
Automatic Measurement of the Root Canal Working Length
Software as a Service
NoCategory
AI & Machine LearningClients
Techstack
Purpose
This thesis aimed to automate a critical step in endodontic treatment: measuring the root canal working length from periapical X-ray images, replacing manual estimation with a repeatable AI-assisted measurement.
Description
A PyQt5 desktop application segments root canals from periapical radiographs using a U-Net model built with segmentation_models_pytorch and trained in PyTorch Lightning. Captured X-ray film is digitized by a Raspberry Pi camera rig served over a Flask API, then preprocessed with cropping and CLAHE contrast enhancement. From each predicted mask, a medial-axis skeleton and graph-based path analysis trace the canal and convert its pixel length to millimeters, with results and annotated images saved per patient.




