{"url":"/dataset/ai-tod","name":"AI-TOD","full_name":"Tiny Object Detection in Aerial Images","description_markdown":"AI-TOD comes with 700,621 object instances for eight categories across 28,036 aerial images. Compared to existing object detection datasets in aerial images, the mean size of objects in AI-TOD is about 12.8 pixels, which is much smaller than others.","description_withheld":null,"homepage":"https://github.com/jwwangchn/AI-TOD","introduced_date":"2021-01-10","introduced_date_note":null,"introduced_by":{"paper":"/paper/tiny-object-detection-in-aerial-images","title":"Tiny Object Detection in Aerial Images","first_author":"Jinwang Wang","url":null},"license":null,"modalities":[{"name":"Images","url":"/datasets/modality/images"}],"tasks":[{"name":"Object Detection","url":"/task/object-detection","datasets_with_task":"/datasets/task/object-detection"},{"name":"Semantic Segmentation","url":"/task/semantic-segmentation","datasets_with_task":"/datasets/task/semantic-segmentation"}],"languages":[],"variants":["AI-TOD"],"data_loaders":[],"num_papers_in_archive":18,"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28"},"benchmarks":[{"leaderboard":"/sota/object-detection-on-ai-tod","task":"Object Detection","dataset_variant":"AI-TOD","rows":7,"metrics":["AP","AP50","AP75","APvt","APt","APs","APm","mAP50","mAP@50-95"],"first_row_in_archive_order":{"model":"DNTR","paper":"/paper/a-denoising-fpn-with-transformer-r-cnn-for","metrics":{"AP":"26.2","AP50":"56.7","AP75":"20.2","APm":"37.0","APs":"31.0","APt":"26.4","APvt":"12.8"},"code_links":[{"title":"hoiliu-0801/dq-detr","url":"https://github.com/hoiliu-0801/dq-detr"},{"title":"hoiliu-0801/dntr","url":"https://github.com/hoiliu-0801/dntr"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"},{"leaderboard":"/sota/semantic-segmentation-on-ai-tod","task":"Semantic Segmentation","dataset_variant":"AI-TOD","rows":2,"metrics":["Dice"],"first_row_in_archive_order":{"model":"Unet++(ResNet-50)","paper":"/paper/unet-a-nested-u-net-architecture-for-medical","metrics":{"Dice":"70.19"},"code_links":[{"title":"qubvel/segmentation_models.pytorch","url":"https://github.com/qubvel/segmentation_models.pytorch"},{"title":"PaddlePaddle/PaddleSeg","url":"https://github.com/PaddlePaddle/PaddleSeg"},{"title":"MrGiovanni/Nested-UNet","url":"https://github.com/MrGiovanni/Nested-UNet"},{"title":"MrGiovanni/UNetPlusPlus","url":"https://github.com/MrGiovanni/UNetPlusPlus"},{"title":"bigmb/Unet-Segmentation-Pytorch-Nest-of-Unets","url":"https://github.com/bigmb/Unet-Segmentation-Pytorch-Nest-of-Unets"},{"title":"4uiiurz1/pytorch-nested-unet","url":"https://github.com/4uiiurz1/pytorch-nested-unet"},{"title":"yingkaisha/keras-unet-collection","url":"https://github.com/yingkaisha/keras-unet-collection"},{"title":"frgfm/Holocron","url":"https://github.com/frgfm/Holocron"},{"title":"1044197988/TF.Keras-Commonly-used-models","url":"https://github.com/1044197988/TF.Keras-Commonly-used-models"},{"title":"AlphaJia/keras_unet_plus_plus","url":"https://github.com/AlphaJia/keras_unet_plus_plus"},{"title":"Burf/tfdetection","url":"https://github.com/Burf/tfdetection"},{"title":"marccoru/marinedebrisdetector","url":"https://github.com/marccoru/marinedebrisdetector"},{"title":"CarryHJR/Nested-UNet","url":"https://github.com/CarryHJR/Nested-UNet"},{"title":"sauravmishra1710/UNet-Plus-Plus---Brain-Tumor-Segmentation","url":"https://github.com/sauravmishra1710/UNet-Plus-Plus---Brain-Tumor-Segmentation"},{"title":"amirfaraji/LowDoseCTPytorch","url":"https://github.com/amirfaraji/LowDoseCTPytorch"},{"title":"mrgiovanni/dissertation","url":"https://github.com/mrgiovanni/dissertation"},{"title":"gallegi/T4E_MICCAI_BrainTumor","url":"https://github.com/gallegi/T4E_MICCAI_BrainTumor"},{"title":"tuvovan/Unet-with-EfficientnetB7-Backbone","url":"https://github.com/tuvovan/Unet-with-EfficientnetB7-Backbone"},{"title":"aqbewtra/Multi-Class-Aerial-Segmentation","url":"https://github.com/aqbewtra/Multi-Class-Aerial-Segmentation"},{"title":"16xccheng/keras-unet","url":"https://github.com/16xccheng/keras-unet"},{"title":"nnzzll/networks","url":"https://github.com/nnzzll/networks"},{"title":"Tu-kun/cultivated_landdivision","url":"https://github.com/Tu-kun/cultivated_landdivision"},{"title":"2023-MindSpore-1/ms-code-217","url":"https://github.com/2023-MindSpore-1/ms-code-217/tree/main/UNet3%2B"},{"title":"mfp0610/semantic-segmentaion","url":"https://github.com/mfp0610/semantic-segmentaion"},{"title":"ShawFromAttock/FYP-Medical-Image-Segmentation-using-Nested-UNet-Architecture","url":"https://github.com/ShawFromAttock/FYP-Medical-Image-Segmentation-using-Nested-UNet-Architecture"},{"title":"TheUser0815/unetpp-pytorch","url":"https://github.com/TheUser0815/unetpp-pytorch"},{"title":"2023-MindSpore-4/Code10","url":"https://github.com/2023-MindSpore-4/Code10/tree/main/Unet"},{"title":"2023-MindSpore-1/ms-code-7","url":"https://github.com/2023-MindSpore-1/ms-code-7/tree/main/UNet3%2B"},{"title":"yangyucheng000/MSpaper","url":"https://github.com/yangyucheng000/MSpaper/tree/main/UNet3%2B"},{"title":"JoelPendleton/Edge-Detection-CNN","url":"https://github.com/JoelPendleton/Edge-Detection-CNN"},{"title":"MS-Mind/MS-Code-08","url":"https://github.com/MS-Mind/MS-Code-08/tree/main/MIMO-UNet"},{"title":"MindSpore-paper-code-2/code3","url":"https://github.com/MindSpore-paper-code-2/code3/tree/main/UNet3%2B"},{"title":"JoelPendleton/Object-Contour-Detection-CNN","url":"https://github.com/JoelPendleton/Object-Contour-Detection-CNN"},{"title":"sjosias/Tsetse-Flies","url":"https://github.com/sjosias/Tsetse-Flies"}]},"note":"rows are the archive's own order at snapshot; nothing here re-ranks them"}],"papers_with_a_benchmark_row":[{"paper":"/paper/a-denoising-fpn-with-transformer-r-cnn-for","title":"A DeNoising FPN With Transformer R-CNN for Tiny Object Detection","date":"2024-06-09","rows_on_this_dataset":1,"code_links":2,"syntology":null},{"paper":"/paper/inner-iou-more-effective-intersection-over","title":"Inner-IoU: More Effective Intersection over Union Loss with Auxiliary Bounding Box","date":"2023-11-06","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/rfla-gaussian-receptive-field-based-label","title":"RFLA: Gaussian Receptive Field based Label Assignment for Tiny Object Detection","date":"2022-08-18","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/a-normalized-gaussian-wasserstein-distance","title":"A Normalized Gaussian Wasserstein Distance for Tiny Object Detection","date":"2021-10-26","rows_on_this_dataset":1,"code_links":3,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":2,"samples_unverified":0,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/tiny-object-detection-in-aerial-images","title":"Tiny Object Detection in Aerial Images","date":"2021-01-10","rows_on_this_dataset":1,"code_links":1,"syntology":null},{"paper":"/paper/detectors-detecting-objects-with-recursive-1","title":"DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution","date":"2020-06-03","rows_on_this_dataset":1,"code_links":6,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":4,"samples_ran":0,"samples_unverified":4,"pointer_only_for_licence":0,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/unet-a-nested-u-net-architecture-for-medical","title":"UNet++: A Nested U-Net Architecture for Medical Image Segmentation","date":"2018-07-18","rows_on_this_dataset":1,"code_links":34,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":28,"samples_ran":5,"samples_unverified":23,"pointer_only_for_licence":2,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/encoder-decoder-with-atrous-separable","title":"Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation","date":"2018-02-07","rows_on_this_dataset":1,"code_links":78,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":72,"samples_ran":43,"samples_unverified":29,"pointer_only_for_licence":40,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}},{"paper":"/paper/cascade-r-cnn-delving-into-high-quality","title":"Cascade R-CNN: Delving into High Quality Object Detection","date":"2017-12-03","rows_on_this_dataset":1,"code_links":8,"syntology":{"read_at":"2026-09-24T18:15:14+00:00","samples_harvested":2,"samples_ran":2,"samples_unverified":0,"pointer_only_for_licence":2,"claim":"Per-sample execution on synthesized fixtures; not a correctness claim."}}],"syntology_totals":{"read_at":"2026-09-24T18:15:14+00:00","papers_with_samples":5,"samples_harvested":108,"samples_ran":52,"samples_unverified":56,"pointer_only_for_licence":44,"papers_with_no_sample_that_ran":1,"note":"the per-paper counts above, summed; not a rate"},"papers_note":"The archive never published its papers-using-dataset list; these are papers with a leaderboard row on this dataset's benchmarks."}