Image Tampering Detection with ELA Transform and Convolutional Neural Network

With the advancement in technology, it has become easier to tamper images and change minute details. Detecting forgeries is getting next to impossible with naked eye. In the recent years we have seen an exponential growth in the field of convolutional neural network (CNN) and their use cases. In order to accurately determine if an image is tampered or authentic, CNN has proven itself to be quite resourceful. In this manuscript, we propose a CNN based model in combination with Error level analysis (ELA) transformation using Discrete Cosine Transform (DCT) coefficients, for detection of forgery like copy-move, image splicing, image retouching. Previously existing models were distinguished with the proposed model and vibrant outcomes depict that our advanced model strikes better accuracy and error detection at a time-saving and thrifty pace.