{"success":true,"data":{"author":"Rahnemoonfar, Chowdhury, Murphy (Lehigh University, Texas A&M); TACO repackaging by ISP, Universitat de Valencia","bucketAccessType":"public","bucketName":"isp-uv-es/rescuenet-taco","bucketPrefix":"","category":"Earth Observation & Geospatial","cloudProvider":"huggingface","createdAt":"2026-11-16T14:11:25.704Z","description":"RescueNet (Scientific Data, 2023) is a high-resolution UAV benchmark for post-disaster damage assessment, collected with small drones after Hurricane Michael struck the Florida panhandle in October 2018. It has 4,494 RGB frames of 4000x3000 or 4056x3040 pixels, each with a pixel-level mask in 11 classes that grade buildings by damage (from no damage to total destruction) and separate clear from blocked roads, alongside water, trees, vehicles, pools and background, plus a whole-frame neighbourhood damage level for scene classification. Splits are 3,595 train, 449 validation and 450 test. This TACO copy was built from the authors' Springer Nature figshare collection (6647354); JPEG frames are copied byte-for-byte and PNG masks re-encoded as tiled COGs. There is no georeferencing or flight altitude. About 23.7 GB, read with tacoreader. The original segmentation and classification sets are released under CC0 on figshare; cite Rahnemoonfar et al. 2023.","displayTitle":"RescueNet: UAV Semantic Segmentation of Hurricane Michael Damage (TACO)","doi":"10.1038/s41597-023-02799-4","id":"f98b2e04-0a35-52c1-8d88-f4822134630a","license":"CC0-1.0","licenseEvidenceUrl":"https://springernature.figshare.com/articles/dataset/RescueNet_Semantic_Segmentation_Train_Set/22825511","objectCountEstimate":13,"probeStatus":"ok","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"isp-uv-es/rescuenet-taco\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"isp-uv-es/rescuenet-taco\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=isp-uv-es, license:cc-by-4.0, size_categories:1K<n<10K, downloads=414"],"format":"mixed","generatedAt":"2026-11-16T14:11:32.641Z","layout":{"example":"assets/logos.png","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"89504e470d0a1a0a0000000d49484452000002220000008c0806000000413dc29d00009fbb49444154789cecfd77986c5775e78d7fd7dafb84ca9d6e0e0a57010524a1608288064cce2667f03018637bb0fd9b796dbf1ef0cc388dc7f1c7181b838d8d6dd2788c49c6649011c2422040a09ca51bfa76aaae74c2de6bbd7fecaa9baf7415407dafcee779eae9eeea3aa74eed3a55fb7bd65eebbb808a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8a8ac32002981944742fffff311f58454545454545c589cda1e2e34041424460e67b7c7c45454545454545c53d32110f3c1616fbfe1e8b8c288a70f1a3ce6da449b26f1b73800089e318173feadc461c47076d37112acc74d0f35454545454543c5ca866be7b8188a0aab0c6c0797fd8ffb76fdd6c5ff7ca175e7afa8e931e79f7ce3db77cfd9b575df5f52bbe3dbfd2ed69a35ec3a58fb970ee89975e72d1f66d9b4ebbe996db7ff8c10f7fe2df6ebdedaef2d0fd1863e0bddff77c1515151515150f072a2132e6a068842a1480310cef05975c785ef3a79fffcc9fbce5b63b6eb9f196db77de78cb6dabbbf72cb8c7fdc485532f7ff1339f17c751321c64fd24896b499ad4f7ec5db8637ecfddf3b3731b66b76cda78525114f928cf87f57ad274ce97fff8cf5ff8f417bf72f9c2bab9693e","encoding":"hex","key":"assets/logos.png","truncatedAt":512}],"sampleObjects":["assets/logos.png","assets/sample_0.png","assets/sample_1.png","assets/sample_2.png","assets/sample_3.png"],"sidecars":[{"bytes":"---\nlicense: cc-by-4.0\npretty_name: \"RescueNet\"\nsize_categories:\n- 1K<n<10K\ntask_categories:\n- image-classification\n- image-segmentation\ntags:\n- earth-observation\n- remote-sensing\n- taco\n---\n\n<p align=\"center\"><img src=\"assets/logos.png\" width=\"546\" alt=\"Image and Signal Processing Group, Universitat de València · ELLIOT\"></p>\n\n# RescueNet\n\n> **This is a repackaging, not a new dataset.** It is **RescueNet** by Rahnemoonfar, Chowdhury and Murphy (figshare collection 6647354), converted to TACO. Pixel values and labels are kept as released except where the description below says otherwise. All credit belongs to the original authors: if you use it, please cite them and follow their licence.