JPEG-Phase-Aware Convolutional Neural Network for Steganalysis of JPEG Images

Chen, Mo; Sedighi, Vahid; Boroumand, Mehdi; Fridrich, Jessica

doi:10.1145/3082031.3083248

Cited by 156 publications

(90 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The BN normalizes the distribution of each feature to a zero-mean and a unit-variance, and eventually, scales and translates the distribution. The benefit of using a BN layer is that it desensitizes the training to the parameters initialization [18], allows the use of a larger learning rate which speeds up the learning, and improves the detection accuracy [7]. Note that similarly to ResNet [9], and in contrast to Xu-Net, we provide a BN layer accompanied by a scale layer.…”

Section: Yedroudj-netmentioning

confidence: 99%

See 1 more Smart Citation

Yedroudj-Net: An Efficient CNN for Spatial Steganalysis

Yedroudj

Comby

Chaumont

2018

2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

205

152

View full text Add to dashboard Cite

For about 10 years, detecting the presence of a secret message hidden in an image was performed with an Ensemble Classifier trained with Rich features. In recent years, studies such as Xu et al. have indicated that well-designed convolutional Neural Networks (CNN) can achieve comparable performance to the two-step machine learning approaches.In this paper, we propose a CNN that outperforms the state-ofthe-art in terms of error probability. The proposition is in the continuity of what has been recently proposed and it is a clever fusion of important bricks used in various papers. Among the essential parts of the CNN, one can cite the use of a pre-processing filterbank and a Truncation activation function, five convolutional layers with a Batch Normalization associated with a Scale Layer, as well as the use of a sufficiently sized fully connected section. An augmented database has also been used to improve the training of the CNN.Our CNN was experimentally evaluated against S-UNIWARD and WOW embedding algorithms and its performances were compared with those of three other methods: an Ensemble Classifier plus a Rich Model, and two other CNN steganalyzers.

show abstract

Section: Yedroudj-netmentioning

confidence: 99%

“…The results were close to those of the state-of-the-art. In [7], the network is built with a phase-split inspired by the JPEG compression process. An ensemble of CNNs was required to obtain results that were slightly better than those of the state-of-the-art.…”

Section: Introductionmentioning

confidence: 99%

Yedroudj-Net: An Efficient CNN for Spatial Steganalysis

Yedroudj

Comby

Chaumont

2018

2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

205

152

View full text Add to dashboard Cite

show abstract

“…In channel-wise convolution, each input channel corresponds to standalone K output channels and is convolved with an array of K kernels. As a result with J input channels we can get J × K output channels: The existing deep-learning steganalyzers [20]- [28], [30] incorporate the domain knowledge behind rich models, and initialize the kernels in the bottom convolutional layer as highpass filters to increase SNR (Signal-to-Noise Ratio) 3 . Fixed weights in the bottom kernels are adopted in most existing deep-learning steganalyzers [20]- [24], [26]- [28], [30], while learnable weights are adopted in Ye's model [25].…”

Section: A Preliminariesmentioning

confidence: 99%

“…Softmax / -recipe of prior works [25]- [28], [30]. The size of the output feature maps of the normal convolutional layers from bottom to top in this stage is 256 × 256, 128 × 128, and 32 × 32, respectively.…”

Section: Fully Connection /mentioning

confidence: 99%

WISERNet: Wider Separate-Then-Reunion Network for Steganalysis of Color Images

Zeng

Tan

Liu

et al. 2019

IEEE Trans.Inform.Forensic Secur.

View full text Add to dashboard Cite

Until recently, deep steganalyzers in spatial domain have been all designed for gray-scale images. In this paper, we propose WISERNet (the wider separate-then-reunion network) for steganalysis of color images. We provide theoretical rationale to claim that the summation in normal convolution is one sort of "linear collusion attack" which reserves strong correlated patterns while impairs uncorrelated noises. Therefore in the bottom convolutional layer which aims at suppressing correlated image contents, we adopt separate channel-wise convolution without summation instead. Conversely, in the upper convolutional layers we believe that the summation in normal convolution is beneficial. Therefore we adopt united normal convolution in those layers and make them remarkably wider to reinforce the effect of "linear collusion attack". As a result, our proposed wide-and-shallow, separate-then-reunion network structure is specifically suitable for color image steganalysis. We have conducted extensive experiments on color image datasets generated from BOSSBase raw images and another large-scale dataset which contains 100,000 raw images, with different demosaicking algorithms and downsampling algorithms. The experimental results show that our proposed network outperforms other state-of-the-art color image steganalytic models either hand-crafted or learned using deep networks in the literature by a clear margin. Specifically, it is noted that the detection performance gain is achieved with less than half the complexity compared to the most advanced deeplearning steganalyzer as far as we know, which is scarce in the literature.

show abstract

“…Xu et al [10] designed a compact and effective CNN architecture with multiple batch normalization layers. This network now becomes a basic model for several more complex CNN models [11][12][13]. Wu et al [14][15] proposed a novel CNN model to detect steganography based on residual learning and achieved low detection error rates when cover images and stego images are paired.…”

Section: Introductionmentioning

confidence: 99%

A Novel Convolutional Neural Network for Image Steganalysis With Shared Normalization

Zhong

Liu

2020

IEEE Trans. Multimedia

View full text Add to dashboard Cite

Deep learning based image steganalysis has attracted increasing attentions in recent years. Several Convolutional Neural Network (CNN) models have been proposed and achieved state-of-the-art performances on detecting steganography. In this paper, we explore an important technique in deep learning, the batch normalization, for the task of image steganalysis. Different from natural image classification, steganalysis is to discriminate cover images and stego images which are the result of adding weak stego signals into covers. This characteristic makes a cover image is more statistically similar to its stego than other cover images, requiring steganalytic methods to use paired learning to extract effective features for image steganalysis. Our theoretical analysis shows that a CNN model with multiple normalization layers is hard to be generalized to new data in the test set when it is well trained with paired learning. To hand this difficulty, we propose a novel normalization technique called Shared Normalization (SN) in this paper. Unlike the batch normalization layer utilizing the mini-batch mean and standard deviation to normalize each input batch, SN shares same statistics for all training and test batches. Based on the proposed SN layer, we further propose a novel neural network model for image steganalysis. Extensive experiments demonstrate that the proposed network with SN layers is stable and can detect the state of the art steganography with better performances than previous methods.

show abstract

JPEG-Phase-Aware Convolutional Neural Network for Steganalysis of JPEG Images

Cited by 156 publications

References 15 publications

Yedroudj-Net: An Efficient CNN for Spatial Steganalysis

Yedroudj-Net: An Efficient CNN for Spatial Steganalysis

WISERNet: Wider Separate-Then-Reunion Network for Steganalysis of Color Images

A Novel Convolutional Neural Network for Image Steganalysis With Shared Normalization

Contact Info

Product

Resources

About