"Dynamic Pupillary Dataset for Autonomic Neuropathy in Type 1 Diabetes"

"This dataset contains dynamic pupillometry images collected from 49 patients diagnosed with Type 1 Diabetes Mellitus. The primary objective of this dataset is to provide high-resolution data for the assessment of Diabetic Autonomic Neuropathy (specifically Cardiovascular Autonomic Neuropathy) through the analysis of the pupillary light reflex (PLR). The data was acquired using a custom-built monocular pupillometer integrated into a virtual reality headset, equipped with a Raspberry Pi NoIR Camera V2 capturing at 30 frames per second. The experimental protocol involved a 2-minute dark adaptation period to allow the pupil to reach maximum dilation, followed by a 10-millisecond white light flash stimulus to trigger parasympathetic constriction. In total, the dataset includes 90 pupillometry videos, resulting in 13,500 grayscale images of the right eye with a resolution of 640x480 pixels. A subset of 4,264 images was manually annotated by a specialist, providing bounding boxes for both the iris and the pupil to serve as a reliable ground truth for computer vision applications."