From Image to Imuge: Immunized Image Generation, official code, implemented by PyTorch, ACMMM 2021 paper
ACMMM 2021 (Poster), Chengdu, China
We have just released the source code of the submitted supplementary material to ACMMM 2021. Feel free to contact me via shinydotcom@163.com if you encounter any problem during using this project.
For simplicity, you can view it as a robust watermarking scheme for image protection.
Digital images are vulnerable to nefarious tampering attacks such as content addition or removal that severely alters the original meaning. It is somehow like a person without protection that is open to various kinds of viruses. Image immunization is a technology of protecting the images by introducing trivial perturbation, so that the protected images are immune to the viruses in that the tampered contents can be auto-recovered. Our method achieves promising results in real-world tests where experiments show accurate tamper localization as well as high-fidelity content recovery. Additionally, we show superior performance on tamper localization compared to state-of-the-art schemes based on passive forensics.