGG16 trained by the VGG network. We use the VGG16 model with 16 classes. With the test model mentioned above, the results were given with an accuracy of up to 94.64%.