{"id":13444128,"url":"https://github.com/QingyongHu/RandLA-Net","last_synced_at":"2025-03-20T18:30:41.036Z","repository":{"id":37677859,"uuid":"224029310","full_name":"QingyongHu/RandLA-Net","owner":"QingyongHu","description":"🔥RandLA-Net in Tensorflow (CVPR 2020, Oral \u0026 IEEE TPAMI 2021)","archived":false,"fork":false,"pushed_at":"2023-07-11T22:42:23.000Z","size":393416,"stargazers_count":1369,"open_issues_count":190,"forks_count":325,"subscribers_count":31,"default_branch":"master","last_synced_at":"2025-03-18T06:49:46.342Z","etag":null,"topics":["3d-vision","computer-vision","s3dis","semantic-segmentation","semantic3d","semantickitti"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/QingyongHu.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null}},"created_at":"2019-11-25T19:51:01.000Z","updated_at":"2025-03-16T16:13:48.000Z","dependencies_parsed_at":"2022-07-14T00:50:36.851Z","dependency_job_id":"f474d852-c431-4466-8bc2-9f3c729b83f4","html_url":"https://github.com/QingyongHu/RandLA-Net","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/QingyongHu%2FRandLA-Net","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/QingyongHu%2FRandLA-Net/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/QingyongHu%2FRandLA-Net/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/QingyongHu%2FRandLA-Net/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/QingyongHu","download_url":"https://codeload.github.com/QingyongHu/RandLA-Net/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":244669858,"owners_count":20490869,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["3d-vision","computer-vision","s3dis","semantic-segmentation","semantic3d","semantickitti"],"created_at":"2024-07-31T03:02:19.785Z","updated_at":"2025-03-20T18:30:41.030Z","avatar_url":"https://github.com/QingyongHu.png","language":"Python","funding_links":[],"categories":["Sensor Processing","Point Cloud","Python"],"sub_categories":["Lidar and Point Cloud Processing","Examples"],"readme":"[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/191111236/semantic-segmentation-on-semantic3d)](https://paperswithcode.com/sota/semantic-segmentation-on-semantic3d?p=191111236)\n[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/191111236/3d-semantic-segmentation-on-semantickitti)](https://paperswithcode.com/sota/3d-semantic-segmentation-on-semantickitti?p=191111236)\n[![License CC BY-NC-SA 4.0](https://img.shields.io/badge/license-CC4.0-blue.svg)](https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode)\n\n# RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds (CVPR 2020)\n\nThis is the official implementation of **RandLA-Net** (CVPR2020, Oral presentation), a simple and efficient neural architecture for semantic segmentation of large-scale 3D point clouds. For technical details, please refer to:\n \n**RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds** \u003cbr /\u003e\n[Qingyong Hu](https://www.cs.ox.ac.uk/people/qingyong.hu/), [Bo Yang*](https://yang7879.github.io/), [Linhai Xie](https://www.cs.ox.ac.uk/people/linhai.xie/), [Stefano Rosa](https://www.cs.ox.ac.uk/people/stefano.rosa/), [Yulan Guo](http://yulanguo.me/), [Zhihua Wang](https://www.cs.ox.ac.uk/people/zhihua.wang/), [Niki Trigoni](https://www.cs.ox.ac.uk/people/niki.trigoni/), [Andrew Markham](https://www.cs.ox.ac.uk/people/andrew.markham/). \u003cbr /\u003e\n**[[Paper](https://arxiv.org/abs/1911.11236)] [[Video](https://youtu.be/Ar3eY_lwzMk)] [[Blog](https://zhuanlan.zhihu.com/p/105433460)] [[Project page](http://randla-net.cs.ox.ac.uk/)]** \u003cbr /\u003e\n \n \n\u003cp align=\"center\"\u003e \u003cimg src=\"http://randla-net.cs.ox.ac.uk/imgs/Fig3.png\" width=\"100%\"\u003e \u003c/p\u003e\n\n\n\t\n### (1) Setup\nThis code has been tested with Python 3.5, Tensorflow 1.11, CUDA 9.0 and cuDNN 7.4.1 on Ubuntu 16.04.