🤖 AI Summary
To address the critical clinical need for joint modeling of tooth instance segmentation and spatial orientation in panoramic radiographs, this paper proposes an enhanced FUSegNet framework. Methodologically, it introduces a grid-based attention gating mechanism embedded within skip connections to strengthen multi-scale feature fusion, and integrates a PCA-driven oriented bounding box (OBB) generation strategy conditioned on segmentation masks to achieve geometrically consistent tooth orientation estimation. Evaluated on the DNS dataset (543 images), the method achieves 82.43% IoU, 90.37% Dice, and 82.82% RIoU—surpassing state-of-the-art approaches in both segmentation accuracy and orientation estimation. This work is the first to unify grid-based attention mechanisms and OBB-based orientation estimation within an end-to-end segmentation architecture, enabling high-fidelity anatomical parsing essential for dental implant planning and personalized treatment.
📝 Abstract
Accurate teeth segmentation and orientation are fundamental in modern oral healthcare, enabling precise diagnosis, treatment planning, and dental implant design. In this study, we present a comprehensive approach to teeth segmentation and orientation from panoramic X-ray images, leveraging deep learning techniques. We build our model based on FUSegNet, a popular model originally developed for wound segmentation, and introduce modifications by incorporating grid-based attention gates into the skip connections. We introduce oriented bounding box (OBB) generation through principal component analysis (PCA) for precise tooth orientation estimation. Evaluating our approach on the publicly available DNS dataset, comprising 543 panoramic X-ray images, we achieve the highest Intersection-over-Union (IoU) score of 82.43% and Dice Similarity Coefficient (DSC) score of 90.37% among compared models in teeth instance segmentation. In OBB analysis, we obtain the Rotated IoU (RIoU) score of 82.82%. We also conduct detailed analyses of individual tooth labels and categorical performance, shedding light on strengths and weaknesses. The proposed model's accuracy and versatility offer promising prospects for improving dental diagnoses, treatment planning, and personalized healthcare in the oral domain. Our generated OBB coordinates and codes are available at https://github.com/mrinal054/Instance_teeth_segmentation.
