Domain Adaptation for Efficiently Fine-tuning Vision Transformer with Encrypted Images

TL;DR

This paper introduces a domain adaptation technique for fine-tuning Vision Transformers with encrypted images, maintaining high accuracy and preventing performance degradation in privacy-preserving scenarios.

Contribution

It presents a novel domain adaptation method that leverages ViT's embedding structure to fine-tune models on encrypted images without accuracy loss.

Findings

Prevents accuracy degradation with encrypted images on CIFAR datasets.

Effective domain adaptation for privacy-preserving image analysis.

Maintains model performance despite data transformation.

Abstract

In recent years, deep neural networks (DNNs) trained with transformed data have been applied to various applications such as privacy-preserving learning, access control, and adversarial defenses. However, the use of transformed data decreases the performance of models. Accordingly, in this paper, we propose a novel method for fine-tuning models with transformed images under the use of the vision transformer (ViT). The proposed domain adaptation method does not cause the accuracy degradation of models, and it is carried out on the basis of the embedding structure of ViT. In experiments, we confirmed that the proposed method prevents accuracy degradation even when using encrypted images with the CIFAR-10 and CIFAR-100 datasets.