AutoVFX: Physically Realistic Video Editing from Natural Language Instructions
This work addresses the inaccessibility of VFX creation for everyday users by automating the process with natural language control.
The authors tackled the problem of laborious and complex visual effects (VFX) creation by developing AutoVFX, a framework that automatically generates realistic and dynamic VFX videos from a single video and natural language instructions, outperforming all competing methods by a large margin in quality, alignment, versatility, and plausibility.
Modern visual effects (VFX) software has made it possible for skilled artists to create imagery of virtually anything. However, the creation process remains laborious, complex, and largely inaccessible to everyday users. In this work, we present AutoVFX, a framework that automatically creates realistic and dynamic VFX videos from a single video and natural language instructions. By carefully integrating neural scene modeling, LLM-based code generation, and physical simulation, AutoVFX is able to provide physically-grounded, photorealistic editing effects that can be controlled directly using natural language instructions. We conduct extensive experiments to validate AutoVFX's efficacy across a diverse spectrum of videos and instructions. Quantitative and qualitative results suggest that AutoVFX outperforms all competing methods by a large margin in generative quality, instruction alignment, editing versatility, and physical plausibility.