NA-MIC Project Weeks

Back to Projects List

Third molar extraction classification

Key Investigators

• Roberto Veraldi (Magna Graecia University of Catanzaro, Italy)
• Amerigo Giudice (Magna Graecia University of Catanzaro, Italy)
• Paolo Zaffino (Magna Graecia University of Catanzaro, Italy)
• Maria Francesca Spadea (Karlsruhe Institute of Technology, Germany)

Presenter location: In-person

Project Description

The classification of third molar extraction is a key factor in oral surgery. Developing a deep learning model to classify the difficulty score of extraction would be useful for surgeons and dentists. This project aims to create a Slicer module that allows clinicians to obtain an extraction-difficulty grade by providing just the patient CT.

Objective

To expose an already developed deep learning classifier in Slicer.

Approach and Plan

1. Identification of optimal classification parameters
2. Expose weights into Slicer
3. Generate extension

Progress and Next Steps

Done during this week:

1. Obtained pth file with the model for deep learning classification.
2. Implemented module extention in Slicer.
3. Tested if the same label obtained in testing was the same that appeared in output in Slicer.

Future steps:

1. Integrating weight files for the specific classification (maybe giving to the clinicians the possibility to download locally the right weights for their specific tasks).
2. Specify what label score means.
3. Other modifications for a general usage of the extention.

Illustrations

Background and References

1. GitHub Project Page: https://github.com/robsver/3DSlicerClassificator
2. Classification score table for third molar extraction: Juodzbalys, Gintaras, and Povilas Daugela. “Mandibular third molar impaction: review of literature and a proposal of a classification.” Journal of oral & maxillofacial research 4.2 (2013).