An Access Control Method with Secret Key for Semantic Segmentation Models
This addresses security concerns for model owners in computer vision, but it is incremental as it extends existing access control methods from classification to segmentation tasks.
The paper tackles the problem of unauthorized access to semantic segmentation models by proposing a secret key-based encryption method using Vision Transformers, achieving the same accuracy for authorized users and severely degraded accuracy for unauthorized ones.
A novel method for access control with a secret key is proposed to protect models from unauthorized access in this paper. We focus on semantic segmentation models with the vision transformer (ViT), called segmentation transformer (SETR). Most existing access control methods focus on image classification tasks, or they are limited to CNNs. By using a patch embedding structure that ViT has, trained models and test images can be efficiently encrypted with a secret key, and then semantic segmentation tasks are carried out in the encrypted domain. In an experiment, the method is confirmed to provide the same accuracy as that of using plain images without any encryption to authorized users with a correct key and also to provide an extremely degraded accuracy to unauthorized users.