Generation Challenge Result
There are five submissions in total. Paper #3346 did not generate the target expression on the source face. Therefore, it was removed from the evaluation.
Here are the results of three experts' evaluations of four submitted videos in three sheets. The three experts are FACS AU coders from UK, Finland and China, each labeling the results without interfering with each other.
The submitted generated videos could be downloaded from https://drive.google.com/drive/folders/1vgbHOXhKgD_7Gvrrg3AF2i59yJB2nuaw?usp=sharing
The final decision were made based on the result performance and the paper quality.
Updates: The GitHub links for the top three winning methods
# 3295
https://github.com/Necolizer/Facial-Prior-Based-FOMM
# 3282
https://github.com/xinqi-fan/FMEG. All pretained weights can be found in OneDrive.
# 3320
https://github.com/QAQFrank/MEGC2021