https://github.com/hw140701/videoto3dposeandbvh

Convert video to the bvh motion file
https://github.com/hw140701/videoto3dposeandbvh

bvh human-pose-estimation pose-estimation video-to-3d-pose

Last synced: 2 months ago
JSON representation

Convert video to the bvh motion file

• Host: GitHub
• URL: https://github.com/hw140701/videoto3dposeandbvh
• Owner: HW140701
• License: other
• Created: 2019-12-19T02:49:52.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2023-10-03T23:07:40.000Z (over 1 year ago)
• Last Synced: 2025-03-29T18:07:32.292Z (3 months ago)
• Topics: bvh, human-pose-estimation, pose-estimation, video-to-3d-pose
• Language: Python
• Homepage: https://www.stubbornhuang.com/613
• Size: 86 MB
• Stars: 401
• Watchers: 10
• Forks: 65
• Open Issues: 35
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Video to 3DPose and Bvh motion file









This project integrates some project working, example as [VideoPose3D](https://github.com/facebookresearch/VideoPose3D),[video-to-pose3D](https://github.com/zh-plus/video-to-pose3D) , [video2bvh](https://github.com/KevinLTT/video2bvh "video2bvh"), [AlphaPose](https://github.com/MVIG-SJTU/AlphaPose), [Higher-HRNet-Human-Pose-Estimation](https://github.com/HRNet/Higher-HRNet-Human-Pose-Estimation),[openpose](https://github.com/CMU-Perceptual-Computing-Lab/openpose), thanks for the mentioned above project.

The project extracted the 2d joint key point from the video by using [AlphaPose](https://github.com/MVIG-SJTU/AlphaPose),[HRNet](https://github.com/HRNet/Higher-HRNet-Human-Pose-Estimation) and so on. Then transform the 2d point to the 3D joint point by using [VideoPose3D](https://github.com/facebookresearch/VideoPose3D). Finally We convert the 3d joint point to the bvh motion file.

# Environment

- Windows 10

- Anaconda

- Python > 3.6

# Dependencies

>You can refer to the project dependencies of [video-to-pose3D](https://github.com/zh-plus/video-to-pose3D) for setting.

This is the dependencies of the project of [video-to-pose3D](https://github.com/zh-plus/video-to-pose3D), and modifyed by me to solve some bug.

- **Packages**

  - Pytorch > 1.1.0 (I use the Pytorch1.1.0 - GPU)

  - [torchsample](https://github.com/MVIG-SJTU/AlphaPose/issues/71#issuecomment-398616495)

  - [ffmpeg](https://ffmpeg.org/download.html) (note:you must copy the ffmpeg.exe to the directory of python install)

  - tqdm

  - pillow

  - scipy

  - pandas

  - h5py

  - visdom

  - nibabel

  - opencv-python (install with pip)

  - matplotlib

- **2D Joint detectors**

     - Alphapose (Recommended)

       - Download **duc_se.pth** from ([Google Drive](https://drive.google.com/open?id=1OPORTWB2cwd5YTVBX-NE8fsauZJWsrtW) | [Baidu pan](https://pan.baidu.com/s/15jbRNKuslzm5wRSgUVytrA)),

         place to `./joints_detectors/Alphapose/models/sppe`

       - Download **yolov3-spp.weights** from ([Google Drive](https://drive.google.com/open?id=1D47msNOOiJKvPOXlnpyzdKA3k6E97NTC) | [Baidu pan](https://pan.baidu.com/s/1Zb2REEIk8tcahDa8KacPNA)),

         place to `./joints_detectors/Alphapose/models/yolo`

     - HR-Net (Bad 3d joints performance in my testing environment)

       - Download **pose_hrnet*** from [Google Drive](https://drive.google.com/drive/folders/1nzM_OBV9LbAEA7HClC0chEyf_7ECDXYA), 

         place to `./joints_detectors/hrnet/models/pytorch/pose_coco/`

       - Download **yolov3.weights** from [here](https://pjreddie.com/media/files/yolov3.weights),

         place to `./joints_detectors/hrnet/lib/detector/yolo`

- **3D Joint detectors**

      - Download **pretrained_h36m_detectron_coco.bin** from [here](https://dl.fbaipublicfiles.com/video-pose-3d/pretrained_h36m_detectron_coco.bin),

        place it into `./checkpoint` folder

- **2D Pose trackers (Optional)**

      - PoseFlow (Recommended)

        No extra dependences

      - LightTrack (Bad 2d tracking performance in my testing environment)

        - See [original README](https://github.com/Guanghan/lighttrack), and perform same *get started step* on `./pose_trackers/lighttrack`

# How to Use it

Please place your video to the .\outputs\inputvideo, and setting the path to the videopose.py, like this 

    inference_video('outputs/inputvideo/kunkun_cut.mp4', 'alpha_pose')

Waiting some minute, you can see the output video in the \outputs\outputvideo directory,and the bvh file in the \outputs\outputvideo\alpha_pose_kunkun_cut\bvh directory.