Privacy-Preserving Federated Vision Transformer Learning Leveraging Lightweight Homomorphic Encryption in Medical AI
It addresses privacy risks in multi-institutional medical AI by preventing model inversion attacks, though it is incremental as it builds on existing federated learning and encryption methods.
This paper tackles the problem of privacy vulnerabilities in federated learning for medical AI by proposing a framework that combines Vision Transformers with homomorphic encryption, achieving 96.12% classification accuracy in the unencrypted domain and 90.02% in the encrypted domain while reducing communication by 30-fold compared to gradient encryption.
Collaborative machine learning across healthcare institutions promises improved diagnostic accuracy by leveraging diverse datasets, yet privacy regulations such as HIPAA prohibit direct patient data sharing. While federated learning (FL) enables decentralized training without raw data exchange, recent studies show that model gradients in conventional FL remain vulnerable to reconstruction attacks, potentially exposing sensitive medical information. This paper presents a privacy-preserving federated learning framework combining Vision Transformers (ViT) with homomorphic encryption (HE) for secure multi-institutional histopathology classification. The approach leverages the ViT CLS token as a compact 768-dimensional feature representation for secure aggregation, encrypting these tokens using CKKS homomorphic encryption before transmission to the server. We demonstrate that encrypting CLS tokens achieves a 30-fold communication reduction compared to gradient encryption while maintaining strong privacy guarantees. Through evaluation on a three-client federated setup for lung cancer histopathology classification, we show that gradients are highly susceptible to model inversion attacks (PSNR: 52.26 dB, SSIM: 0.999, NMI: 0.741), enabling near-perfect image reconstruction. In contrast, the proposed CLS-protected HE approach prevents such attacks while enabling encrypted inference directly on ciphertexts, requiring only 326 KB of encrypted data transmission per aggregation round. The framework achieves 96.12 percent global classification accuracy in the unencrypted domain and 90.02 percent in the encrypted domain.