A Quality-Controlled Computational Workflow for Analysing Articulating Marks and Cross-Condition Occlusal Contact Changes: Development and Pilot Technical Evaluation

Background/Objectives: Articulating marks indicate contact locations but do not reliably quantify occlusal force. This pilot study developed and technically evaluated a quality-controlled workflow for analysing scanned pigment marks across LIGHT and STRONG clenching conditions. Methods: Twenty-one dentate participants underwent colour intraoral scanning after foil marking. Eighteen formed the developmental pool; three were reserved for held-out technical workflow and quality control evaluation. Detector assessment and random forest candidate-validity classification used annotations from three participants, with internal evaluation grouped by participant. Comparisons of anatomical regions were descriptive. Alignments were classified by registration quality control as PASS, REVIEW or FAIL. Results: Among 337 detected candidates, 271 were annotated as genuine marks, 65 as artefacts and one as indeterminate; 30 additional marks were missed. Relative to detector-assisted manual reference annotations, developmental detector precision, recall and F1-score were 80.65%, 90.03% and 85.09%, respectively. Across 152 anatomical regions, marks were recorded in both conditions in 38, only in STRONG in 52 and only in LIGHT in 62. These descriptive differences do not establish individual physical-mark correspondence. Held-out registration classifications comprised one PASS, three REVIEW and two FAIL. Evaluation involved small participant subsets with clustered observations, limiting generalisability beyond the evaluated acquisition conditions. Conclusions: Conceived as a proof of concept, the workflow demonstrated technical feasibility. Under predefined quality control criteria, automated interpretation was withheld for FAIL registrations, affecting 37 probable LIGHT contacts. Independent predictive validity and clinical effectiveness remain unestablished; force and tooth displacement were not measured. The workflow provides a foundation for future predictive models incorporating tooth mobility and force-dependent occlusal dynamics, requiring calibrated measurements and independent validation to establish clinical relevance.