Welcome to our dataset repository. Here, you'll find a collection of high-quality datasets that we've carefully collected and annotated. Please cite our paper if you use our dataset.

Important Notice
  • Our datasets are available only for non-commercial and academic research purposes. Any form of commercial use is strictly prohibited.
  • Access to all datasets is granted through an application and review process. To request access, please complete the Data Access Application Form and email it to the first author of the corresponding paper and Dr. Zhiming Cui, while also cc your supervisor. Upon approval, you will receive the download link.

01 CBCT Dataset

This is the dataset for our 'A fully automatic AI system for tooth and alveolar bone segmentation from cone-beam CT images' paper in Nature Communication 2022. We released partial data (50 raw data of CBCT scans collected from dental clinics) to support the results in this study with permission from respective data centers. The full datasets are protected because of privacy issues and regulation policies in hospitals.

Download Link

Reference:

@article{cui2022fully,
title={A fully automatic AI system for tooth and alveolar bone segmentation from cone-beam CT images},
author={Cui, Zhiming and Fang, Yu and Mei, Lanzhuju and Zhang, Bojun and Yu, Bo and Liu, Jiameng and Jiang, Caiwen and Sun, Yuhang and Ma, Lei and Huang, Jiawei and others},
journal={Nature communications},
volume={13},
number={1},
pages={2096},
year={2022},
publisher={Nature Publishing Group UK London}
}

02 CephAdoAdu Dataset

We collected a new benchmark dataset, named CephAdoAdu, with both adolescent and adult cases, distinguishing it from existing datasets that solely consist of either adolescent or most adult cases. CephAdoAdu has a total of 1000 (500 adult cases, 500 adolescent cases) cephalometric X-ray images, acquired from eight clinical centers. Every cephalometric image underwent manual annotations to mark 10 typical landmarks, by an experienced dental radiologist with over ten years of expertise.

Download Link

Reference:

@inproceedings{wu2024cephalometric,
title={Cephalometric Landmark Detection across Ages with Prototypical Network},
author={Wu, Han and Wang, Chong and Mei, Lanzhuju and Yang, Tong and Zhu, Min and Shen, Dinggang and Cui, Zhiming},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
pages={155--165},
year={2024},
organization={Springer}
}

03 Tooth Alignment Dataset

We acquire 2,224 CBCT scans from Shanghai NinthPeople’s Hospital under consistent acquisition parameters (100 kVsource voltage, 0.3 mm voxel size, 468 × 468 × 250 resolution). Eachscan undergoes rigid registration to align it to a standardized jawcoordinate system for uniform orientation and field of view, and weapply stringent quality control to remove scans with metal or motionartifacts and incomplete dentition coverage. The resulting dataset comprises1,955 clinically validated high-quality sets of 3D tooth models.

Download Link

Reference:

@article{DOU2025103746,
title = {CLIK-Diffusion: Clinical Knowledge-informed Diffusion Model for Tooth Alignment},
journal = {Medical Image Analysis},
volume = {106},
pages = {103746},
year = {2025},
issn = {1361-8415},
doi = {https://doi.org/10.1016/j.media.2025.103746},
url = {https://www.sciencedirect.com/science/article/pii/S1361841525002932},
author = {Yulong Dou and Han Wu and Changjian Li and Chen Wang and Tong Yang and Min Zhu and Dinggang Shen and Zhiming Cui}
}

04 DVCT Dataset

We release DVCT, an open benchmark dataset consisting of 2,000 panoramic dental X-ray images with high-precision annotations of dental caries. The dataset is established as a gold standard, with annotations dual-verified using both panoramic X-rays and intra-oral images, and further cross-checked by experienced dental radiologists. DVCT covers subjects across different age groups and includes caries at various stages of dentition, providing a reliable resource for advancing automated dental analysis.

Download Link

Reference:

@InProceedings{LuoTao_Adapting_MICCAI2025,
        author = { Luo, Tao and Wu, Han and Yang, Tong and Shen, Dinggang and Cui, Zhiming},
        title = { { Adapting Foundation Model for Dental Caries Detection with Dual-View Co-Training } },
        booktitle = {proceedings of Medical Image Computing and Computer Assisted Intervention -- MICCAI 2025},
        year = {2025},
        publisher = {Springer Nature Switzerland},
        volume = {LNCS 15975},
        month = {September},
        page = {44 -- 53}
}