Automated PAR Index Calculation System
Team
- E/22/151, M.I.M. Imaadh, e22151@eng.pdn.ac.lk
- E/22/060, R. DAKSHAYANI, e22060@eng.pdn.ac.lk
- E/22/405, D.A. Victor, e22405@eng.pdn.ac.lk
- E/22/056, P.G.D.D. Chandrawansha, e22056@eng.pdn.ac.lk
Supervisors
- Dr. Asitha Bandaranayake, asithab@eng.pdn.ac.lk
- S. Ganathipan, e21148@eng.pdn.ac.lk
Table of Contents
Introduction
Orthodontics is a specialised field focusing on diagnosing and treating irregularities of the teeth and jaws. Central to this field is the Peer Assessment Rating (PAR) Index, a widely accepted quantitative metric used internationally to measure malocclusion and evaluate treatment outcomes.
However, conventional PAR scoring is performed manually. This process is time-consuming, labour-intensive, and subject to significant variability between different examiners.
The Solution This project builds upon previous research to develop a Web-Based Software System for Automated PAR Index Calculation. While earlier work demonstrated the feasibility of using Machine Learning to detect dental landmarks, this project focuses on transitioning that research prototype into a robust, usable, and extensible software system suitable for real-world clinical use.
Key Objectives:
- Clinical Usability: Developing a clinician-friendly web interface for uploading 3D dental models and visualizing scores.
- MLOps Integration: Implementing retraining pipelines to allow the AI models to improve continuously as new data becomes available.
- Security: Ensuring secure data handling and encryption suitable for hospital environments.
- Scalability: Designing an architecture that supports multi-institution access, moving beyond a single local machine.
The system is being developed in collaboration with Dr. H.S.K. Ratnatilake (Faculty of Dental Sciences) to ensure it meets strict clinical workflows and practitioner needs.
Conclusion
This project aims to solve a genuine clinical bottleneck in orthodontic diagnostics. By automating the PAR Index calculation, the system significantly improves diagnostic objectivity and consistency while saving clinicians substantial time during audits.
Beyond immediate clinical use, the system enables large-scale outcome analysis which was previously unfeasible due to manual constraints. The project has already attracted international interest from orthodontic groups in Japan and South Korea, demonstrating its potential to evolve from an academic prototype into a globally usable clinical software product.