Generative Image Dynamics
This work addresses the challenge of generating realistic motion from static images for applications in video creation and interactive media, representing an incremental advance in image-based motion modeling.
The paper tackles the problem of modeling scene motion from a single image by learning a prior from real video sequences of oscillatory dynamics, resulting in a method that can generate seamless looping videos and enable realistic user interaction with objects in pictures.
We present an approach to modeling an image-space prior on scene motion. Our prior is learned from a collection of motion trajectories extracted from real video sequences depicting natural, oscillatory dynamics such as trees, flowers, candles, and clothes swaying in the wind. We model this dense, long-term motion prior in the Fourier domain:given a single image, our trained model uses a frequency-coordinated diffusion sampling process to predict a spectral volume, which can be converted into a motion texture that spans an entire video. Along with an image-based rendering module, these trajectories can be used for a number of downstream applications, such as turning still images into seamlessly looping videos, or allowing users to realistically interact with objects in real pictures by interpreting the spectral volumes as image-space modal bases, which approximate object dynamics.