Transfer Deep Learning for Dental and Maxillofacial Imaging Modality Classification: A Preliminary Study.

Area of Science:

• Dentistry
• Artificial Intelligence
• Medical Imaging

Background:

• Accurate classification of dental X-ray modalities is crucial for diagnosis.
• Small datasets pose a challenge for traditional machine learning models.

Purpose of the Study:

• To apply transfer deep learning for automatic classification of dental X-ray modalities.
• To evaluate the effectiveness of pre-trained convolutional neural networks (CNNs) on a small dataset.

Main Methods:

• Utilized VGG16, NASNetLarge, and Xception CNN architectures pre-trained on ImageNet.
• Trained models on a dataset of 496 panoramic and cephalometric X-ray images.
• Employed NVIDIA GeForce GTX 1080 Ti GPU for model training.

Main Results:

• Achieved 100% accuracy in classifying X-ray modalities on a test set of 124 images.
• Models demonstrated high performance despite the small, un-preprocessed test dataset.
• Other statistical metrics were deemed irrelevant due to near-perfect accuracy.

Conclusions:

• Transfer deep learning is highly effective for automatic classification of dental X-ray modalities.
• This approach shows significant potential to aid dentists in daily practice and enhance diagnostic quality.
• Further research into modality and sub-modality classification can further refine applications.