Submitted by Lani Ko 46 3DreamBooth: High-Fidelity 3D Subject-Driven Video Generation Model Yonsei University 23 3