Hybrid optimization enabled Eff-FDMNet for Parkinson's disease detection and classification in federated learning

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Imaging

Background:

• Parkinson's Disease (PD) is a progressive neurodegenerative disorder.
• Early diagnosis is critical for effective symptom management and slowing disease progression.

Purpose of the Study:

• To propose a novel framework, FedL_WSSO based Eff-FDMNet, for accurate PD detection and classification.
• To leverage federated learning and advanced deep learning models for improved diagnostic reliability.

Main Methods:

• Image preprocessing using Gaussian filters and augmentation.
• Feature extraction followed by PD detection using ShCNN-Fuzzy-ZFNet.
• PD classification via WSSO-trained Eff-FDMNet with CAViaR-based server updates.

Main Results:

• Achieved highest accuracy of 0.927.
• Obtained mean average precision of 0.905.
• Reported lowest false positive rate (FPR) of 0.082, loss of 0.073, MSE of 0.213, and RMSE of 0.461.

Conclusions:

• The developed FedL_WSSO based Eff-FDMNet framework demonstrates high accuracy and low error rates for PD detection.
• This potent framework has the potential to enhance patient outcomes through reliable and personalized diagnosis.