This study presents the development of a novel algorithm for the dynamic evaluation of pupil diameter (dynamic pupillometry) over time, aimed at enhancing the assessment of diabetic neuropathy. Benefiting from the principles of computer vision, this research addresses the critical need for precise and non-invasive diagnostic methods for autonomic disorders associated with diabetes. The methodology integrates image processing techniques, including the Hough Transform and moving average algorithms, to accurately measure and analyze pupillary responses from video data captured during pupillometry exams. A custom post-processing framework identifies and corrects outliers, employing a moving average to smooth the data and generate stable trends over time. The results are visualized through time-series plots and 3D representations, offering insightful and comprehensible analyses of pupillary dynamics.