A New Approach to JPEG Tampering Detection Using Convolutional Neural Networks
Andrey V. Kuznetsov · 2019 International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON) · 2019
The digital world is overloaded with information nowadays. It is quite hard to distinguish between original and fake news. That is why digital image forgery detection algorithms became very popular and develop actively. One of the ways to hide important area of an image is to replace it with another image part. If the original image was not compressed and the new image part was JPEG compressed, then we need some algorithm to distinguish between compressed and non-compressed blocks. That is why in this paper we propose a new algorithm for JPEG compressed local tampering detection using deep convolutional neural networks. The algorithm description and comparison of the proposed approach with existing algorithms are presented in the paper. Our approach shows better quality results in comparison to existing methods and can be used for detection of JPEG compressed areas with different quality factors.