Related Experiment Video
Updated: Jun 6, 2025
05:56
Objectification of Tongue Diagnosis in Traditional Medicine, Data Analysis, and Study Application
Published on: April 14, 2023
2.4K
Research and Application of Deep Learning Models with Multi-Scale Feature Fusion for Lesion Segmentation in Oral
Rui Zhang1,2,3, Miao Lu4, Jiayuan Zhang4
1Zhejiang Provincial Key Laboratory of Internet Multimedia Technology, College of Biomedical Engineering & Instrument Science, Zhejiang University, Hangzhou 310027, China.
Bioengineering (Basel, Switzerland)
|November 27, 2024
Summary
This study introduces an AI diagnostic tool using SegFormer semantic segmentation for precise oral mucosal disease detection. The AI model accurately segments lesions in images, aiding clinical diagnosis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Vision
Background:
- Diagnosing oral mucosal diseases is complex, with traditional methods lacking precision.
- Automated segmentation of lesions in oral mucosal disease images is crucial for objective diagnosis.
Purpose of the Study:
- To develop a high-accuracy AI-assisted diagnostic approach for oral mucosal diseases.
- To automatically segment lesion areas in white-light images using the SegFormer model.
Main Methods:
- A dataset of 838 oral mucosal disease images (oral lichen planus, leukoplakia, submucous fibrosis) was curated and pixel-level annotated.
- A SegFormer model was designed and compared against UNet and DeepLabV3 using cross-validation.
- Performance was evaluated using Dice coefficient and mean Intersection over Union (mIoU).
Main Results:
- The SegFormer-B2 model achieved the best performance with a Dice coefficient of 0.710 and mIoU of 0.786.
- SegFormer significantly outperformed traditional segmentation models like UNet and DeepLabV3.
- Visual analysis confirmed accurate segmentation of lesion areas for three common oral mucosal diseases.
Conclusions:
- The SegFormer model provides precise automatic segmentation for common oral mucosal diseases.
- This AI approach offers a reliable auxiliary tool to enhance clinical diagnosis efficiency and accuracy.
- The study highlights the potential clinical application value of AI in oral medicine.
