search

YuDeng / Portrait-4D

Portrait4D: Learning One-Shot 4D Head Avatar Synthesis using Synthetic Data (CVPR 24); Portrait4D-v2: Pseudo Multi-View Data Creates Better 4D Head Synthesizer (ECCV 2024)
MIT License
211 stars 8 forks source link

Congrats #1

Closed Inferencer closed 6 months ago

Inferencer commented 6 months ago

No issues, just wanted to say congratulations it's a clever approach and the results look amazing, There was a recent paper called InvertAvatar: Incremental GAN Inversion for Generalized Head Avatars, and another called EMO: Emote Portrait Alive but I think yours is SOTA so a round of applause for you!

I look forward to release and can't wait to tinker with the code & train new models where applicable

Inferencer commented 6 months ago

actually one issue, there seems to be a difference between Portrait-4D and Portait-4D V2 with no paper link on the project page of v2, Should I refer to Portrait-4D's paper or will you be releasing a separate paper for V2?

YuDeng commented 6 months ago

actually one issue, there seems to be a difference between Portrait-4D and Portait-4D V2 with no paper link on the project page of v2, Should I refer to Portrait-4D's paper or will you be releasing a separate paper for V2?

Thanks for your interest. The paper for Portrait4D-v2 can be found on arxiv now: https://arxiv.org/abs/2403.13570

Inferencer commented 6 months ago

Oh one more thing, are you planning to release v1 & v2 in April or just v1 ? Because v2 is what has peaked my interest and its better to let me down now than shock me in April if it's just v1

YuDeng commented 6 months ago

Currently we only plan to release v1 in April. But we will promote the open source of v2 as soon as possible.

skyler14 commented 5 months ago

Is there any updates?