Shape-Aware Text-Driven Layered Video Editing | Read Paper on Bytez