{"id":13443469,"url":"https://github.com/yifanzhang713/IA-SSD","last_synced_at":"2025-03-20T16:31:38.230Z","repository":{"id":37700124,"uuid":"468900362","full_name":"yifanzhang713/IA-SSD","owner":"yifanzhang713","description":"Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Clouds (CVPR 2022, Oral)","archived":false,"fork":false,"pushed_at":"2025-03-03T11:33:27.000Z","size":51286,"stargazers_count":390,"open_issues_count":23,"forks_count":51,"subscribers_count":4,"default_branch":"main","last_synced_at":"2025-03-03T12:33:23.122Z","etag":null,"topics":["3d-object-detection","autonomous-driving","ia-ssd","object-detection","point-cloud","pytorch"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"apache-2.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/yifanzhang713.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}},"created_at":"2022-03-11T21:08:08.000Z","updated_at":"2025-03-03T11:33:31.000Z","dependencies_parsed_at":"2022-07-12T16:43:39.421Z","dependency_job_id":null,"html_url":"https://github.com/yifanzhang713/IA-SSD","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/yifanzhang713%2FIA-SSD","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yifanzhang713%2FIA-SSD/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yifanzhang713%2FIA-SSD/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/yifanzhang713%2FIA-SSD/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/yifanzhang713","download_url":"https://codeload.github.com/yifanzhang713/IA-SSD/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":244649884,"owners_count":20487514,"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-object-detection","autonomous-driving","ia-ssd","object-detection","point-cloud","pytorch"],"created_at":"2024-07-31T03:02:01.197Z","updated_at":"2025-03-20T16:31:38.224Z","avatar_url":"https://github.com/yifanzhang713.png","language":"Python","funding_links":[],"categories":["Python"],"sub_categories":[],"readme":"[![arXiv](https://img.shields.io/badge/arXiv-Paper-\u003cCOLOR\u003e.svg)](https://arxiv.org/abs/2203.11139)\n[![GitHub Stars](https://img.shields.io/github/stars/yifanzhang713/IA-SSD?style=social)](https://github.com/yifanzhang713/IA-SSD)\n![visitors](https://visitor-badge.glitch.me/badge?page_id=yifanzhang713/IA-SSD)\n\n# Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Clouds (CVPR 2022, Oral)\n\nThis is the official implementation of ***IA-SSD*** (CVPR 2022), a simple and highly efficient point-based detector for 3D LiDAR point clouds. For more details, please refer to:\n\n**Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Clouds** \u003cbr /\u003e\nYifan Zhang, [Qingyong Hu*](https://qingyonghu.github.io/), Guoquan Xu, Yanxin Ma, Jianwei Wan, [Yulan Guo](http://yulanguo.me/)\u003cbr /\u003e\n\n**[[Paper](https://arxiv.org/abs/2203.11139)] [[Video](https://www.youtube.com/watch?v=3jP2o9KXunA)]** \u003cbr /\u003e\n\n\u003cp align=\"center\"\u003e \u003cimg src=\"docs/imgs/IA-SSD.png\" width=\"100%\"\u003e \u003c/p\u003e\n\n\n## Getting Started\n### Installation\n\na. Clone this repository\n```shell\ngit clone https://github.com/yifanzhang713/IA-SSD.git \u0026\u0026 cd IA-SSD\n```\nb. Configure the environment\n\nWe have tested this project with the following environments:\n* Ubuntu18.04/20.04\n* Python = 3.7\n* PyTorch = 1.1\n* CUDA = 10.0\n* CMake \u003e= 3.13\n* spconv = 1.0 \n    ```shell\n    # install spconv=1.0 library\n    git clone https://github.com/yifanzhang713/spconv1.0.git\n    cd spconv1.0\n    sudo apt-get install libboostall-dev\n    python setup.py bdist_wheel\n    pip install ./dist/spconv-1.0*   # wheel file name may be different\n    cd ..