PANDIA: Multimodal AI System for Infant Pain Assessment

A dataset likely containing metadata or summary information related to the PANDIA AI system for infant pain assessment. The system was evaluated on 2,847 infants across four datasets and achieves 87.3% accuracy. The dataset was authored by Oussama El Othmani and last updated in May 2026.

Use Cases

• Benchmarking multimodal AI systems for infant pain assessment based on the described evaluation across four datasets.
• Studying privacy-preserving federated learning frameworks for clinical data based on the described multi-site collaboration.
• Analyzing clinician acceptance of AI explanations based on the reported 92.1% acceptance rate.
• Evaluating lightweight model architectures for edge deployment based on the system's fewer than 30M parameters.

Strengths

• System performance is quantified with specific metrics: 87.3% accuracy and 92.1% clinician acceptance rate.
• Evaluation scale is documented: 2,847 infants across four datasets.
• Model efficiency is specified: fewer than 30M parameters for edge deployment.
• All code, models, and pipelines are publicly available without restriction.

Limitations

• Dataset access is restricted; the NICU-MM dataset requires an ethical data use agreement.
• The validation design is retrospective, and dataset heterogeneity across collection sites is noted as a limitation.
• Row count and column-level documentation are unknown, limiting suitability assessment.

Provenance

Source

figshare

Freshness

Last updated 2026-05-26 17:32:02