MICCAI STSR 2025 Challenge: Semi-Supervised Teeth and Pulp Segmentation and CBCT-IOS Registration

arXiv — cs.CVWednesday, December 3, 2025 at 5:00:00 AM
  • The MICCAI STSR 2025 Challenge has been announced, focusing on semi-supervised learning for teeth and pulp segmentation using Cone-Beam Computed Tomography (CBCT) and Intraoral Scanning (IOS). The challenge includes two main tasks: segmenting teeth and pulp canals in CBCT images and registering CBCT with IOS data. Participants will have access to a dataset comprising 60 labeled and 640 unlabeled IOS samples, alongside 30 labeled and 250 unlabeled CBCT scans.
  • This initiative is significant as it addresses the critical issue of data scarcity in digital dentistry, which hampers the development of automated solutions for pulp canal segmentation and cross-modal registration. By fostering community participation and encouraging the submission of open-source solutions, the challenge aims to enhance the capabilities of semi-supervised learning in this specialized field.
  • The challenge reflects a growing trend in the application of deep learning techniques in medical imaging, particularly in the realm of dental diagnostics. Innovations such as nnU-Net and Mamba-like State Space Models are being utilized to improve segmentation accuracy, while frameworks like nnActive are emerging to evaluate active learning in 3D biomedical segmentation. This convergence of technology and healthcare highlights the ongoing efforts to refine automated processes and improve patient outcomes in digital dentistry.
— via World Pulse Now AI Editorial System

Was this article worth reading? Share it

Recommended apps based on your readingExplore all apps
Continue Readings
Detecting Dental Landmarks from Intraoral 3D Scans: the 3DTeethLand challenge
PositiveArtificial Intelligence
The 3DTeethLand challenge was announced in collaboration with the International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2024, focusing on the detection of dental landmarks from intraoral 3D scans. This initiative aims to enhance the precision of orthodontic diagnostics and treatment monitoring by leveraging advanced deep learning techniques to address the complexities of individual tooth geometries.
U-Mamba2: Scaling State Space Models for Dental Anatomy Segmentation in CBCT
PositiveArtificial Intelligence
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.
CT Synthesis with Conditional Diffusion Models for Abdominal Lymph Node Segmentation
PositiveArtificial Intelligence
A new pipeline integrating conditional diffusion models and nnU-Net has been developed to enhance the segmentation of abdominal lymph nodes, addressing challenges such as complex environments and limited annotated data. This approach utilizes LN-DDPM, a conditional denoising diffusion probabilistic model, to generate diverse and realistic lymph node data, improving segmentation performance.
The MICCAI Federated Tumor Segmentation (FeTS) Challenge 2024: Efficient and Robust Aggregation Methods for Federated Learning
PositiveArtificial Intelligence
The MICCAI Federated Tumor Segmentation (FeTS) Challenge 2024 has been announced, focusing on federated learning for glioma sub-region segmentation in multi-parametric MRI. The challenge evaluated six teams using a standardized setup and a dataset from the BraTS glioma benchmark, which included 1,251 training cases and 570 hidden test cases. Teams were ranked based on segmentation performance and communication efficiency, with a PID-controller-based method achieving the top score.