Home

Awesome

The code of 《Fast-Full-frame-Video-Stabilization-with-Iterative-Optimization》

You can run python assets/save_training_video_to_disk.py to prepare your synthetic dataset. Of course, the corresponding datasets should be prepared firstly (illustrated in the main paper). Then, you should also adjust the directory path in save_training_video_to_disk.py, including --image_data_path, --csv_path, --save_dir, coco_path.

You can run pre_video_flow_process.py to generate a confidence map sequence for the input video. Before running, please prepare the PDCNet code and make sure it runs successfully.

You can find the code for the network model corresponding to the paper in the core_model folder, including model.py, dataset.py, and loss.py, etc.

If you find this code helpful, please cite:

@inproceedings{zhao2023fast,
  title={Fast full-frame video stabilization with iterative optimization},
  author={Zhao, Weiyue and Li, Xin and Peng, Zhan and Luo, Xianrui and Ye, Xinyi and Lu, Hao and Cao, Zhiguo},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={23534--23544},
  year={2023}
}