\n    ```\n\n*You are encouraged to try to install higher versions above, please refer to the [official github repository](https://github.com/open-mmlab/OpenPCDet) for more information. **Note that the maximum number of parallel frames during inference might be slightly decrease due to the larger initial GPU memory footprint with updated `Pytorch` version.**\n\nc. Install `pcdet` toolbox.\n```shell\npip install -r requirements.txt\npython setup.py develop\n```\n\nd. Prepare the datasets. \n\nDownload the official KITTI with [road planes](https://drive.google.com/file/d/1d5mq0RXRnvHPVeKx6Q612z0YRO1t2wAp/view?usp=sharing) and Waymo datasets, then organize the unzipped files as follows:\n```\nIA-SSD\n├── data\n│   ├── kitti\n│   │   ├── ImageSets\n│   │   ├── training\n│   │   │   ├──calib \u0026 velodyne \u0026 label_2 \u0026 image_2 \u0026 (optional: planes)\n│   │   ├── testing\n│   │   ├── calib \u0026 velodyne \u0026 image_2\n│   ├── waymo\n│   │   │── ImageSets\n│   │   │── raw_data\n│   │   │   │── segment-xxxxxxxx.tfrecord\n|   |   |   |── ...\n|   |   |── waymo_processed_data_v0_5_0\n│   │   │   │── segment-xxxxxxxx/\n|   |   |   |── ...\n│   │   │── waymo_processed_data_v0_5_0_gt_database_train_sampled_1/\n│   │   │── waymo_processed_data_v0_5_0_waymo_dbinfos_train_sampled_1.pkl\n│   │   │── waymo_processed_data_v0_5_0_gt_database_train_sampled_1_global.npy (optional)\n│   │   │── waymo_processed_data_v0_5_0_infos_train.pkl (optional)\n│   │   │── waymo_processed_data_v0_5_0_infos_val.pkl (optional)\n├── pcdet\n├── tools\n```\nGenerate the data infos by running the following commands:\n```python \n# KITTI dataset\npython -m pcdet.datasets.kitti.kitti_dataset create_kitti_infos tools/cfgs/dataset_configs/kitti_dataset.yaml\n\n# Waymo dataset\npython -m pcdet.datasets.waymo.waymo_dataset --func create_waymo_infos \\\n    --cfg_file tools/cfgs/dataset_configs/waymo_dataset.yaml\n```\n\n\n### Quick Inference\nWe provide the pre-trained weight file so you can just run with that:\n```shell\ncd tools \n# To achieve fully GPU memory footprint (NVIDIA RTX2080Ti, 11GB).\npython test.py --cfg_file cfgs/kitti_models/IA-SSD.yaml --batch_size 100 \\\n    --ckpt IA-SSD.pth --set MODEL.POST_PROCESSING.RECALL_MODE 'speed'\n\n# To reduce the pressure on the CPU during preprocessing, a suitable batchsize is recommended, e.g. 16. (Over 5 batches per second on RTX2080Ti)\npython test.py --cfg_file cfgs/kitti_models/IA-SSD.yaml --batch_size 16 \\\n    --ckpt IA-SSD.pth --set MODEL.POST_PROCESSING.RECALL_MODE 'speed' \n```\n* Then detailed inference results can be found [here](docs/imgs/quick_inference.txt).\n\n\n\n\n### Training\nThe configuration files are in ```tools/cfgs/kitti_models/IA-SSD.yaml``` and ```tools/cfgs/waymo_models/IA-SSD.yaml```, and the training scripts are in ```tools/scripts```.\n\nTrain with single or multiple GPUs: (e.g., KITTI dataset)\n```shell\npython train.py --cfg_file cfgs/kitti_models/IA-SSD.yaml\n\n# or \n\nsh scripts/dist_train.sh ${NUM_GPUS} --cfg_file cfgs/kitti_models/IA-SSD.yaml\n```\n\n\n### Evaluation\n\nEvaluate with single or multiple GPUs: (e.g., KITTI dataset)\n```shell\npython test.py --cfg_file cfgs/kitti_models/IA-SSD.yaml  --batch_size ${BATCH_SIZE} --ckpt ${PTH_FILE}\n\n# or\n\nsh scripts/dist_test.sh ${NUM_GPUS} \\\n    --cfg_file cfgs/kitti_models/IA-SSD.yaml --batch_size ${BATCH_SIZE} --ckpt ${PTH_FILE}\n```\n\n### Experimental results\n\n#### KITTI