Related Experiment Video
Updated: May 9, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.6K
A Deep Learning Approach for Mandibular Condyle Segmentation on Ultrasonography.
Gaye Keser1, Hakan Yülek2, Ayşe Gül Öner Talmaç3
1Department of Oral Diagnosis and Radiology, Faculty of Dentistry, Marmara University, Başıbüyük Sağlık Yerleşkesi Başıbüyük Yolu 9/3, 34854, Maltepe, Istanbul, Turkey. gayekeser@hotmail.com.
Journal of Imaging Informatics in Medicine
|May 6, 2025
Summary
This study developed an AI model for segmenting mandibular condyles in ultrasound images. The deep learning approach achieved high accuracy, offering a faster diagnostic tool for medical specialists.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oral and Maxillofacial Radiology
Background:
- Deep learning shows promise in medical image processing, particularly for segmentation tasks.
- Accurate segmentation of the mandibular condyle is crucial for diagnosis in oral and maxillofacial radiology.
Purpose of the Study:
- To develop and evaluate a computer-based diagnostic software for mandibular condyle segmentation in ultrasound images.
- To assess the performance of the YOLOv8 deep learning model for this specific application.
Main Methods:
- Retrospective analysis of 668 adult mandibular condyle ultrasound images.
- Annotation using the CranioCatch labeling program, validated by experts.
- Segmentation performed using the YOLOv8 deep learning artificial intelligence (AI) model.
Main Results:
- The YOLOv8 model achieved an F1 score of 0.93.
- Sensitivity was 0.90, and precision was 0.96 for image segmentation.
- The AI model successfully detected and segmented mandibular condyles in test images.
Conclusions:
- Automatic segmentation of the mandibular condyle from ultrasound images using AI is a viable and promising application.
- This AI-driven approach can significantly reduce diagnostic time for surgeons and radiologists.
- The developed software demonstrates potential for improving efficiency in medical diagnostics.
