CAD&CG, Zhejiang University; University of Oxford; Ant Research; Pixelwise AI; Bytedance Seed

Yuxi Xiao, Jianyuan Wang, Nan Xue, Nikita Karaev, Iurii Makarov, Bingyi Kang, Xing Zhu, Hujun Bao, Yujun Shen, Xiaowei Zhou

@inproceedings{xiao2025spatialtrackerv2,
      title={SpatialTrackerV2: 3D Point Tracking Made Easy}, 
      author={Yuxi Xiao and Jianyuan Wang and Nan Xue and Nikita Karaev and Yuri Makarov and Bingyi Kang and Xing Zhu and Hujun Bao and Yujun Shen and Xiaowei Zhou},
      year={2025},
      booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
      url={https://arxiv.org/abs/2507.12462}, 
}

• [July 17, 2025]: Our paper is on arXiv
• [June 27, 2025]: SpatialTrackerV2 accepted by ICCV 2025
• [June 23, 2025]: Huggingface Space Demo launched! Try it out: 🤗 Huggingface Space

• Release quick start of SpaTrack2-offline
• Final version of Paper at PAPER.md
• Release SpaTrack2-online
• Training & Evaluation Codes.
• More supports for other Depth Model, e.g., DepthAnything, StereoFoundation, UniDepth, Metric3D.
• Ceres Python Bindings designed for SpatialTracker and Dynamic Reconstruction.

To set up the environment for running the SpaTrack model, follow these steps:

1. Clone the Repository:

   # clone the code
   git clone https://github.com/henry123-boy/SpaTrackerV2.git
   cd SpaTrackerV2
   
   # optionally download the example data to run the examples
   # Note: This will slowdown the clonining process, as it includes large files.
   git submodule update --init --recursive

2. Create a Virtual Environment: It's recommended to use a virtual environment to manage dependencies.

   conda create -n SpaTrack2 python=3.11
   conda activate SpaTrack2

3. Install Dependencies:

   Install the torch dependencies pip (tested with torch2.4).

   python -m pip install torch==2.4.1 torchvision==0.19.1 torchaudio==2.4.1 --index-url https://download.pytorch.org/whl/cu124

   Install the required Python packages using pip.

   python -m pip install -r requirements.txt

By following these steps, you should have a working environment ready to run the SpaTrack model.

Here are two examples of how to use SpaTrack2.

python inference.py --data_type="RGB" --data_dir="examples" --video_name="protein" --fps=3

We provide an example with Depth and Camera poses from MegaSAM. First, download the examples via:

sh scripts/download.sh

Then run inference with the command with below:

python inference.py --data_type="RGBD" --data_dir="assets/example1" --video_name="snowboard" --fps=1

Guidance will be displayed in the terminal after running inference.py.

Please follow the instructions in the app_3rd README to configure the dependencies. Then, install the required packages:

python -m pip install gradio==5.31.0 pako

Our Gradio demo enables users to easily track points on the target object. Just try:

python app.py