DentIRO: 5,300 Single-Tooth Intraoral Radiographs from Bangladesh

Bangladesh is the source for 5,300 tooth-centered intraoral radiographs collected from 3,243 patients at two dental clinics between 2025 and 2026. The dataset, created by Md. Mehedi Hasan Shoib and released on figshare, provides expert-labeled images categorized into four diagnostic classes: Healthy, Caries, Crowned, and Root Canal. It includes structured metadata on patient demographics and acquisition site to support machine learning research for automated dental diagnosis.

Use Cases

• Training multi-class image classifiers for dental conditions based on the four diagnostic labels (Healthy, Caries, Crowned, Root Canal).
• Developing patient-aware models for dental diagnosis based on the provided patient demographic metadata.
• Benchmarking machine learning methods for automated tooth-level diagnosis using the expert-labeled, tooth-centric radiographs.
• Studying multi-site generalization in medical imaging models based on data collected from two different dental clinics.

Strengths

• 5,300 expert-labeled images across four clinically relevant diagnostic classes.
• Data originates from 3,243 patients, providing a patient-aware foundation for analysis.
• Images were collected from two clinical sites, which may support studies of multi-site robustness.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Data is sourced exclusively from two clinics in Bangladesh, which may reflect geographic bias.
• The dataset size of 5,300 images may be limited for training very large, data-hungry models.

Provenance

Source

Dhaka Dental Care (Dhaka) and Ma Dental Care (Cumilla) in Bangladesh.

Collection Method

Retrospectively collected from routine clinical practice, with diagnostic labels assigned by experienced dental practitioners through independent review and consensus.

Time Range

2025 to 2026

Freshness

Last updated 2026-05-02 02:27:59; freshness should be verified.

Geography

Bangladesh