LCA is accepted at CVPR 2026! 🚀
We introduce a pre/post-training paradigm for 3D avatars (1M in-the-wild videos ➡️ studio data).
The result? High-fidelity full-body avatars with emergent relightability and zero-shot stylization.
Project: https://t.co/liddQrFIqF
#CVPR2026
@sentfromspacevr love the work! could you also ask to cite the original research paper on your gumroad webpage if people use the add-on for academic/research work? Thank you! https://t.co/DpcDhk2qtF
Today at @SIGGRAPHAsia , I will be presenting a simple and robust method for transferring skin weights from rigged body meshes to any clothing meshes, thus helping automate the tedious process of manual weight painting.
https://t.co/hSpTlNNWz8
@_AlecJacobson@SIGGRAPHAsia@JustinMSolomon thank you! Just read the paper, I should definitely try it. In extreme cases where the inpainting regions are very large, properly enforcing non-negativity is probably the right thing to do even if the computation cost is a bit higher.
All you need to implement this is to compute the Laplacian/Mass matrices and do the closest point queries. This technique is already part of the experimental Chaos Panel Cloth Editor shipped in UE 5.3.
We do an initial transfer by copying weights from the source mesh only for those vertices on the target mesh where we have high confidence in obtaining the ground truth weights. And "inpaint" weights for all other vertices.
Unlike skin binding techniques (i.e., compute initial per-vertex weights given a skeleton), there were no good transfer techniques (i.e., compute per-vertex weight given another mesh that has the artist-painted weights) either in commercial software or research literature, especially for non-skin-tight clothing.
@stevebrownett@UltHackKeyboard Just ordered mine a few days ago with the same attachments. Does your thumb get tired after long hours of using attachments? Is it hard to reach them?