2026-10-07T08:47:29Z https://api.figshare.com/v2/oai

oai:figshare.com:article/34003191 2026-09-26T21:24:30Z category_24781 category_29161 category_26275 item_type_3 month_year_09_2026

Tumor Tissues FLIM Image Segmentation Dataset: Human Tumors and Xenograft Mouse Models. Garance Boesinger (25114629) Victoria Fay (22366753) Cancer diagnosis Deep learning Biomedical imaging head and neck cancer treatment Tumor Margin Delineation image segmentation-deep learning Fluorescence Lifetime Imaging Microscopy-FLIM label-free imaging method digital pathology solutions Metabolic imaging <h2 dir="ltr">Dataset Description</h2><p dir="ltr">This dataset contains the data associated with the study <b>"The role of spatial resolution in cellular-scale </b><b>fluorescence lifetime imaging deep-learning </b><b>segmentation of head and neck cancer </b><b>cryosections"</b>, which investigates deep learning-based segmentation of FLIM tumor tissue images using different neural network architectures.</p><p dir="ltr">The dataset includes imaging data from both <b>human tumor samples</b> and <b>xenograft mouse models</b>, together with the corresponding segmentation data and reconstructed images used in the study.</p><h3 dir="ltr">Dataset contents</h3><p dir="ltr">The dataset is organized into the following main categories:</p><ul><li><b>Human tumor data (\human):</b> imaging data and corresponding segmentation data obtained from human tumor samples, each tumor in a separate file.</li><li><b>Xenograft mouse model data (\xenograft):</b> imaging data and corresponding segmentation data obtained from xenograft mouse models, all tumors in the same file.</li><li><b>Reconstructed images (\human_reconstructions; \xenograft_reconstructions) and histopathological stainings (H&E and E-cadherin; \human_histopathological_stainings_and_masks):</b> reconstructed image data and visualization outputs generated as part of the analyses presented in the study.</li><li><b>Metadata:</b> information describing the individual samples, data types, dataset splits, and relevant technical characteristics.</li></ul><p dir="ltr">The numerical data are provided in NumPy (<code>.npy</code>) format. Detailed information regarding the structure, dimensions, data types, channel ordering, normalization, and interpretation of the individual arrays is provided in the accompanying <code>DATA_DESCRIPTION.md</code> file.</p><h3 dir="ltr">Data organization</h3><p dir="ltr">The dataset distinguishes between human tumor samples and xenograft mouse model samples.</p><p dir="ltr">Each data file is associated with a unique sample identifier. The accompanying metadata provide the correspondence between samples, image data, segmentation masks, and other derived data products.</p><h3 dir="ltr">Data format</h3><p dir="ltr">The primary numerical datasets are provided as NumPy arrays (<code>.npy</code>). These files can be loaded using standard Python/NumPy functionality. The exact structure and interpretation of each array are described in the accompanying data documentation.</p><h3 dir="ltr">Relation to the associated code</h3><p dir="ltr">The source code used for data preprocessing, model training, inference, segmentation, reconstruction, and evaluation is available in the associated GitHub repository:<b> </b>https://github.com/GaranceBo/flim-head-neck-cancer-segmentation.git. An archived version of the source code is available through Zenodo: https://doi.org/10.5281/zenodo.22983044.</p><h3 dir="ltr">Relation to the associated publication</h3><p dir="ltr">This dataset supports the results presented in:<b> </b><b>The role of spatial resolution in cellular-scale </b><b>fluorescence lifetime imaging deep-learning </b><b>segmentation of head and neck cancer </b><b>cryosections</b></p><h3 dir="ltr">Reuse</h3><p dir="ltr">These data are provided to facilitate reproducibility of the published results and to enable further research on tumor tissue image analysis and segmentation using machine learning and deep learning approaches. Users are encouraged to cite both this dataset and the associated publication when reusing these data.</p><h3 dir="ltr">License</h3><p dir="ltr">This dataset is distributed under the <b>[LICENSE, e.g. Creative Commons Attribution 4.0 International (CC BY 4.0)]</b> license.</p><h3 dir="ltr">Version</h3><p dir="ltr">Dataset version: <b>1.0</b></p><p dir="ltr">This version corresponds to the data used for the analyses reported in the associated publication.</p> 2026-09-26T21:24:30Z Dataset Dataset 10.6084/m9.figshare.34003191.v2 https://figshare.com/articles/dataset/Tumor_Tissues_FLIM_Image_Segmentation_Dataset_Human_Tumors_and_Xenograft_Mouse_Models_/34003191 CC BY 4.0