{"id":23593064,"url":"https://github.com/mrnerf/colmap-frame-processor-for-neural-3d-video","last_synced_at":"2025-11-04T10:30:40.539Z","repository":{"id":268375208,"uuid":"904150644","full_name":"MrNeRF/COLMAP-Frame-Processor-for-Neural-3D-Video","owner":"MrNeRF","description":"Script for preprocceing the Neural 3D Video dataset on a frame basis","archived":false,"fork":false,"pushed_at":"2024-12-16T10:57:55.000Z","size":5,"stargazers_count":1,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2024-12-27T08:18:58.789Z","etag":null,"topics":[],"latest_commit_sha":null,"homepage":null,"language":"Python","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/MrNeRF.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":"2024-12-16T10:49:34.000Z","updated_at":"2024-12-17T10:18:32.000Z","dependencies_parsed_at":"2024-12-16T11:41:00.852Z","dependency_job_id":null,"html_url":"https://github.com/MrNeRF/COLMAP-Frame-Processor-for-Neural-3D-Video","commit_stats":null,"previous_names":["mrnerf/colmap-frame-processor-for-neural-3d-video"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MrNeRF%2FCOLMAP-Frame-Processor-for-Neural-3D-Video","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MrNeRF%2FCOLMAP-Frame-Processor-for-Neural-3D-Video/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MrNeRF%2FCOLMAP-Frame-Processor-for-Neural-3D-Video/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/MrNeRF%2FCOLMAP-Frame-Processor-for-Neural-3D-Video/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/MrNeRF","download_url":"https://codeload.github.com/MrNeRF/COLMAP-Frame-Processor-for-Neural-3D-Video/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":239432761,"owners_count":19637798,"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":"2024-12-27T08:17:47.522Z","updated_at":"2025-11-04T10:30:40.498Z","avatar_url":"https://github.com/MrNeRF.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# COLMAP Frame Processor for Neural 3D Video\n\nThis script processes multi-view video data from the [Neural 3D Video](https://github.com/facebookresearch/Neural_3D_Video) dataset for COLMAP reconstruction. It processes each frame independently, creating a separate COLMAP reconstruction for each timestamp across all camera views.\n\n## Overview\n\nThe script takes synchronized multi-view video frames and performs the following operations:\n- Organizes data by frame timestamps\n- Runs COLMAP reconstruction for each frame independently\n- Creates undistorted images and sparse reconstruction data\n- Maintains a clean directory structure for further processing\n\n## Requirements\n\n- Python 3.7+\n- COLMAP (with CUDA support recommended)\n- numpy\n- The [Neural 3D Video](https://github.com/facebookresearch/Neural_3D_Video) dataset\n\n## Installation\n\n1. Clone this repository:\n```bash\ngit clone git@github.com:MrNeRF/COLMAP-Frame-Processor-for-Neural-3D-Video \ncd COLMAP-Frame-Processor-for-Neural-3D-Video\n```\n\n2. Install Python dependencies:\n```bash\npip install numpy\n```\n\n3. Install COLMAP:\n```bash\n# Ubuntu/Debian\nsudo apt-get install colmap\n\n# For other systems, see COLMAP documentation:\n# https://colmap.github.io/install.html\n```\n\n## Directory Structure\n\nInput data should follow the Neural 3D Video dataset structure. The script will create the following structure for each frame:\n\n```\nframe_XXXX/\n├── images/            # Undistorted images from all cameras\n└── sparse/\n    └── 0/            # COLMAP reconstruction data\n        ├── cameras.bin\n        ├── images.bin\n        └── points3D.bin\n```\n\n## Usage\n\nBasic usage:\n```bash\n# Process existing frames\npython process_frames.py /path/to/scene_data\n\n# Extract frames from videos and then process\npython process_frames.py /path/to/scene_data --extract-frames\n\n# Disable GPU usage\npython process_frames.py /path/to/scene_data --no-gpu\n```\n\n### Options\n- `--extract-frames`: Extract frames from MP4 videos before processing\n- `--no-gpu`: Disable GPU usage in COLMAP processing\n\n## Input Data Format\n\nThe script expects either:\n1. A directory containing MP4 files (when using `--extract-frames`)\n2. A directory with already extracted frames in the `images` subdirectory\n\nWhen using `--extract-frames`, the script will:\n1. Look for `.mp4` files in the input directory\n2. Extract frames using ffmpeg with format `{camera_name}_%04d.png`\n3. Place extracted frames in the `images` subdirectory\n4. Proceed with COLMAP processing\n\n## Processing Pipeline\n\nFor each frame, the script:\n1. Extracts images from all camera views\n2. Performs feature extraction using COLMAP\n3. Matches features across views\n4. Runs bundle adjustment\n5. Creates undistorted images\n6. Organizes output in a clean directory structure\n\n## Citation\n\nIf you use this code, please cite the original Neural 3D Video paper:\n\n```bibtex\n@inproceedings{gao2023neural3dvideo,\n  title={Neural 3D Video Synthesis from a Single Video},\n  author={Gao, Chen and Rong, Gengshan and Chen, Boyi and Mu, Ye and Wang, Yang and Black, Michael J. and Szeliski, Richard},\n  booktitle={CVPR},\n  year={2023}\n}\n```\n\n## License\n\nThis code is released under the MIT License. See [LICENSE](LICENSE) for details.\n\n## Acknowledgments\n\nThis script is designed to work with the [Neural 3D Video](https://github.com/facebookresearch/Neural_3D_Video) dataset from Meta Research.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmrnerf%2Fcolmap-frame-processor-for-neural-3d-video","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmrnerf%2Fcolmap-frame-processor-for-neural-3d-video","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmrnerf%2Fcolmap-frame-processor-for-neural-3d-video/lists"}