1 paper
Jiahui Li, Shengeng Tang, Jingxuan He +7
Reconstructing dynamic 3D scenes from monocular video remains fundamentally challenging due to the need to jointly infer motion, structure, and appearance from limited observations…