Research on image steganography based on conditional Invertible Neural Network DOI Creative Commons

Menghua Liang,

Hongtu Zhao

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Dec. 13, 2024

Abstract To improve the security of image steganography, an steganography method based on conditional Invertible Neural Network is proposed in this paper. First, we design a to obtain high-quality stego images with rich high-level semantic information and clear spatial details. Based directivity Network, can accurately adjust ensure controllability content. We introduce dual cross-attention module into network structure. The integration modules enhances feature extraction captures complex details steganographic accuracy. In addition, introduction convolutional block attention layer direct model's focus key regions, refining quality. increase number blocks, which improves efficiency reuse. A large experiments are carried out datasets. For cover pairs, PSNR value reached 43.62dB, for secret recovery 46.48dB. Experimental results show that quality better than other state-of-the-art methods.

Language: Английский

A transformer-based adversarial network framework for steganography DOI
Chaoen Xiao, Sufang Peng, Lei Zhang

et al.

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: 269, P. 126391 - 126391

Published: Jan. 10, 2025

Language: Английский

Citations

0
DKiS: Decay weight invertible image steganography with private key DOI
Hang Yang, Yitian Xu, Xuhua Liu

et al.

Neural Networks, Journal Year: 2025, Volume and Issue: 185, P. 107148 - 107148

Published: Jan. 16, 2025

Language: Английский

Citations

0
Research on image steganography based on a conditional invertible neural network DOI

Menghua Liang,

Hongtu Zhao

Signal Image and Video Processing, Journal Year: 2025, Volume and Issue: 19(3)

Published: Jan. 28, 2025

Language: Английский

Citations

0
High-Capacity Image Hiding via Compressible Invertible Neural Network DOI
Chang-Guang Wang,

Haoyi Shi,

Qingru Li

et al.

Lecture notes in computer science, Journal Year: 2025, Volume and Issue: unknown, P. 36 - 53

Published: Jan. 1, 2025

Language: Английский

Citations

0
MFI-Net: multi-level feature invertible network image concealment technique DOI Creative Commons
Dapeng Cheng,

Minghui Zhu,

Bo Yang

et al.

PeerJ Computer Science, Journal Year: 2025, Volume and Issue: 11, P. e2668 - e2668

Published: Feb. 14, 2025

The utilization of deep learning and invertible networks for image hiding has been proven effective secure. These methods can conceal large amounts information while maintaining high quality security. However, existing often lack precision in selecting the hidden regions primarily rely on residual structures. They also fail to fully exploit low-level features, such as edges textures. issues lead reduced model generation results, a heightened risk network overfitting, diminished generalization capability. In this article, we propose novel method based networks, called MFI-Net. introduces new upsampling convolution block (UCB) combines it with dense that employs parametric rectified linear unit (PReLU) activation function, effectively utilizing multi-level (low-level high-level features) image. Additionally, frequency domain loss (FDL) is introduced, which constrains secret be cover are more suitable concealing data. Extensive experiments DIV2K, COCO, ImageNet datasets demonstrate MFI-Net consistently outperforms state-of-the-art methods, achieving superior metrics. Furthermore, apply proposed digital collection images, significant success.

Language: Английский

Citations

0
A robust high-resolution remote sensing image hiding network DOI
Shiyuan Wang, Fumin Wang, Mingqiang Guo

et al.

Neurocomputing, Journal Year: 2025, Volume and Issue: unknown, P. 130178 - 130178

Published: April 1, 2025

Language: Английский

Citations

0
JSN: Design and Analysis of JPEG Steganography Network DOI Open Access
Po-Chyi Su,

Y.S. Cheng,

Tien-Ying Kuo

et al.

Electronics, Journal Year: 2024, Volume and Issue: 13(23), P. 4821 - 4821

Published: Dec. 6, 2024

Image steganography involves hiding a secret message within an image for covert communication, allowing only the intended recipient to extract hidden from “stego” image. The can also be itself enable transmission of more information, resulting in applications where one is concealed another. While existing techniques embed similar size into cover with minimal distortion, they often overlook effects lossy compression during transmission, such as when saving images commonly used JPEG format. This oversight hinder extraction To address challenges posed by steganography, we propose Steganography Network (JSN) that leverages reversible deep neural network its backbone, integrated encoding process. We utilize 8×8 Discrete Cosine Transform (DCT) and consider quantization step specified create JPEG-compliant stego discuss various design considerations conduct extensive testing on JSN validate performance practicality real-world applications.

Language: Английский

Citations

1
Sensitivity of a Convolutional Neural Network for Different Pooling Layers in Spatial Domain Steganalysis DOI Creative Commons

Yoggy Harisusilo Putra,

Bayu Aditya Triwibowo,

Erick Delenia

et al.

Ingénierie des systèmes d information, Journal Year: 2024, Volume and Issue: 29(4), P. 1653 - 1665

Published: Aug. 21, 2024

In the modern era, numerous research studies consistently affirm superior performance of Convolutional Neural Networks (CNNs) over traditional machine learning methods in steganalysis, a technique used to detect hidden data through steganography.Deep Learning (DL), particularly CNNs, is powerful tool for steganalysis because it can handle large datasets effectively.Despite CNNs being widely various areas, previous have primarily focused on improving image classification (cover or stego), often neglecting thorough exploration experimental setup.This aims assess sensitivity CNN-based model by investigating impact different pooling layers state-of-the-art models.The experiments involve five recently proposed models.Significantly, choice goes beyond mere improvement; also addresses overfitting.The results reveal significant diversity based selected layers, namely maximum, average, and mixed pooling, emphasizing importance optimizing objectives when choosing particular approach.This highlights evolving nature this field study need careful consideration layer selection effective steganalysis.

Language: Английский

Citations

0
Image deraining via invertible disentangled representations DOI
Xueling Chen, Xuan Zhou, Wei Sun

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 137, P. 109207 - 109207

Published: Aug. 30, 2024

Citations

0
Robust steganographic approach using generative adversarial network and compressive autoencoder DOI
Malik Qasaimeh,

Alaa Abu Qtaish,

Shadi Aljawarneh

et al.

Multimedia Tools and Applications, Journal Year: 2024, Volume and Issue: unknown

Published: Nov. 13, 2024

Language: Английский

Citations

0