本地上传-avi格式-test-标注数据集

\n# Road Anomaly dataset\n\n\n\nThis dataset contains images of unusual dangers which can be encountered by a vehicle on the road - animals, rocks, traffic cones and other obstacles.\nIts purpose is testing autonomous driving perception algorithms in rare but safety-critical circumstances.\n\nThe dataset contains images with associated per-pixel labels.\nThe labeling has been performed with our [LabelGrab](github.com/cvlab-epfl/LabelGrab) tool. Most frames retain the editor's files and can be further edited; some are missing because the labeling was done at a different resolution and rescaled. \nIf you need the editor files and instance labels for all frames, please [contact us](mailto:krzysztof.lis@epfl.ch).\n\nFor any additional information, please contact [Krzysztof Lis](mailto:krzysztof.lis@epfl.ch).\n\nThe images are provided for research purposes.\nWe do not own the images; the authors and sources are listed in `credits.txt`.\n\n### Directory structure\n\n`RoadAnomaly`\n- `readme.md`\n- `frame_list.json` - list of all image file names\n- `credits.txt` - sources of the images\n- `frames/`\n - `frame_name.webp` - image\n - `frame_name.labels/labels_semantic.png` - per pixel label, background = `0` and anomaly = `2`\n - `frame_name.labels/labels_instance.png` - instance id of objects, some are missing at this time.\n\n### References\nIf you find this dataset useful, please cite:\n\n```\n@inproceedings{DetectingTheUnexpected2019,\n\ttitle = {Detecting the Unexpected via Image Resynthesis},\n\tauthor = {Krzysztof Lis and Krishna Nakka and Pascal Fua and Mathieu Salzmann},\n\tbooktitle = {International Conference on Computer Vision},\n\tyear = 2019,\n\turl = {https://infoscience.epfl.ch/record/269093?ln=en}\n}\n```\n