Enhanced Edge Detection and Segmentation Framework with Modified Yen's Thresholding for Parkinson Disease Diagnosis

Parkinson's Disease (PD) analysis is presently based on using accurate image evaluation strategies to identify pathological adjustments in images from MRI scans and DaT scans. Int his work, a new approach that exploits the combination of Gaussian Blur, Canny Edge Detection, Morphological Closing, and Modified Yen's Thresholding. The framework seeks to enhance the accuracy and reliability of PD analysis through a better fixed boundary separation of precise images. The preliminary processing step of Gaussian Blur is introduced to reduce noise and smoothen the image, at the same time as Canny Edge Detection is employed to discover edges correctly within the picture. Morphological Closing is a manner that makes the edges smoother through connecting the disconnected edges. Moreover, Modified Yen's Thresholding, a dynamic thresholding technique, determines the top-quality threshold values based on the neighborhood versions of depth, accordingly, it improves the accuracy of feature detection and segmentation. The evaluation of the proposed scheme becomes achieved on a dataset which covered 78 MRI scans and DaT Scans. The proposed work gives the accuracy of 91%, the precision of 88%, recall of 92%, the F1 score of 89%, and the low loss of 0.22. Thus the proposed model executed without over fitting and therefore it's far suitable for accurate and reliable PD diagnosis.