Detecting Tampered Regions in JPEG Images via CNN

In this work, we will detect the tampered regions in JPEG images using deep learning techniques (U Net architecture). Often, digital pictures are used as evidence in criminal investigations. Therefore, it is essential to check whether they have been tampered with or not. DCT coefficients play an important role in the detection of Tampered regions of images and these DCT coefficients are input to the CNN. 

Convolutional neural network (U Net architecture) has been successfully used to achieve good performance in detecting tampered regions of images. U-net is a convolutional network architecture for fast and precise segmentation of images mainly. Experiments will demonstrate that our model will provide the best detection performance compared to the state-of-the-art methods.