arXiv:2509.12069v3 Announce Type: replace 
Abstract: Cone-Beam Computed Tomography (CBCT) is a widely used 3D imaging technique in dentistry, providing volumetric information about the anatomical structures of jaws and teeth. Accurate segmentation of these anatomies is critical for clinical applications such as diagnosis and surgical planning, but remains time-consuming and challenging. In this paper, we present U-Mamba2, a new neural network architecture designed for multi-anatomy CBCT segmentation in the context of the ToothFairy3 challenge. U-Mamba2 integrates the Mamba2 state space models into the U-Net architecture, enforcing stronger structural constraints for higher efficiency without compromising performance. In addition, we integrate interactive click prompts with cross-attention blocks, pre-train U-Mamba2 using self-supervised learning, and incorporate dental domain knowledge into the model design to address key challenges of dental anatomy segmentation in CBCT. Extensive experiments, including independent tests, demonstrate that U-Mamba2 is both effective and efficient, securing first place in both tasks of the Toothfairy3 challenge. In Task 1, U-Mamba2 achieved a mean Dice of 0.84, HD95 of 38.17 with the held-out test data, with an average inference time of 40.58s. In Task 2, U-Mamba2 achieved the mean Dice of 0.87 and HD95 of 2.15 with the held-out test data. The code is publicly available at https://github.com/zhiqin1998/UMamba2.

تم تقديم بنية جديدة لشبكة عصبية، U-Mamba2، لعملية تقسيم تشريح الأسنان في التصوير المقطعي المحوسب باستخدام شعاع مخروطي (CBCT)، حيث تعالج التحديات المتعلقة بالتمثيل التشريحي الدقيق، وهو أمر حاسم للتطبيقات السريرية. يدمج هذا النموذج نماذج الفضاء الحالة في بنية U-Net، مما يعزز الكفاءة دون المساس بالأداء.

Se ha presentado una nueva arquitectura de red neuronal, U-Mamba2, para la segmentación de la anatomía dental en la tomografía computarizada de haz cónico (CBCT), abordando los desafíos de la representación anatómica precisa, que es crítica para las aplicaciones clínicas. Este modelo integra modelos de espacio de estado en la arquitectura U-Net, mejorando la eficiencia sin comprometer el rendimiento.

Une nouvelle architecture de réseau de neurones, U-Mamba2, a été introduite pour la segmentation de l'anatomie dentaire dans la tomographie à faisceau conique (CBCT), répondant aux défis de la représentation anatomique précise, essentielle pour les applications cliniques. Ce modèle intègre des modèles d'état dans l'architecture U-Net, améliorant l'efficacité tout en maintenant la performance.

A new neural network architecture, U-Mamba2, has been introduced for the segmentation of dental anatomy in Cone-Beam Computed Tomography (CBCT), addressing the challenges of accurate anatomical representation critical for clinical applications. This model integrates state space models into the U-Net architecture, enhancing efficiency while maintaining performance.

U-Mamba2: Scaling State Space Models for Dental Anatomy Segmentation in CBCT

Was this article worth reading? Share it

LucidQuery AI

Augmeta

Zemith-3bda3b