Improving the Robustness of JPEG Steganography With Robustness Cost
| Autor: | Xiaolei He, Xianfeng Zhao, Hong Zhang, Jimin Zhang |
|---|---|
| Rok vydání: | 2022 |
| Předmět: |
Distortion function
Steganography Pixel Computer science Applied Mathematics ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Data_CODINGANDINFORMATIONTHEORY computer.file_format Function (mathematics) JPEG Reduction (complexity) Robustness (computer science) Signal Processing Discrete cosine transform Electrical and Electronic Engineering computer Algorithm |
| Zdroj: | IEEE Signal Processing Letters. 29:164-168 |
| ISSN: | 1558-2361 1070-9908 |
| DOI: | 10.1109/lsp.2021.3129419 |
| Popis: | Due to a large number of user-uploaded images, social networks have become secure channels for covert communication. However, the JPEG recompression of social networks changes the DCT coefficients of stego images, resulting in the failure of adaptive steganography. To achieve steganography in lossy channels, robust steganography has been proposed. In this letter, the ability against JPEG recompression of robust steganography is further improved by introducing a robustness cost function. For calculating the robustness cost, a robust- ness model based on the spatial domain calculated from DCT coefficients is firstly proposed. Then the robustness cost is acquired by measuring the distance between the spatial pixels calculated from modified DCT coefficients and the robustness model adjusted spatial pixels. Combining the distortion function and the robustness cost function, the method proposed has a good performance on both robustness and security. Experimental results show that with the maximum reduction of 4.04% on security performance, the algorithm proposed has a significant improvement on robustness performance compared with state- of-the-art robust steganography. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|