Self-Supervised Vision Transformers for CBCT-Based Detection of Temporomandibular Joint Osteoarthritis
arXiv:2606.08364v1 Announce Type: cross Abstract: Temporomandibular joint osteoarthritis (TMJ OA) is a prevalent degenerative condition whose osseous changes are often subtle on cone-beam CT (CBCT), making automated detection challenging. We study how well the DINO family of self-supervised vision transformers -- DINOv1, DINOv2, DINOv2+reg, and RAD-DINO (a radiology-pretrained variant) -- transfers to CBCT, asking how much backbone adaptation is needed and of what kind. We propose a simple slice-based pipeline using Vision Transformer (ViT) backbones: axial CBCT slices are encoded per-slice by
The continuous advancements in self-supervised learning and vision transformers are enabling more specialized and efficient AI applications in medical imaging, coinciding with the growing availability of medical datasets.
This development indicates a significant step towards automated, more accurate detection of complex conditions like TMJ OA, which can improve diagnostic workflows and patient outcomes in dentistry and oral medicine.
The reliance on highly specialized human interpretation of medical scans for subtle conditions could decrease as AI models demonstrate robust performance, potentially democratizing access to high-quality diagnostics.
- · Medical AI developers
- · Radiology software providers
- · Dental and oral medicine clinics
- · Patients with TMJ disorders
- · Traditional diagnostic imaging interpretation services (long-term)
Improved early detection and treatment planning for Temporomandibular Joint Osteoarthritis.
Increased adoption of AI-powered diagnostic tools across various medical specialties due to proven efficacy in complex tasks.
Revolutions in medical training, shifting focus from raw image interpretation to validating AI diagnoses and complex case management.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.AI