\n>\n> [original dataset](https://doi.org/10.6084/m9.figshare.c.6647354) · [paper](https://doi.org/10.1038/s41597-023-02799-4) · licence: CC-BY-4.0\n>\n> Repackaged into TACO by the [Image and Signal Processing Group (ISP)](https://isp.uv.es/), Universitat de València, within the [ELLIOT](https://elliot-ai.eu) project.\n\n## Citation\n\nPlease cite the original work:\n\n```bibtex\n@article{rahnemoonfar2023rescuenet,\n  title   = {RescueNet: A High Resolution UAV Semantic Segmentation Dataset for Natural Disaster Damage Assessment},\n  author  = {Rahnemoonfar, Maryam and Chowdhury, Tashnim and Murphy, Robin},\n  journal = {Scientific Data},\n  volume  = {10},\n  number  = {1},\n  pages   = {913},\n  year    = {2023},\n  doi     = {10.1038/s41597-023-02799-4}\n}\n```\n\n## About the data\n\n4494 high-resolution UAV frames of post-Hurricane-Michael Florida, each carrying both a pixel-level 11-class damage map and a whole-frame neighbourhood damage level.\n\n**4,494 samples** · splits: `test` 450 · `train` 3,595 · `validation` 449 · tasks: `semantic-segmentation`, `scene-classification`\n\nPackaged as [TACO v3](https://asterisk.coop/taco/spec/).\n\n<details>\n<summary>Full description</summary>\n\n**Legend.** The segmentation legend grades buildings and roads by damage rather than by object class (Building_No_Damage through Building_Total_Destruction, Road-Clear against Road-Blocked), which is what makes it a damage-assessment benchmark. It has eleven entries although the release's own note says ten and then lists eleven; the enumeration is what the pixels hold, and a scan of all 4,494 masks found exactly the values 0..10 with every class occurring. `Background` (0) stays a class rather than an ignore label: nothing on disk distinguishes unannotated from none-of-the-other-ten.\n\n**One container.** The masks and the classification tables cover identical frame sets in all three splits, checked both ways, and two containers would mean two copies of 12-megapixel JPEGs.\n\n**Imagery.** Frames are JPEG and copied byte-for-byte, being lossy in with nothing to recover; masks are PNG and re-encoded as COGs at a measured 1.09x the PNG, which buys tiling on a 12 Mpx raster. Two frame sizes, 4000x3000 and 4056x3040, every mask header-checked against its own frame. UAV RGB with no geotransform and no published per-frame altitude: radiometry `render`, no CRS, no declared GSD.\n\n**Source.** The authors' figshare release (Springer Nature collection 6647354, three clean zips, each extracting to exactly its own entry count), not the ResQAI/RescueNet Hub mirror, whose archive is incomplete.\n\n**Splits.** Official, preserved: 3595 train, 449 val, 450 test.\n\n</details>\n\n## Getting started\n\n```bash\ngit clone --recursive https://github.com/OscarPellicer/taco\npip install -e \"taco/python[ml]\"     # builds the reader: C++23, CMake, Ninja, pkg-config, libcurl >= 7.83, OpenSSL >= 3\n```\n\nRead it straight from the Hub:\n\n```python\nfrom huggingface_hub import snapshot_download\nfrom taco.ml import Dataset, plot_sample\n\npath = snapshot_download(\"isp-uv-es/rescuenet-taco\", repo_type=\"dataset\")\nds = Dataset(path)                  # the repository is the container: parts + .tacocat\nplot_sample(ds[0])\n```\n\nor from a local copy:\n\n```python\nds = Dataset(\"rescuenet\")\nsample = ds[0]                      # {slot name: SlotValue}, arrays decoded\nsample[\"image\"].array.shape\n```\n\nMetadata without decoding anything:\n\n","encoding":"utf8","key":"README.md","name":"README.md","size":5270,"truncated":true}],"sourceObjectKey":"assets/logos.png","starterCell":"from huggingface_hub import hf_hub_download\npath = hf_hub_download(repo_id=\"isp-uv-es/rescuenet-taco\", filename=\"assets/logos.png\", repo_type=\"dataset\")\nprint(path)","starterSamplePath":"assets/logos.png","totalObjects":13},"summary":"RescueNet (Scientific Data, 2023) is a high-resolution UAV benchmark for post-disaster damage assessment, collected with small drones after Hurricane Michael struck the Florida…","title":"RescueNet: UAV Semantic Segmentation of Hurricane Michael Damage (TACO)","updatedAt":"2026-11-16T15:47:24.885Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/isp-uv-es/rescuenet-taco"}}