AI Diagnostic Reliability for Dental Features from 306 Intraoral Photographs

Saif Abdelaziz's research dataset from 2026 evaluates a prototype AI software for detecting clinical features in dental photographs. The dataset likely contains diagnostic performance metrics for features like caries, gingival recession, calculus, retained roots, bleeding, and staining. It is based on a cross-sectional study of 34 patients at Ajman University's College of Dentistry.

Use Cases

• Benchmarking AI diagnostic performance based on sensitivity and specificity metrics for dental conditions.
• Analyzing trade-offs between sensitivity and positive predictive value for automated dental feature detection.
• Training or refining machine learning models for detecting gingival recession, calculus, or retained roots from intraoral images.

Strengths

• Provides specific diagnostic metrics (e.g., 71.43% sensitivity for bleeding, 98.77% specificity for staining) with 95% confidence intervals.
• Ground truth was established by clinical examiners, providing a reference for AI output evaluation.
• Dataset is openly available under a CC-BY-4.0 license, facilitating reuse and verification.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data is from a single study of 34 patients, which may limit generalizability.

Provenance

Source

figshare, authored by Saif Abdelaziz.

Collection Method

Cross-sectional study comparing AI software output against clinical examiner ground truth for 306 standardized intraoral photographs.

Time Range

Study date is not specified; dataset was last updated in 2026.

Freshness

Last updated 2026-05-08 05:51:45; freshness should be verified.

Geography

Data was collected at Ajman University's College of Dentistry; specific geographic scope is not detailed.