Video textures

Abstract

This paper introduces a new type of medium, called a video texture, which has qualities somewhere between those of a photograph and a video. A video texture provides a continuous infinitely varying stream of images. While the individual frames of a video texture may be repeated from time to time, the video sequence as a whole is never repeated exactly. Video textures can be used in place of digital photos to infuse a static image with dynamic qualities and explicit action. We present techniques for analyzing a video clip to extract its structure, and for synthesizing a new, similar looking video of arbitrary length. We combine video textures with view morphing techniques to obtain 3D video textures. We also introduce videobased animation, in which the synthesis of video textures can be guided by a user through high-level interactive controls. Applications of video textures and their extensions include the display of dynamic scenes on web pages, the creation of dynamic backdrops for special effects and games, and the interactive control of video-based animation.

Webcam Synopsis: Peeking Around the World

“…In [23] infinite video is generated from a short video clip. While having a different goal, objects (video sprites) are separated from the background and rendered at arbitrary video locations to create novel videos.…”

Probabilistic motion sequence generation

“…Following the work of Schödl et al [20] and Lee et al [16] we also model the data as a first-order Markov process, hence the transition between states depends only on the current state, which in our case are the given key frames in A. The Markov process is represented as a matrix P ij storing the probability of a transition from frame i to frame j.…”

“…The basic idea to search for coherence in frames of image sequences was introduced by Schödl et al [20] with the term video textures. The goal of this work was to create endless or fixed length image sequences with non-trivial repetitions.…”

“…The main idea is, that similarities between frames are transferred into the PCA weights. As in Schödl et al [20] we compute a coherence and transition matrix, which can be used to create either infinite random or length controlled loops during rendering, as shown in figure 1. The rest of the paper is organized as follows.…”

“…On the other hand creating a large motion database, as it is sometimes done in the gaming industry, is also a costly task. In this paper we extend the idea of video textures as introduced by Schödl et al [20] to geometry and generalized attributes like vertex normals, velocities or reflection properties. Similar to Lee et al [16] control of the avatar can be achieved, but in contrast to the latter we are not restricted to motion capture data.…”