A Hybrid Deep Learning Framework for Automated Dental Disorder Diagnosis from X-Ray Images

Area of Science:

• Artificial Intelligence
• Medical Imaging
• Computer Vision

Background:

• Dental disorders like cavities and periodontal disease are significant global health concerns.
• Current diagnostic methods, relying on manual X-ray interpretation, are subjective and time-consuming.
• Early detection is crucial for effective treatment, cost reduction, and preventing systemic complications.

Purpose of the Study:

• To develop an automated deep learning framework for precise dental disorder diagnosis from X-ray images.
• To improve diagnostic accuracy and accessibility, especially in resource-limited settings.
• To leverage hybrid feature fusion and sequence-based classification for enhanced performance.

Main Methods:

• A hybrid feature-fusion framework combining handcrafted Histogram of Oriented Gradients (HOG) features with DenseNet-201 and Swin Transformer deep learning models.
• Utilizing a Long Short-Term Memory (LSTM) classifier to learn sequential dependencies from fused features.
• Evaluating the framework on the Dental Radiography Analysis and Diagnosis (DRAD) dataset after preprocessing including CLAHE enhancement.

Main Results:

• The proposed LSTM-based hybrid model achieved high performance metrics: 96.47% accuracy, 91.76% specificity, 94.92% precision, 91.76% recall, and 93.14% F1-score.
• Demonstrated the effectiveness of integrating handcrafted and deep learning features for dental image analysis.
• The model's performance indicates significant potential for automated dental diagnostics.

Conclusions:

• The developed framework offers a flexible, interpretable, and high-performing solution for image-based recognition tasks in dentistry.
• It serves as a reproducible model for future research in hybrid feature fusion and sequence-based classification.
• This AI-driven approach has the potential to revolutionize dental diagnostics, improving efficiency and patient care.