Litcius/Paper detail

An Intriguing Struggle of CNNs in JPEG Steganalysis and the OneHot Solution

Yassine Yousfi, Jessica Fridrich

2020IEEE Signal Processing Letters52 citationsDOI

Abstract

Deep convolutional neural networks (CNNs) have become the tool of choice for steganalysis because they outperform older feature-based detectors by a large margin. However, recent work points at cases where feature-based detectors perform better than CNNs due to their failure to compute simple statistics of DCT coefficients. We introduce a shallow “OneHot” CNN, which encodes DCT coefficients using clipped one-hot encoding into a binary volumetric representation of the DCT plane fed to a convolutional block designed to learn relevant intra-block and inter-block relationships using vanilla and dilated convolutions. Methodology for plugging the “OneHot” network into conventional steganalysis CNNs is also introduced for an end-to-end learnable detector with improved performance.

Topics & Concepts

SteganalysisComputer scienceJPEGArtificial intelligenceSteganographyPattern recognition (psychology)Computer securityData compressionEmbeddingAdvanced Steganography and Watermarking TechniquesDigital Media Forensic DetectionGenerative Adversarial Networks and Image Synthesis
An Intriguing Struggle of CNNs in JPEG Steganalysis and the OneHot Solution | Litcius