About the re-implemented model of YoutubeHand

[EAST-J]

Hi Xingyu, Thanks for your great work. I see in the Experiment part of the paper, you use YouTubeHand as the baseline. I wonder would you release this part of code and model?

[SeanChenxy]

Hi, I have re-implement YoutubeHand (https://github.com/SeanChenxy/HandMesh/blob/main/cmr/ytbhand.py). But I have not re-train it. You could use it according to your needs.

[yuyu19970716]

Hello, I can't find the ytbhand.py file in the code, can you provide it?

[EAST-J]

Hello, I can't find the ytbhand.py file in the code, can you provide it?

You can find in cmr/models/ytbhand.py

[yuyu19970716]

Oh I found it! Thank you so much, this question is so sloppy!

[yuyu19970716]

Hello, I have another question to ask. This question may also be rather silly. I want to correspond the model code of DenseStack_Backnone with the paper, but I feel that I am looking at it in a mess. May I ask where the 2d encoder-decoder architecture and cascade are reflected in the model? At the same time, is the final generated latent Fe? Is uv-reg Lp? So sorry,I'm new to this direction.
Looking forward to your reply! Thanks in advance!