Early stage of Parkinson’s Disease Identification Using Advanced Image Processing Techniques

Abstract

Parkinson’s Disease (PD) is a kind of neurodegenerative disorder. There is an imperative need for identifying the early stage of disease as it keeps on affecting the human mid-brain. The incipient level of the disorder is identified with the help of sixteen volume rendering image slices (VRIS) which are taken from a Single Photon Emission Computed Tomography (SPECT) image as a novel tool. These image slices are selected on account of striated intake from the striatum. The shape and texture attributes of segmented VRIS and Striatal Binding Ratio (SBR) values are considered as a feature set for the analysis. These two different features (attribute) are synthesized to identify the difference between Healthy Control (HC) and the early stage of Parkinson’s disease (EPD). The various classifier models like Extreme Learning Machine (ELM), Support Vector Machine (SVM) and Artificial Neural Network (ANN) with different kernel functions are solely designed for the study the impact of single and multi-features to identify EPD. The performance of the present work is investigated and found that the Polynomial ELM offers an appreciated outcome with reference to the accuracy of 99.3%. The outcome has been compared with the previous work to underline the efficacy of the present work. Hence, the present work could be of a great aid to the experts in neurology to protect the neurons from the impairment.