diffusion-rig
diffusion-rig copied to clipboard
Question about transferring expression.
https://github.com/adobe-research/diffusion-rig/blob/d9b266102e1900032efc62e66a7f6c66dde5ff79/scripts/inference.py#L82
Hi, thanks for sharing your great work. Why first three dimension of pose codes are changed when transferring expression.
The pose parameters represent both neck and jaw poses. The last three are for jaw rotations, therefore we include them in the expression. You can see the figure 2 in the FLAME paper.