PyTorch Lightning implementation of a 5-class dental pathology segmentation system using a U-Net with a pre-trained ResNet34 backbone, joint Dice-CE loss, and power-law smoothed class weighting.
python computer-vision deep-learning tensorflow keras pytorch transfer-learning image-segmentation class-imbalance unet semantic-segmentation clinical-data medical-image-segmentation medical-ai pytorch-lightning albumentations dental-ai dental-imaging dental-segmentation caries-detection
-
Updated
Jul 18, 2026 - Jupyter Notebook