3D Vision in a Changing World

3D reconstruction from images has been a tremendous success-story of computer vision, with city-scale reconstruction now a reality. However, these successes apply almost exclusively in a static world, where the only motion is that of the camera. Even with the advent of realtime depth cameras, full 3D modelling of dynamic scenes lags behind the rigid-scene case, and for many objects of interest (e.g. animals moving in natural environments), depth sensing remains challenging. In this talk, I will discuss a range of recent work in the modelling of nonrigid real-world 3D shape from 2D images, for example building generic animal models from internet photo collections. While the state of the art depends heavily on dense point tracks from textured surfaces, I will talk about recovering shape from largely textureless objects such as dolphins, by incorporating the strong constraints given by the object’s silhouette.