dataset\n\nQuantitative results of different approaches on KITTI dataset (*test* set):\n\u003cp align=\"center\"\u003e \u003cimg src=\"docs/imgs/kitti_test.png\" width=\"100%\"\u003e \u003c/p\u003e\n\nQualitative results of our IA-SSD on KITTI dataset: \n| ![z](docs/imgs/kitti_1.gif)    | ![z](docs/imgs/kitti_2.gif)   |\n| -------------------------------- | ------------------------------- |\n| ![z](docs/imgs/kitti_3.gif)    | ![z](docs/imgs/kitti_4.gif)   |\n\n\nQuantitative results of different approaches on Waymo dataset (*validation* set):\n\u003cp align=\"center\"\u003e \u003cimg src=\"docs/imgs/waymo_val.png\" width=\"100%\"\u003e \u003c/p\u003e\n\nQualitative results of our IA-SSD on Waymo dataset:\n\n| ![z](docs/imgs/waymo_1.gif)    | ![z](docs/imgs/waymo_2.gif)   |\n| -------------------------------- | ------------------------------- |\n| ![z](docs/imgs/waymo_3.gif)    | ![z](docs/imgs/waymo_4.gif)   |\n\n\nQuantitative results of different approaches on ONCE dataset (*validation* set):\n\u003cp align=\"center\"\u003e \u003cimg src=\"docs/imgs/once_val.png\" width=\"100%\"\u003e \u003c/p\u003e\n\nQualitative result of our IA-SSD on ONCE dataset:\n\u003cp align=\"center\"\u003e \u003cimg src=\"docs/imgs/once.gif\" width=\"90%\"\u003e \u003c/p\u003e\n\n\n\n## Citation \nIf you find this project useful in your research, please consider citing:\n\n```\n@inproceedings{zhang2022not,\n  title={Not All Points Are Equal: Learning Highly Efficient Point-based Detectors for 3D LiDAR Point Clouds},\n  author={Zhang, Yifan and Hu, Qingyong and Xu, Guoquan and Ma, Yanxin and Wan, Jianwei and Guo, Yulan},\n  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},\n  year={2022}\n}\n```\n\n## Acknowledgement\n-  This work is built upon the `OpenPCDet` (version `0.5`), an open source toolbox for LiDAR-based 3D scene perception. Please refer to the [official github repository](https://github.com/open-mmlab/OpenPCDet) for more information.\n\n-  Parts of our Code refer to \u003ca href=\"https://github.com/qiqihaer/3DSSD-pytorch-openPCDet\"\u003e3DSSD-pytorch-openPCDet\u003c/a\u003e library and the the recent work \u003ca href=\"https://github.com/blakechen97/SASA\"\u003eSASA\u003c/a\u003e.\n\n\n\n## License\n\nThis project is released under the [Apache 2.0 license](LICENSE).\n\n\n## Related Repos\n1. [Highly Efficient and Unsupervised Framework for Moving Object Detection in Satellite Videos](https://github.com/ChaoXiao12/Moving-object-detection-in-satellite-videos-HiEUM)![GitHub stars](https://img.shields.io/github/stars/ChaoXiao12/Moving-object-detection-in-satellite-videos-HiEUM.svg?style=flat\u0026label=Star)\n2. [RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds](https://github.com/QingyongHu/RandLA-Net) ![GitHub stars](https://img.shields.io/github/stars/QingyongHu/RandLA-Net.svg?style=flat\u0026label=Star)\n3. [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)\n4. [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)\n5. [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)\n6. [SQN: Weakly-Supervised Semantic Segmentation of Large-Scale 3D Point Clouds](https://github.com/QingyongHu/SQN) ![GitHub stars](https://img.shields.io/github/stars/QingyongHu/SQN.svg?style=flat\u0026label=Star)\n7. [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)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fyifanzhang713%2FIA-SSD","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fyifanzhang713%2FIA-SSD","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fyifanzhang713%2FIA-SSD/lists"}