Automated Detection and Classification of Oral Lesions Using Deep Learning to Detect Oral Potentially Malignant...

Area of Science:

• Oncology
• Computer Science
• Medical Imaging

Background:

• Oral cancer is a leading cause of cancer deaths globally, with late diagnosis significantly reducing survival rates.
• Early detection of oral potentially malignant disorders (OPMDs) is crucial for improving patient outcomes.
• Current screening methods are limited by public awareness and delayed specialist referrals.

Purpose of the Study:

• To explore computer vision applications for detecting OPMDs in photographic images.
• To investigate the potential of an automated system for early oral cancer detection.
• To develop a deep learning model for real-time identification and classification of oral lesions.

Main Methods:

• A two-stage deep learning model was developed for oral lesion detection and classification.
• The first stage employed a detector network to identify oral lesions.
• The second stage utilized a classifier network to categorize lesions as benign, OPMD, or carcinoma.

Main Results:

• Preliminary results show the feasibility of deep learning for automated oral lesion detection and classification.
• The proposed model demonstrated real-time processing capabilities.
• The system showed potential for accurate identification of OPMDs.

Conclusions:

• Deep learning-based approaches show promise for automated oral lesion analysis.
• The developed model can serve as a low-cost, non-invasive tool to aid oral cancer screening.
• This technology has the potential to improve the early detection of OPMDs and enhance oral cancer outcomes.