RodinHD: High-Fidelity 3D Avatar Generation with Diffusion Models
This improves 3D avatar generation for applications in gaming, VR, and digital humans, but is incremental as it builds on existing diffusion model frameworks.
The paper tackles the problem of generating high-fidelity 3D avatars from portrait images, specifically addressing failures in capturing intricate details like hairstyles, and achieves notably better details than previous methods by training on 46K avatars.
We present RodinHD, which can generate high-fidelity 3D avatars from a portrait image. Existing methods fail to capture intricate details such as hairstyles which we tackle in this paper. We first identify an overlooked problem of catastrophic forgetting that arises when fitting triplanes sequentially on many avatars, caused by the MLP decoder sharing scheme. To overcome this issue, we raise a novel data scheduling strategy and a weight consolidation regularization term, which improves the decoder's capability of rendering sharper details. Additionally, we optimize the guiding effect of the portrait image by computing a finer-grained hierarchical representation that captures rich 2D texture cues, and injecting them to the 3D diffusion model at multiple layers via cross-attention. When trained on 46K avatars with a noise schedule optimized for triplanes, the resulting model can generate 3D avatars with notably better details than previous methods and can generalize to in-the-wild portrait input.