Deep Residual Network for Steganalysis of Digital Images

By Mehdi Boroumand, Mo Chen, Jessica J. Fridrich. IEEE Transactions on Information Forensics and Security, 2019

911 citations recorded by the source corpus.

Steganography detectors built as deep convolutional neural networks have firmly established themselves as superior to the previous detection paradigm – classifiers based on rich media models. Existing network architectures, however, still contain elements designed by hand, such as fixed or constrained convolutional kernels, heuristic initialization of kernels, the thresholded linear unit that mimics truncation in rich models, quantization of feature maps, and awareness of JPEG phase. In this work, we describe a deep residual architecture designed to minimize the use of heuristics and externally enforced elements that is universal in the sense that it provides state-of-the-art detection accuracy for both spatial-domain and JPEG steganography. The key part of the proposed architecture is a significantly expanded front part of the detector that “computes noise residuals” in which pooling has been disabled to prevent suppression of the stego signal. Extensive experiments show the superior performance of this network with a significant improvement, especially in the JPEG domain. Further performance boost is observed by supplying the selection channel as a second channel.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/04fde97b6bfc08fddbd57e36c84c9b60abe04bf0

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Depth-Wise Separable Convolutions and Multi-Level Pooling for an Efficient Spatial CNN-Based Steganalysis (2020, methodology): “Our proposed CNN network is compared with five popular networks: Xu-Net [17], Ye-Net [19] (without the selectionchannel information), Yedroudj-Net [21], SRNet [39], and SRM + EC standing for the hand-crafted feature set named Spatial-Rich-Model [13] and Ensemble Classifier [32].”
Detection of AI-Generated Synthetic Faces (2022, methodology): “…extraction as often performed in image forensics strategies [52] (residual), do not perform down-sampling in the ﬁrst layer as suggested by [5] (no-down), perform a stronger augmentation (strong-aug) by including Gaussian noise adding, geometric transformations, cut-out, and brightness and…”
HiNet: Deep Image Hiding by Invertible Network (2021, methodology): “Following [32], we gradually increase the amount of training images to re-train the SRNet.”