Online image annotation and 3D scene understanding
Bryan Russell
University of Washington
Talk
Bryan Russell is a research scientist as part of the Intel Science and Technology Center on Visual Computing and collaborates within the Department of Computer Science and Engineering at the University of Washington. His research is in the area of computer vision, with particular interests in object recognition and scene understanding. Bryan received his PhD in 2007 from MIT and was a postdoctoral fellow in the INRIA Willow team at the Ecole Normale Supérieure in Paris,France.
An early goal of computer vision was to build a system that could
automatically understand a 3D scene just by looking. This requires not
only the ability to extract 3D information from visual information
alone, but also to handle the large variety of different environments
and objects that comprise our visual world. To address the latter and
motivated by the availability of images on the Internet, we introduced
LabelMe, a web-based annotation tool that allows online users to label
objects and their spatial extent in images. To date, over 1 million
object annotations have been entered through the LabelMe online
annotation tool, with the annotations spanning a variety of different
scene and object classes. This has allowed progress on tasks such as
scene understanding and object recognition and enabled applications in
areas such as computer graphics and cognitive science. In this talk, I
will show the contents of the database, its growth over time, and
statistics of its usage. In addition, I will describe recent work in
recovering 3D structure for a single view by reasoning about 3D
geometric primitives, 3D spatial relations, and view
dependent/independent contours.