A Convolutional Neural Network Based Auto-Positioning Method For Dental Arch In Rotational Panoramic Radiography

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence

Background:

• Dental panoramic radiography (DPR) is a common diagnostic tool.
• Patient positioning is critical for DPR image quality but challenging due to anatomical variations and posture.
• Image blurring, particularly in the anterior dental arch, compromises diagnostic accuracy.

Purpose of the Study:

• To develop a novel method for estimating and correcting patient positioning errors in DPR.
• To reduce image blur and enhance the diagnostic quality of panoramic dental images.
• To improve the stability and reliability of reconstructed dental arch images.

Main Methods:

• A convolutional neural network (CNN) was employed to estimate dental arch positioning errors.
• The CNN-based method reconstructs panoramic images with corrected dental curvature.
• The approach aims to mitigate blurring caused by positioning inaccuracies.

Main Results:

• The proposed CNN method effectively estimates patient positioning errors.
• Reconstructed panoramic images exhibit reduced blur, especially in the anterior dental arch.
• The technique leads to more stable image quality for subsequent diagnostic interpretation.

Conclusions:

• The novel CNN-based method offers a promising solution for improving DPR image quality.
• Accurate estimation and correction of positioning errors enhance diagnostic reliability.
• This approach has the potential to benefit dental diagnosis by providing clearer panoramic images.