•  
  •  
 

Abstract

This work introduces a new approach to protecting the data in the healthcare applications of federated learning based on the classification of skin cancer. The recommended solution established and prevents the data poisoning attacks by using deep learning and CNN architectures namely VGG16. In a federated learning system which comprises of ten healthcare facilities, the approach enables the training of models in a collaborative way without compromising the medical data or the patients’ information. Data is meticulously prepared and preprocessed using the Skin Cancer MNIST: According to the HAM10000 dataset. As for the federated learning approach, VGG16’s feature extraction capability is employed to classify skin cancer. A powerful mechanism to identify the threats of data poisoning in federated learningis proposed in the work. Heuristic and rigorous methods identify and assess undesirable changes in the models through the outlier detection techniques. The performance evaluation confirms that the proposed model works and is accurate, private, and immuneto data poisoning. This work provides a federated learning skin cancer categorisation for the healthcare domain that is both secure and exact. The presented strategy enhances the diagnostic of the healthcare system and pays attention to protecting data privacy and security in FL environments when addressing data poisoning attacks.

Download

Share

COinS