ICDepth: Taming Video Diffusion Models for Video Depth Estimation via In-Context Conditioning figure
AlphaXiv 中文概览(可滚动查看)