Healthcare Privacy-Preserving Federated Transfer Learning using CKKS-Based Homomorphic Encryption and PYHFEL Tool

Digitization of healthcare data has shown an urgent necessity to deal with privacy concerns withinthe field of deep learning for healthcare organizations. A promising approach is federated transfer learning,enabling medical institutions to train deep learning models collaboratively through sharing model parameters ratherthan raw data. The objective of this research is to improve the current privacy-preserving federated transferlearning systems that use medical data by implementing homomorphic encryption utilizing PYthon forHomomorphic Encryption Libraries (PYFHEL). The study leverages a federated transfer learning model to classifycardiac arrhythmia. The procedure begins by converting raw Electrocardiogram (ECG) scans into 2-D ECGimages. Then, these images are split and fed into the local models for extracting features and complex patternsthrough a finetuned ResNet50V2 pre-trained model. Optimization techniques, including real-time augmentationand balancing, are also applied to maximize model performance. Deep learning models can be vulnerable toprivacy attacks that aim to access sensitive data. By encrypting only model parameters, the Cheon-Kim-Kim-Song(CKKS) homomorphic scheme protects deep learning models from adversary attacks and prevents sensitive rawdata sharing. The aggregator uses a secure federated averaging method that averages encrypted parameters toprovide a global model protecting users’ privacy. The system achieved an accuracy rate of 84.49% when evaluatedusing the MIT-BIH arrhythmia dataset. Furthermore, other comprehensive performance metrics were computed togain deeper insights, including a precision of 72.84%, recall of 51.88%, and an F1-score of 55.13%, reflecting abetter understanding of the adopted framework. Our findings indicate that employing the CKKS encryption schemein a federated environment with transfer cutting-edge technology achieves relatively high accuracy but at the costof other performance metrics, which is lower in the encrypted settings when compared to the plain one, anacceptable trade-off to ensure data privacy through encryption with achieving an optimal model performance.