Deep learning for dental implant planning using CBCT
A deep learning approach for dental implant planning in cone-beam computed tomography images
Authors: Sevda Kurt Bayrakdar, Kaan Orhan, Ibrahim Sevki Bayrakdar, Elif Bilgir, Matvey Ezhov, Maxim Gusarev, Eugene Shumilov
Summary:
This study assessed the diagnostic performance of a deep convolutional neural network (Diagnocat) for implant planning using CBCT images. Manual and AI measurements of bone height and thickness were compared. The AI achieved strong agreement for bone height and canal detection but showed lower accuracy in bone thickness evaluation. Findings demonstrate that AI can facilitate implant planning workflows, supporting clinicians in radiographic interpretation.
Check the original publication:
Key Words: Artificial intelligence, Deep learning, Cone-beam computed tomography, Implant planning, Dental implant, DCNN, Diagnocat
Extracted Data
Year: 2021
Modality: Cone-Beam Computed Tomography (CBCT)
Dataset: 75 CBCT images (508 regions)
Dataset Split: Not specified
Network Architecture: 3D Fully Convolutional U-Net (Diagnocat)
Metrics: ICC (0.995–0.996), Bland–Altman, Wilcoxon test
AP – Professional Qty: 1
AP – Supervisor Presence: Noinformation
AP – Experience Level: 8 years
AP – Expertise Area: Oral and Maxillofacial Radiology
AP – Tool or System: InvivoDental 6.0, Diagnocat AI
ML Task: Segmentation
Project Objective: Evaluate AI reliability and accuracy in implant planning from CBCT images
Clinical Relevance
AI-based systems can enhance precision and efficiency in dental implant planning, acting as a supportive mechanism for clinicians. Continuous refinement of bone measurement algorithms is required for widespread clinical use.
Watch full video on YouTube
