ViT Enhanced Privacy-Preserving Secure Medical Data Sharing and Classification

TL;DR

This paper proposes a novel secure framework combining a learnable encryption method with Vision Transformer (ViT) to enable privacy-preserving medical data sharing and classification with high security and minimal performance loss.

Contribution

It introduces a learnable encryption scheme integrated with ViT, enhancing data privacy and security specifically for medical image analysis.

Findings

Robust against bit and difference attacks

Maintains high classification accuracy

Ensures data privacy during sharing

Abstract

Privacy-preserving and secure data sharing are critical for medical image analysis while maintaining accuracy and minimizing computational overhead are also crucial. Applying existing deep neural networks (DNNs) to encrypted medical data is not always easy and often compromises performance and security. To address these limitations, this research introduces a secure framework consisting of a learnable encryption method based on the block-pixel operation to encrypt the data and subsequently integrate it with the Vision Transformer (ViT). The proposed framework ensures data privacy and security by creating unique scrambling patterns per key, providing robust performance against leading bit attacks and minimum difference attacks.