A Closed-Loop System for Dental Plaque Management: Integrating Automated UV Image Analysis with Clinical Decision Feedback

Abstract

Current dental plaque assessment relies on subjective visual scoring (e.g., Turesky index), leading to high inter-operator variability (reported k = 0.4-0.6) and poor feedback loops for longitudinal monitoring. This represents a failure in the healthcare system's ability to provide consistent, actionable data for preventive care. This work presents a human-in-the-loop system for plaque quantification that integrates UV illumination with user-assisted tooth delineation and CIE-LAB color-space analysis. The system implements a five-level nested feedback architecture (pixel, tooth, session, longitudinal, population), positioning the clinician as an adaptive agent within the loop. User-drawn contours define the system boundary, while adaptive thresholding (Otsu's method) on the enhanced b* channel provides core quantification. Real-time adjustable parameters establish a clinician-guided feedback loop, consistent with Ashby’s Law of Requisite Variety, allowing clinician adjustments to match environmental variability in enamel reflectivity, saliva-induced specular reflections, and plaque maturation. A continuous plaque-coverage percentage P is computed prior to ordinal mapping to the Turesky modification of the Quigley-Hein Index (grades 0-5). Validation on clinical images demonstrated discrimination of 13.4% differences among teeth receiving identical Grade 3 scores revealing differences undetectable in routine visual scoring. By converting visual impressions into a quantitative, repeatable process embedded within a cybernetic human-machine framework, this system enables a Learning Health System where longitudinal data can inform preventive policies. The architecture generalizes to other domains that rely on visual assessment (e.g., wound healing, skin lesions), and its structured metrics can support population-level surveillance. The proposed approach offers a practical pathway to integrate objective oral hygiene metrics into digital health records, improving patient feedback, supporting clinical research, and enabling population-level public health surveillance.