\n \n- Clone the repository \n```\ngit clone --depth=1 https://github.com/QingyongHu/RandLA-Net \u0026\u0026 cd RandLA-Net\n```\n- Setup python environment\n```\nconda create -n randlanet python=3.5\nsource activate randlanet\npip install -r helper_requirements.txt\nsh compile_op.sh\n```\n\n**Update 03/21/2020, pre-trained models and results are available now.** \nYou can download the pre-trained models and results [here](https://drive.google.com/open?id=1iU8yviO3TP87-IexBXsu13g6NklwEkXB).\nNote that, please specify the model path in the main function (e.g., `main_S3DIS.py`) if you want to use the pre-trained model and have a quick try of our RandLA-Net.\n\n### (2) S3DIS\nS3DIS dataset can be found \n\u003ca href=\"https://docs.google.com/forms/d/e/1FAIpQLScDimvNMCGhy_rmBA2gHfDu3naktRm6A8BPwAWWDv-Uhm6Shw/viewform?c=0\u0026w=1\"\u003ehere\u003c/a\u003e. \nDownload the files named \"Stanford3dDataset_v1.2_Aligned_Version.zip\". Uncompress the folder and move it to \n`/data/S3DIS`.\n\n- Preparing the dataset:\n```\npython utils/data_prepare_s3dis.py\n```\n- Start 6-fold cross validation:\n```\nsh jobs_6_fold_cv_s3dis.sh\n```\n- Move all the generated results (*.ply) in `/test` folder to `/data/S3DIS/results`, calculate the final mean IoU results:\n```\npython utils/6_fold_cv.py\n```\n\nQuantitative results of different approaches on S3DIS dataset (6-fold cross-validation):\n\n![a](http://randla-net.cs.ox.ac.uk/imgs/S3DIS_table.png)\n\nQualitative results of our RandLA-Net:\n\n| ![2](imgs/S3DIS_area2.gif)   | ![z](imgs/S3DIS_area3.gif) |\n| ------------------------------ | ---------------------------- |\n\n\n\n### (3) Semantic3D\n7zip is required to uncompress the raw data in this dataset, to install p7zip:\n```\nsudo apt-get install p7zip-full\n```\n- Download and extract the dataset. First, please specify the path of the dataset by changing the `BASE_DIR` in \"download_semantic3d.sh\"    \n```\nsh utils/download_semantic3d.sh\n```\n- Preparing the dataset:\n```\npython utils/data_prepare_semantic3d.py\n```\n- Start training:\n```\npython main_Semantic3D.py --mode train --gpu 0\n```\n- Evaluation:\n```\npython main_Semantic3D.py --mode test --gpu 0\n```\nQuantitative results of different approaches on Semantic3D (reduced-8):\n\n![a](http://randla-net.cs.ox.ac.uk/imgs/Semantic3D_table.png)\n\nQualitative results of our RandLA-Net:\n\n| ![z](imgs/Semantic3D-1.gif)    | ![z](http://randla-net.cs.ox.ac.uk/imgs/Semantic3D-2.gif)   |\n| -------------------------------- | ------------------------------- |\n| ![z](imgs/Semantic3D-3.gif)    | ![z](imgs/Semantic3D-4.gif)   |\n\n\n\n**Note:** \n- Preferably with more than 64G RAM to process this dataset due to the large volume of point cloud\n\n\n### (4) SemanticKITTI\n\nSemanticKITTI dataset can be found \u003ca href=\"http://semantic-kitti.org/dataset.html#download\"\u003ehere\u003c/a\u003e. Download the files\n related to semantic segmentation and extract everything into the same folder. Uncompress the folder and move it to \n`/data/semantic_kitti/dataset`.