Forecasting Future Videos from Novel Views via Disentangled 3D Scene Representation | Read Paper on Bytez