Image vaccine against steganography in encrypted domain
Abstract This paper investigates on the defense against steganography, and the overall purpose of the study is to design a satisfactory defense scheme in encrypted domain. Image vaccine against steganography is an effective technique to discover the utilization of steganography with extremely high d...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-02-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-88384-8 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract This paper investigates on the defense against steganography, and the overall purpose of the study is to design a satisfactory defense scheme in encrypted domain. Image vaccine against steganography is an effective technique to discover the utilization of steganography with extremely high detection accuracy. However, the image owner and vaccine provider are not the same person usually. To meet the requirements of steganography defense and privacy protection simultaneously, this paper proposes a vaccine scheme against steganography for encrypted images. After encrypting the entire data of a original image using a stream cipher, the vaccine data can be injected into the image without knowing the image content. With an encrypted image containing vaccine data, one can decrypt it to obtain the vaccinated image. When steganography is executed on vaccinated image, the utilization of steganography can be discovered in encrypted domain. Experimental results show that the detection accuracy of our scheme on steganography is 100% for all cases. That means the utilization of steganography can be always detected using our scheme. Integrate image vaccine into the imaging process of digital cameras in IoT systems is a potential practical application of our scheme. Non-universal detection mechanism is the potential limitations of this study, and it may be solved by pre-processing original image instead of injecting specific data. |
|---|---|
| ISSN: | 2045-2322 |