Analysis of Parkinson's Disease SPECT Images Using Geometric Measures and Orthogonal Moments

Abstract

Parkinson's disease being a neurodegenerative disorder commonly seen in older age group has become an alarming health issue. The reduced dopaminergic activity in the striatum region leads to serious complications when not diagnosed at the early stage. Symptomatic changes can be seen in these patients which includes both motor and non motor indications. In this work, an attempt is made to discriminate between the normal and Parkinson's disease subjects by morphological changes in the striatum. SPECT neuroimaging modality has been used for this study. DAT SPECT images were downloaded and analyzed. The striatum region on both the left and right hemispheres is segmented. Geometrical changes seen in the segmented region is quantified by various parameters. Geometric features such as area, axis length, extent and circularity measures show significant changes differentiating between normal and Parkinson's disease. The orthogonal image moments extracted from the striatum discriminates between HC and PD with higher significant values. Thus the experimental results show a decreased dopamine activity region on the striatum in Parkinson's disease patients.