Anticipating the future by watching unlabeled video.
Year: 2015
Type: preprint
Abstract: In many computer vision applications, machines will need to reason beyond the present, and predict the future. This task is challenging because it requires leveraging ex-tensive commonsense knowledge of the world that is diffi-cult to write down. We believe that a promising resource for efficiently obtaining this knowledge is through the mas-sive amounts of readily available unlabeled video. In this paper, we present a large scale framework that capitalizes on temporal structure in unlabeled video to learn to antic-ipate both actions and objects in the future. The key idea behind our approach is that we can train deep networks to predict the visual representation of images in the future. We experimentally validate this idea on two challenging “in the wild ” video datasets, and our results suggest that learning with unlabeled videos significantly helps forecast actions and anticipate objects. 1.
Source: arXiv (Cornell University)
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1599058448