On deblurring after video synthesis

[Explore008]

Hello, author. Your work in this field is excellent. I am currently studying a work about video-driven two-dimensional human image movement, and finally a video is generated, but the frame rate is low and the artifacts are very serious. I want to introduce your work into my article and optimize it directly in the final generated video. How should I call it?

[XiangZ-0]

Hi, thanks for your interest in our work. Assuming you already got the synthesized videos and the corresponding events, training EVDI on them is the same as training on regular videos. All you need to do is first prepare your data via 'Prepare_data.py' and then run 'Train.py' to optimize your EVDI model. Please see the 'Train' section in README.md for detailed instructions. Hope this helps!