zwyking/fast-stab

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

• Synthetic Dataset

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.

• Generate Confidence Maps

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.

fast-stab's Issues

the code of confidence map back-propagation?

can you plz provide the code about confidence map back-propagation? i try to reimplement it according to the supplement of your paper but failed.

Inquiry about Code Environment Requirements

Hello,

I wanted to express my admiration for the valuable contribution you’ve made with your code on GitHub. It’s truly fascinating! However, I noticed that the code lacks information about the necessary installation environment, such as the Python version or PyTorch version. Could you kindly provide details on how to set up the environment to successfully run the code?

Thank you for your time and expertise!

ETA for code release?

excellent work and question about the code

Thanks for your excellent work about fast-stab!
Im a gradulation student worked on video stablization!
and I want to know how could your release the training and testing code!
Looking forward your reply and congradulation for you new work!

加油

请问预计什么时候开源代码呢？