Automatic segmenting teeth in X-ray images: Trends, a novel data set, benchmarking and future perspectives

Gil Jader,Luciano Oliveira,Matheus Pithon

DOI: https://doi.org/10.48550/arXiv.1802.03086

2018-02-09

Abstract:This review presents an in-depth study of the literature on segmentation methods applied in dental imaging. Ten segmentation methods were studied and categorized according to the type of the segmentation method (region-based, threshold-based, cluster-based, boundary-based or watershed-based), type of X-ray images used (intra-oral or extra-oral) and characteristics of the dataset used to evaluate the methods in the state-of-the-art works. We found that the literature has primarily focused on threshold-based segmentation methods (54%). 80% of the reviewed papers have used intra-oral X-ray images in their experiments, demonstrating preference to perform segmentation on images of already isolated parts of the teeth, rather than using extra-oral X-rays, which show tooth structure of the mouth and bones of the face. To fill a scientific gap in the field, a novel data set based on extra-oral X-ray images are proposed here. A statistical comparison of the results found with the 10 image segmentation methods over our proposed data set comprised of 1,500 images is also carried out, providing a more comprehensive source of performance assessment. Discussion on limitations of the methods conceived over the past year as well as future perspectives on exploiting learning-based segmentation methods to improve performance are also provided.

Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The main problems that this paper attempts to solve are the analysis of the current situation in the field of dental X - ray image segmentation and the exploration of future research directions. Specifically, the authors made attempts in the following aspects:

1. **Literature Review**: Conduct an in - depth literature review of the existing dental X - ray image segmentation methods, classify and analyze the types of these methods (region - based, threshold - based, clustering - based, boundary - based or watershed - based), the types of X - ray images used (intra - oral or extra - oral), and the characteristics of the data sets used to evaluate these methods.

2. **Proposing a Data Set**: In order to fill the scientific gaps in existing research, the authors proposed a new data set based on extra - oral X - ray images, which contains 1,500 annotated panoramic X - ray images. This data set aims to provide more image variations and diversities to more comprehensively evaluate the performance of segmentation methods.

3. **Method Evaluation**: Conduct a statistical comparison of 10 image segmentation methods on the newly proposed 1,500 - image data set, providing a more comprehensive source of performance evaluation.

4. **Limitations and Future Prospects**: Discuss the limitations of the methods proposed in the past year and explore future research directions for using learning - based segmentation methods to improve performance.

Through these efforts, the authors hope to promote the development of dental X - ray image segmentation technology, enabling it to more accurately identify and isolate teeth, thereby assisting in automated dental diagnosis and treatment plan formulation.