Pixel-wise Reflectance Spectra of Ex Vivo Porcine Tissues via Hyperspectral Imaging (500–5000 nm)

This dataset contains pixel-wise reflectance spectra (500-5000 nm) extracted from annotated hyperspectral (HS) images of ex vivo porcine tissues. The dataset was generated to support the study “Broadband Spectral Characterization of Ex Vivo Porcine Tissues via Hyperspectral Imaging across 500–5000 nm.” The data can be used to investigate tissue-specific spectral signatures over a wide wavelength range. Note that full hyperspectral cubes are not included in this dataset, but may be shared upon request. The dataset also includes curated visual and spectral summaries designed to facilitate data exploration, comparison across systems, and reproducibility of the reported analyses. Shared material organization The released material is organized in two parts: 1. Data and metadata By data, we mean the reflectance spectra extracted from annotated Regions of Interest (ROIs) of hyperspectral images. HS images were acquired using three complementary systems: visble near-infrared (VNIR, 500-1000 nm), short-wave infrared (SWIR, 1000-2500 nm), mid-wave infrared (MWIR, 2500-5000nm). Note that only pixel-wise spectral information is shared, whereas spatial relationships between pixels are not retrievable with these data. Each observation is a vector, corresponding to a single pixel reflectance spectrum. The following elements are included: Spectral reflectance data Wavelength reference vectors Metadata associated with each observation in the spectral reflectance data Data are organized by spectral range: VNIR SWIR MWIR 2. Data visual inspection Supplementary material is provided to facilitate qualitative exploration of the dataset. For each analysed tissue sample, PDF reports are provided. These contain: Representative reflectance images (central wavelength slice of each HS cube) Annotated ROIs used for spectral data extraction ROI-averaged reflectance spectra ROI-averaged first-derivative reflectance spectra Complementary RGB images are also provided as visual references for the samples at three steps of the sequential HS acquisition session: at the start of VNIR acquisition, and at the start/end of SWIR/MWIR acquisition. Note: some samples may lack one or more reference images. Supplementary material is grouped according to the following structure: 7 tissue classes: brain, adipose tissue (subcutaneous, ventral abdominal wall), myocardium, kidney, liver, skin, and spleen 5 specimens per tissue class (e.g., Organ 1–5) 2 samples per specimen (c1, c2) Data type: Images or Spectra 3 spectral ranges: VNIR, SWIR, MWIR 3 rotations: approximately 0°, 45°, and 90° (rotation around the optical axis) 2 repetitions: two consecutive acquisitions per rotation For Images PDF files, visual representations of the samples are displayed at a representative wavelength. For Spectra PDF files, reflectance spectra and their first derivatives are reported. Further details about the dataset structure and contents are provided in the accompanying README file. This work was supported by the research project "HyperSIGHT" (No. R18SF4YHHS) funded by the Italian Ministry of Education and University; the European Union - NextGenerationEU, M4C2 I1.1, Project PRIN 2022 ”ARTIS”, Prot. 20228H4ZR3, CUP D53D23001080006; the European Research Council (ERC) through Horizon Europe Framework Programme for Research and Innovation under Grant 10121.