2026-11-06 · America/Los_Angeles · 论文 · #11

EDGAR: Efficient Diffusion-Driven Gaussian Avatar Refinement via Synthetic Failure

Recent advances in Gaussian splatting have enabled efficient and photorealistic head avatar reconstruction, yet monocular video-based or image-based methods still suffer from under-constrained geometry and texture, leading to artifacts such as blurring, stretching, and temporal instability. We present EDGAR (Efficient Diffusion-Driven Gaussian Avatar Refinement), a conditional diffusion framework designed to enhance the quality of Gaussian avatars by correcting artifacts while preserving identity fidelity. To train EDGAR, we construct EDGAR-SET, a hybrid triplet dataset composed of…

热度 44.0 / 100；排名与评分保留该期记录。