Giters

zhaoyuzhi / SVCNet

SVCNet: Scribble-based Video Colorization Network with Temporal Aggregation. IEEE TIP, 2023

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Inference on your own data

tomershmul opened this issue · comments

commented

Hi
I would like to use SVCNet on my video with my scribble.
Is there a guide on how to do inference on your own data?
How should I create the scribbles?
Thanks

commented

Hi @tomershmul,

  1. Creating your own scribbles: https://github.com/zhaoyuzhi/SVCNet/tree/main/GCS. You need first provide the first color scribble; then, you can use generate_color_scribbles_video.py script to obtain the following scribbles based on the optical flows of your own video.
  2. Inference with your generated scribbles: https://github.com/zhaoyuzhi/SVCNet/tree/main/SSNet. Please follow the guide in the README file, e.g., running test.py.
commented

I have included a fast inference guide here: https://github.com/zhaoyuzhi/SVCNet#4-fast-inference

Once you generate your own color scribbles, go into the test.py, just replace the --base_root and --scribble_root with your own grayscale frames and your generated color scribbles, respectively.