Image Manipulation Detection using Convolutional Neural Network

Using various methods, an image manipulation can be done not only by the image manipulation itself, but also by the criminals of counterfeiters for the purpose of counterfeiting. Digital forensic techniques are needed to detect the tampering and manipulation of images for such illegal purposes. In this paper, we present an image manipulation detection algorithm using deep learning technology, which has achieved remarkable results in recent researches. First, a convolutional neural network that is verified for image processing is applied. In addition, a high pass filter is used to acquire hidden features in the image rather than semantic information in the image. For the experiments, modified images are generated using median filtering, Gaussian blurring, additive white Gaussian noise addition, and image resizing for 256x256 images that were divided into 4 equal parts of Boss Base 1.01 images. Quantitative performance analysis is performed to test the performance of the proposed algorithm and image manipulation is detected with 95% accuracy.