Related Experiment Video
Updated: Oct 1, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.0K
Trainable WEKA (Waikato Environment for Knowledge Analysis) Segmentation Tool: Machine-Learning-Enabled Segmentation
Nitin Kanuri1, Ahmed Z Abdelkarim2, Sonali A Rathore3
1Oral Diagnostic Sciences, Mills E. Godwin High School, Richmond, USA.
Cureus
|March 7, 2022
Summary
Machine learning automates dental radiograph segmentation, simplifying diagnosis. Trainable WEKA Segmentation (TWS) shows promise, especially for images with more teeth, improving diagnostic efficiency.
Area of Science:
- Oral and Maxillofacial Radiology
- Medical Image Analysis
- Machine Learning in Healthcare
Background:
- Dental radiograph segmentation is crucial for accurate oral diagnosis.
- Automating this process simplifies the work of oral and maxillofacial radiologists.
- Current methods often require manual delineation of anatomical structures.
Purpose of the Study:
- To analyze the performance of Trainable WEKA Segmentation (TWS) for segmenting panoramic dental radiographs.
- To assess the potential for large-scale implementation of automated segmentation to aid radiological diagnosis.
- To establish new benchmarks in oral imaging segmentation.
Main Methods:
- Utilized open-source panoramic radiographs from the UFBA UESC DENTAL IMAGES dataset.
- Simulated realistic imaging conditions by applying Gaussian noise to degrade radiographs.
- Quantified segmentation accuracy using MorphoLibJ, comparing automated results with dentist annotations and expert review.
Main Results:
- TWS achieved a Dice value of 0.66 for radiographs with >= 32 teeth and an F1 score of 0.59 for those with < 32 teeth.
- Statistically significant results were observed, with higher Intersection over Union and F1 scores for radiographs with more teeth due to better tooth alignment.
- Calculated t-values of 2.78 for Jaccard index and 2.81 for Dice score indicate significant findings.
Conclusions:
- Machine learning, specifically TWS, offers a viable alternative to manual segmentation of dental radiographs.
- Automated segmentation can enhance diagnostic efficiency and accuracy in oral radiology.
- Further research can optimize TWS for broader clinical application.
