{"id":18427690,"url":"https://github.com/eldar/snes","last_synced_at":"2025-07-03T17:36:12.769Z","repository":{"id":46748291,"uuid":"515201443","full_name":"eldar/snes","owner":"eldar","description":"SNeS: Learning Probably Symmetric Neural Surfaces from Incomplete Data","archived":false,"fork":false,"pushed_at":"2022-08-15T14:41:31.000Z","size":3212,"stargazers_count":61,"open_issues_count":1,"forks_count":7,"subscribers_count":6,"default_branch":"main","last_synced_at":"2025-04-07T16:42:56.271Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/eldar.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":"2022-07-18T13:43:11.000Z","updated_at":"2024-05-06T00:25:23.000Z","dependencies_parsed_at":"2022-09-22T16:42:25.320Z","dependency_job_id":null,"html_url":"https://github.com/eldar/snes","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/eldar/snes","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eldar%2Fsnes","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eldar%2Fsnes/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eldar%2Fsnes/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eldar%2Fsnes/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/eldar","download_url":"https://codeload.github.com/eldar/snes/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/eldar%2Fsnes/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":263370183,"owners_count":23456426,"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":[],"created_at":"2024-11-06T05:11:31.463Z","updated_at":"2025-07-03T17:36:12.180Z","avatar_url":"https://github.com/eldar.png","language":"Jupyter Notebook","funding_links":[],"categories":[],"sub_categories":[],"readme":"# SNeS: Learning Probably Symmetric Neural Surfaces from Incomplete Data\n\nEldar Insafutdinov*, Dylan Campbell*, Joao F Henriques and Andrea Vedaldi. *ECCV 2022*\n\n[Paper](https://arxiv.org/abs/2206.06340) | [Project page](https://www.robots.ox.ac.uk/~vgg/research/snes/)\n\n![](figures/snes_teaser.png)\n\n## Setup\n\nCreate a conda environment with python 3.9, pytorch 1.11 for CUDA 11.3 and pytorch3d 0.6.2:\n\n```shell\nconda env create -f environment.yml\n```\n\n## Dataset\n\nDownload a CO3D car category (or others) [here](https://ai.facebook.com/datasets/co3d-downloads/) and extract to `data/` such that it is organised like so:\n\n```\ndata/co3d\n|-- car\n    |-- \u003cscene_id\u003e\n        |-- images\n        |-- masks\n        |-- depths\n        |-- depth_masks\n    ...\n|-- toyplane\n...\n```\n\nExtract 3D bounding box fits for selected categories:\n\n```shell\ncd data; tar xzvf co3d_extra_data.tar.gz; cd ..\n```\n\n## Training\n\nRun the following command to train the model on a single scene using the structured train/val split as described in the paper:\n\n```shell\n(EXP=car/structured ID=157_17286_33548; python exp_runner.py gpu=0 mode=train config.file=exp/$EXP/config.yaml config.exp_name=$EXP/${ID} dataset.instance=\\'$ID\\')\n```\n\nTraining logs and model checkpoints of this run will be saved under `exp/car/structured/157_17286_33548`. In order to reproduce the results in the paper repeat the command above for every scene listed in `lists/co3d_car_structured_split.txt`.\n\nSimilarly, to train a model on the official CO3D split replace `car/structured` with `car/official` in the command above.\n\n## Test and evaluate\n\nRender test views:\n\n```shell\npython scripts/render_nvs_predictions.py --gpu=0 --exp=car/structured --instance=157_17286_33548\n```\n\nEvaluate:\n\n```\npython scripts/evaluate.py --gpu=0 --exp=car/structured\n```\n\n## Visualisation\n\n### Extract and display 3D mesh\n\nFirst run the following script:\n\n```shell\npython exp_runner.py gpu=0 mode=visualise_mesh test.web_vis=true config.exp_name=car/structured/157_17286_33548 visualisation.port=8888\n```\n\nAnd then open the link in the browser http://localhost:8888.\n\n### Show epipolar lines\n\nUse the notebook `scripts/vis_epipolar.ipynb` for interactive epipolar line visualisation. Click on anywhere in the left image and the corresponding epipolar line on the right will be shown\n\n### Visualise cameras\n\n\u003c!-- \u003cimg src=\"figures/vis_cameras.png\" width=\"500\"\u003e --\u003e\n![](figures/vis_cameras.png)\n\nUse `scripts/vis_cameras.py` (script adapted from [NeRF++](https://github.com/Kai-46/nerfplusplus)) to visualise cameras.\n\n## Citation\n\nIf you find this work useful consider citing our paper:\n\n```\n@article{insafutdinov2022snes,\n  title={SNeS: Learning Probably Symmetric Neural Surfaces from Incomplete Data},\n  author={Insafutdinov, Eldar and Campbell, Dylan and Henriques, Jo{\\~a}o F and Vedaldi, Andrea},\n  journal={arXiv preprint arXiv:2206.06340},\n  year={2022}\n}\n```\n\n## Acknowledgement\n\nThe starting point for this project was the implementation of [NeuS](https://github.com/Totoro97/NeuS). We thank the authors of this excellent paper.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Feldar%2Fsnes","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Feldar%2Fsnes","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Feldar%2Fsnes/lists"}