Porcine Wound Forecasting Processed Data

This dataset contains 1,650 processed porcine wound images at 256 × 256 pixels, organized into 72 longitudinal wound sequences from six pigs. It was created for the development and evaluation of generative models for longitudinal wound forecasting. The release includes the processed images and a manifest describing their organization by pig, wound, observation day, and within-day imaging burst. For the associated Wound Forecasting study, pigs 1323, 1324, 1326, and 1327 formed the model-development cohort, while pigs 1325 and 1328 were reserved as held-out test subjects. Partitioning was performed at the pig level to prevent subject leakage. The images were derived from the publicly available longitudinal porcine wound-healing dataset published by Zlobina et al and deposited in Dryad (https://doi.org/10.5061/dryad.0rxwdbsbr). The original dataset remains the authoritative source. Processing details, file organization, attribution, and links to the maintained Hugging Face dataset and project code are provided in the accompanying README. This is a preclinical porcine research dataset and has not been validated for clinical decision-making or human wound prognosis. If you use these processed data, please cite this Zenodo record and the original source: Zlobina K, Yang H, Kesapragada M, et al. A high-resolution temporal transcriptomic and imaging dataset of porcine wound healing. Scientific Data. 2025;12:1635. doi:10.1038/s41597-025-05921-w Original data repository: Zlobina K, et al. A high-resolution temporal transcriptomic and imaging dataset of porcine wound healing. Dryad. doi:10.5061/dryad.0rxwdbsbr