Detection of Parkinson’s Disease on DaTSCAN Image Using Multi-kernel Support Vector Machine

Parkinson's disease (PD) is a degenerative illness of central nervous system primarily caused by neuronal degeneration in substantia nigra of the brain.A biomarker for Parkinson's disease (PD) is represented by blood uric acid level.Although, the relationship between Parkinson's disease, and diabetes and the outcome of specific treatments remains unclear.Quantitative analysis of the images can increase the potential of dopamine transporter (DAT) and single-photon emission computed tomography (SPECT) images as a biomarker for tracking PD progression.To predict Parkinson's disease using DaTscan images, an ensemble of Machine Learning (ML) models was developed.In this work, Parkinson's Progressive Markers Initiative (PPMI) datasets were utilized for collecting the data on PD.Initially, feature extraction was done using the VGG-16, and AlexNet to classify Parkinson's illness.In following stage, a predicted classification of Parkinson's cases and healthy controls is generated using the Multi-kernel Support Vector Machine (MSVM) model to enhance overall final output of classification model.PPMI has made a database that is accessible to individuals for use in evaluating the proposed model.Compared with existing methods of Explainable Boosting Machine (EBM), Light Gradient Boosting Machines (LGBM) & Random Forests (RF), Convolutional Neural Network (CNN), Bayesian CNNs, CNN with Fuzzy Rank Level Fusion (CNN+FRLF), and EfficientNet-B0 and MobileNet-V2 models the implemented MSVM achieves high accuracy.When compared with the existing methods the implemented MSVM method obtained 98.60% of accuracy.