Development and Validation of an Artificial Intelligence Algorithm for Automated Detection of Dental Biofilm in Intraoral Photographs

• Grant number: 25/24115-9
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: January 01, 2026
• End date: December 31, 2026
• Field of knowledge: Health Sciences - Dentistry - Periodontology
• Principal Investigator: Mohamed Ahmed Hassan Mahmoud
• Grantee: Eduarda Freitas de Andrade
• Host Institution: Universidade Universus Veritas Guarulhos (Univeritas UNG). Guarulhos , SP, Brazil

Abstract

ABSTRACTThe detection and quantification of dental plaque are fundamental steps for the diagnosis and monitoring of periodontal health. Traditional methods present limitations related to the subjectivity of clinical examination. This study aims to develop an Artificial Intelligence algorithm based on Deep Learning for automated detection of dental plaque from standardized intraoral photographs. One hundred photographs were collected from twenty participants at three clinical timepoints: (T0), baseline image without disclosing agent; (T1), image after application of chemical disclosing agent; and (T2), image after professional plaque removal, totaling 300 high-resolution images (6000 × 4000 pixels). Images from timepoint T1 will be used to generate segmentation masks through manual annotation with the open-source LabelMe tool. The algorithm will be developed employing the U-Net architecture, an encoder-decoder convolutional neural network widely applied in medical image segmentation, with a pre-trained encoder (transfer learning) for training optimization. The model will be trained to delineate the dental plaque area. Development will be conducted using Python with deep learning frameworks (TensorFlow) and image processing libraries. Validation will include evaluation of model sensitivity and specificity, with a performance target above 0.80. The results will provide scientific evidence for Artificial Intelligence systems that use digital photographs in dental plaque assessment, with applications in telemedicine and home oral hygiene monitoring.Keywords: Artificial Intelligence; Deep Learning; Dental Plaque; Intraoral Photography; Automatic Segmentation; Preventive Dentistry. (AU)