Home // PATTERNS 2023, The Fifteenth International Conference on Pervasive Patterns and Applications // View article

Automatic Teeth Segmentation From Panoramic X-ray Images Using Deep Learning Models

Authors:
Shuaa S. Alharbi

Keywords: Convolutional Neural Networks, Deep learning, Deep Neural Networks, Image Segmentation, Medical Image Processing, Semantic Segmentation.

Abstract:
A dentist’s primary objective when screening for X-ray problems is to determine the shape, number, and position of teeth. Computational tools have been proposed to aid specialists in making more accurate diagnoses rather than relying solely on the trained eyes of dentists. Teeth segmentation and object detection are the core functions of these tools when applied to X-ray images. Segmenting and detecting the teeth in images is actually the first step in enabling other automatic processing methods. Medical image segmentation, especially in dentistry field, has been transformed by Deep Learning (DL) in recent years. U-Net with its different extensions and modifications has been among the most popular deep networks developed for medical image segmentation. However, it is difficult to determine which one will work best for teeth segmentation. In this study, different semantic segmentation models are selected based on their common use in medical image segmentation. Models include:UNet++, ResU-Net++ and MultiResU-Net. Using panoramic X-ray dataset, MultiResUNet architecture performed better than the other segmentation models with an accuracy of 97.16%.

Pages: 24 to 29

Copyright: Copyright (c) IARIA, 2023

Publication date: June 26, 2023

Published in: conference

ISSN: 2308-3557

ISBN: 978-1-68558-049-0

Location: Nice, France

Dates: from June 26, 2023 to June 30, 2023