Paired mini-batch training: A new deep network training for image forensics and steganalysis

Abstract

Deep convolutional neural networks (convnets) have recently become popular in many research areas because convnets can extract features automatically and classify them with high accuracy. Researchers in the image forensics and steganalysis field have proposed methods using convnets to develop technologies that work in practical environments. However, they found that the convnets used for computer vision were not suitable for image forensics and steganalysis because these convnets tend to learn features that represent the contents of images rather than forensic or steganalysis features. To overcome this limitation, researchers have proposed various structures, but there are no studies that take into account other factors related to training neural networks for image forensics and steganalysis. In this paper, we clearly represent the training process for image forensics and steganalysis using a training equation and explain why training convnets with the standard mini-batch is inefficient for image forensics and steganalysis. We then propose a new mini-batch, called the paired mini-batch, which is better suited for image forensics and steganalysis.