Key Information of Experimental Datasets for Cross-Modal Medical AI

Chong Liu published a 5.5 KB Excel file on figshare in April 2026. The dataset likely contains key information supporting the evaluation of a cross-modal fusion AI model for medical diagnosis. The model was tested on COVID-19 radiography, skin cancer, and chest X-ray datasets, achieving reported accuracy improvements.

Use Cases

• Benchmarking cross-modal fusion models based on reported accuracy metrics for COVID-19, skin cancer, and chest X-ray diagnosis
• Analyzing model efficiency trade-offs based on reported reductions in parameters and computational complexity
• Studying feature alignment and redundancy pruning techniques in medical AI based on the described ViT-B/16 and SmartTrim modules

Strengths

• The supporting research reports specific model performance improvements of 3.1% to 4.1% over baselines on three medical datasets
• The model evaluation includes concrete efficiency metrics: a 44.2% parameter reduction and over 43% decrease in computational complexity

Limitations

• Row count is unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download