Medical Imaging Dataset for Survival Prediction with Optimal Transport

OTSurv is a dataset for survival prediction from medical images using a multiple instance learning framework. The dataset was created by researchers from Stony Brook University and the University of Florida and was released in June 2025.

Use Cases

• Predict patient survival time from medical image patches using a multiple instance learning framework.
• Apply optimal transport to model heterogeneity in image features for survival risk stratification.
• Benchmark deep learning models for survival analysis against the proposed OTSurv methodology.

Strengths

• Dataset is associated with a peer-reviewed MICCAI 2025 publication.
• Framework addresses feature heterogeneity via optimal transport, a specific methodological strength.

Limitations

• Specific dataset size, row count, and column details are not provided.
• The dataset appears to be relatively new, with limited independent validation or benchmarking.

Provenance

Source

Y-Research-Group on Hugging Face.

Collection Method

Created for the research paper 'OTSurv: A Novel Multiple Instance Learning Framework for Survival Prediction with Heterogeneity-aware Optimal Transport'.

Time Range

null

Freshness

Last updated on June 30, 2025.

Geography

null