{"id":18573568,"url":"https://github.com/prs-eth/scene-recognition-in-3d","last_synced_at":"2025-08-24T14:34:15.997Z","repository":{"id":40969282,"uuid":"243601101","full_name":"prs-eth/Scene-Recognition-in-3D","owner":"prs-eth","description":"[IROS 2020] Indoor Scene Recognition in 3D","archived":false,"fork":false,"pushed_at":"2023-07-06T21:17:05.000Z","size":2896,"stargazers_count":54,"open_issues_count":5,"forks_count":10,"subscribers_count":4,"default_branch":"master","last_synced_at":"2025-04-10T20:48:28.495Z","etag":null,"topics":["3dvision","computer-vision","scene-recognition","sparse-convolution"],"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/prs-eth.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,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2020-02-27T19:37:57.000Z","updated_at":"2024-09-18T18:49:22.000Z","dependencies_parsed_at":"2025-04-10T07:37:54.910Z","dependency_job_id":"3c6504ba-1bc5-4c00-99d7-037c8a52eaa7","html_url":"https://github.com/prs-eth/Scene-Recognition-in-3D","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/prs-eth/Scene-Recognition-in-3D","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FScene-Recognition-in-3D","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FScene-Recognition-in-3D/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FScene-Recognition-in-3D/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FScene-Recognition-in-3D/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/prs-eth","download_url":"https://codeload.github.com/prs-eth/Scene-Recognition-in-3D/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/prs-eth%2FScene-Recognition-in-3D/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":271887736,"owners_count":24839138,"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","status":"online","status_checked_at":"2025-08-24T02:00:11.135Z","response_time":111,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"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":["3dvision","computer-vision","scene-recognition","sparse-convolution"],"created_at":"2024-11-06T23:10:55.853Z","updated_at":"2025-08-24T14:34:15.976Z","avatar_url":"https://github.com/prs-eth.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# Scene-Recognition-in-3D(IROS'20)\nThis repository contains the source code and pretrained models from the paper [Scene Recognition in 3D](https://arxiv.org/abs/2002.12819). \n\n**Indoor Scene Recognition in 3D**\n\nShengyu Huang, Mikhail Usvyatsov, Konrad Schindler\n\nTo the best of our knowledge, we are the first to study the task of indoor scene recognition in 3D.\n![](tmp/arch.png)\n\n## News\n- 2020-11-26 Updates on pretrained weights and a small data pre-processing script\n- 2019-02-28 initial release\n\n## Citation \nIf you find our work useful, please consider citing\n```shell\n@article{huang2020indoor,\n  title={Indoor Scene Recognition in 3D},\n  author={Huang, Shengyu and Usvyatsov, Mikhail and Schindler, Konrad},\n  journal={IROS},\n  year={2020}\n}\n```\n\n## Dependencies\nThe required libraries can be easily installed by runing\n```shell\npip3 install -r requirements.txt\n```\nWe use MinkowskiEngine(v0.4.2) as our 3D sparse convolution framework. If you have problem with compiling it, please refer to [MinkowskiEngine](https://github.com/StanfordVL/MinkowskiEngine) for more details.\n\n## Data\nWe evaluate our model on [ScanNet benchmark](http://kaldir.vc.in.tum.de/scannet_benchmark/), the dataset is released under the [ScanNet Term of Use](http://kaldir.vc.in.tum.de/scannet/ScanNet_TOS.pdf), please contact ScanNet team for access. \n\nWe preprocess the raw data to be pth file for efficient access. We use [torch_cluster](https://github.com/rusty1s/pytorch_cluster) for GPU-based effficient farthest point sampling, you can find a sample under the folder ``tmp``. The train/val/test split can be found under the folder ``split``.\n\n\n## Pretrained model\nYou can download the pretrained models for testing from [here](https://drive.google.com/drive/folders/1L3L4jtUZLFQRo8IdZyba0YztSgGrgPyK?usp=sharing).\n\n## Train and test\n### DGCNN\nPlease change ``base_train``, ``base_val`` to your data folder. Then run \n```shell\npython3 main.py\n```\n\n### PointNet++\nPlease change ``base_train``, ``base_val`` to your data folder. Then run \n```shell\npython3 scene_classification.py --add_color True --num_points 4096\n```\n\n### sparseConv\n1. For Resnet14, change ``path_train`` and ``path_val`` then run \n```shell\npython3 main.py --num_points 4096 --use_color True\n```\n\n2. Follow the following 3 steps to train the multi-task learner, these three parts only differ slightly in our implementation, please refer to three folders respectively for more details: \n\n- train the sparse encoder and semantic segmentation decoder\n- freeze the encoder then train the sparse classification decoder \n- finetune both encoder and decoder with small lr\n\n## References\nHere are some great resources we benefit:\n\n- [Pointnet++](https://github.com/erikwijmans/Pointnet2_PyTorch)\n- [DGCNN](https://github.com/WangYueFt/dgcnn)\n- [SpatioTemporalSegmentation](https://github.com/chrischoy/SpatioTemporalSegmentation)\n\nI personally also recommend these three repositories regarding sparse convolution to you:\n\n- [MinkowskiEngine](https://github.com/StanfordVL/MinkowskiEngine)\n- [SparseConvNet](https://github.com/facebookresearch/SparseConvNet)\n- [ILA-SCNN](https://github.com/TimoHackel/ILA-SCNN)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fprs-eth%2Fscene-recognition-in-3d","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fprs-eth%2Fscene-recognition-in-3d","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fprs-eth%2Fscene-recognition-in-3d/lists"}