\n \n- Preparing the dataset:\n```\npython utils/data_prepare_semantickitti.py\n```\n\n- Start training:\n```\npython main_SemanticKITTI.py --mode train --gpu 0\n```\n\n- Evaluation:\n```\nsh jobs_test_semantickitti.sh\n```\n\nQuantitative results of different approaches on SemanticKITTI dataset:\n\n![s](http://randla-net.cs.ox.ac.uk/imgs/SemanticKITTI_table.png)\n\nQualitative results of our RandLA-Net:\n\n![zzz](imgs/SemanticKITTI-2.gif)    \n\n\n### (5) Demo\n\n\u003cp align=\"center\"\u003e \u003ca href=\"https://youtu.be/Ar3eY_lwzMk\"\u003e\u003cimg src=\"http://randla-net.cs.ox.ac.uk/imgs/demo_cover.png\" width=\"80%\"\u003e\u003c/a\u003e \u003c/p\u003e\n\n\n### Citation\nIf you find our work useful in your research, please consider citing:\n\n\t@article{hu2019randla,\n\t  title={RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds},\n\t  author={Hu, Qingyong and Yang, Bo and Xie, Linhai and Rosa, Stefano and Guo, Yulan and Wang, Zhihua and Trigoni, Niki and Markham, Andrew},\n\t  journal={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},\n\t  year={2020}\n\t}\n\t\n\t@article{hu2021learning,\n\t  title={Learning Semantic Segmentation of Large-Scale Point Clouds with Random Sampling},\n\t  author={Hu, Qingyong and Yang, Bo and Xie, Linhai and Rosa, Stefano and Guo, Yulan and Wang, Zhihua and Trigoni, Niki and Markham, Andrew},\n\t  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},\n\t  year={2021},\n\t  publisher={IEEE}\n\t}\n\n\n### Acknowledgment\n-  Part of our code refers to \u003ca href=\"https://github.com/jlblancoc/nanoflann\"\u003enanoflann\u003c/a\u003e library and the the recent work \u003ca href=\"https://github.com/HuguesTHOMAS/KPConv\"\u003eKPConv\u003c/a\u003e.\n-  We use \u003ca href=\"https://www.blender.org/\"\u003eblender\u003c/a\u003e to make the video demo.\n\n\n### License\nLicensed under the CC BY-NC-SA 4.0 license, see [LICENSE](./LICENSE).\n\n\n### Updates\n* 21/03/2020: Updating all experimental results\n* 21/03/2020: Adding pretrained models and results\n* 02/03/2020: Code available!\n* 15/11/2019: Initial release！\n\n## Related Repos\n1. [SoTA-Point-Cloud: Deep Learning for 3D Point Clouds: A Survey](https://github.com/QingyongHu/SoTA-Point-Cloud) ![GitHub stars](https://img.shields.io/github/stars/QingyongHu/SoTA-Point-Cloud.svg?style=flat\u0026label=Star)\n2. [SensatUrban: Learning Semantics from Urban-Scale Photogrammetric Point Clouds](https://github.com/QingyongHu/SpinNet) ![GitHub stars](https://img.shields.io/github/stars/QingyongHu/SensatUrban.svg?style=flat\u0026label=Star)\n3. [3D-BoNet: Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds](https://github.com/Yang7879/3D-BoNet) ![GitHub stars](https://img.shields.io/github/stars/Yang7879/3D-BoNet.svg?style=flat\u0026label=Star)\n4. [SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registration](https://github.com/QingyongHu/SpinNet) ![GitHub stars](https://img.shields.io/github/stars/QingyongHu/SpinNet.svg?style=flat\u0026label=Star)\n5. [SQN: Weakly-Supervised Semantic Segmentation of Large-Scale 3D Point Clouds with 1000x Fewer Labels](https://github.com/QingyongHu/SQN) ![GitHub stars](https://img.shields.io/github/stars/QingyongHu/SQN.svg?style=flat\u0026label=Star)\n\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FQingyongHu%2FRandLA-Net","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FQingyongHu%2FRandLA-Net","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FQingyongHu%2FRandLA-Net/lists"}