Automated pixel-wise analysis of UV fluorescence for standardized PPE exposure assessment: A simulation study

Qualitative assessment of personal protective equipment (PPE) exposure using ultraviolet (UV) fluorescence is often subjective and lacks the precision required for rigorous exposure analysis. This study developed an automated computer vision-based proof-of-concept framework for the objective quantification of simulated PPE exposure during medical procedures. To evaluate the practical utility of the proposed system, 48 simulation sessions were conducted involving five clinical tasks: intravenous (IV) injection, indwelling urinary catheter insertion, oral suctioning, nasogastric tube insertion, and endotracheal intubation using a medical mannequin and a UV-fluorescent tracer. Participating healthcare workers (HCWs) wore Level D PPE, and their spatial exposure patterns were captured via UV photography within a standardized darkroom. A custom Python-based algorithm was then employed to calculate the Fluorescence Coverage Area (FCA, %) by analyzing the ratio of fluorescent pixels to the total PPE surface area through RGB color space segmentation. The mean overall FCA across all procedures was 2.0% ± 1.2%. The system identified distinct spatial patterns, with significantly higher FCA on the anterior surface compared to the posterior surface (3.4% ± 2.3% vs. 0.6% ± 0.5%, p < 0.001), and regionally high exposure on both arms (> 4.8%, p < 0.001). This computer vision-based method offers a standardized and objective approach to PPE exposure assessment, providing empirical data to refine safety protocols and enhance data-driven infection control training for HCWs.