Video Diffusion Models are Training-free Motion Interpreter and Controller
This work addresses the need for interpretable and training-free motion control in video generation, offering a solution that reduces resource demands and enhances transparency for researchers and practitioners in AI and video synthesis.
The paper tackled the problem of understanding and controlling motion in video diffusion models without requiring training, by identifying and extracting motion-aware features (MOFT) that encode comprehensive motion information with clear interpretability, leading to competitive performance in generating natural and faithful motion across diverse architectures.
Video generation primarily aims to model authentic and customized motion across frames, making understanding and controlling the motion a crucial topic. Most diffusion-based studies on video motion focus on motion customization with training-based paradigms, which, however, demands substantial training resources and necessitates retraining for diverse models. Crucially, these approaches do not explore how video diffusion models encode cross-frame motion information in their features, lacking interpretability and transparency in their effectiveness. To answer this question, this paper introduces a novel perspective to understand, localize, and manipulate motion-aware features in video diffusion models. Through analysis using Principal Component Analysis (PCA), our work discloses that robust motion-aware feature already exists in video diffusion models. We present a new MOtion FeaTure (MOFT) by eliminating content correlation information and filtering motion channels. MOFT provides a distinct set of benefits, including the ability to encode comprehensive motion information with clear interpretability, extraction without the need for training, and generalizability across diverse architectures. Leveraging MOFT, we propose a novel training-free video motion control framework. Our method demonstrates competitive performance in generating natural and faithful motion, providing architecture-agnostic insights and applicability in a variety of downstream tasks.