Automatic Mandible Segmentation Using Convolutional Neural Networks and Feature Sharing

Mandible segmentation is a crucial step in various dental procedures and is the focus of this work. It uses a new neural network for it. The initial results were promising, achieving 98% accuracy. However, due to class imbalance, these results did not fully reflect reality. The customized architecture is still in the testing phase and has not been fully evaluated, so no conclusive results have been obtained thus far. Nevertheless, it is undergoing extensive validation and demonstrates strong potential to significantly advance automation in dental imaging.