Shannon meets Gibson: Actionable Information in Vision
Tues., Oct 27, 2009
2:00 - 3:00 PM
400 Cory (Hughes room)
2:00 - 3:00 PM
400 Cory (Hughes room)
I will discuss a notion of visual information as complexity _not_ of the raw dat
a, but of the images after the effects of nuisance factors such as viewpoint and
illumination are discounted. It is rooted in ideas of J. J. Gibson, and stands
in contrast to traditional information as entropy or coding length of the data r
egardless of its use, and regardless of the nuisance factors affecting it. Its c
omputation is made possible by a recent characterization of the set of images mo
dulo viewpoint and contrast changes, that induce group (invertible) transformati
ons on the domain and range of the image. The non-invertibility of nuisances suc
h as occlusion and quantization induces an ``information gap'' that can only be
bridged by controlling the data acquisition process.
Measuring visual information entails early vision operations, tailored to the st
ructure of the nuisances so as to be ``lossless'' with respect to visual decisio
n and control tasks (as opposed to data transmission and storage tasks implicit
in traditional information theory). I illustrate these ideas on visual explorati
on, whereby a ``Shannonian Explorer'' navigates unaware of the structure of the
physical space surrounding it, while a ``Gibsonian Explorer'' is guided by the t
opology of the environment, despite measuring only images of it, without perform
ing 3D reconstruction. This operational definition of visual information suggest
s desirable properties that a visual representation should possess to best accom
plish vision-based decision and control tasks.
Bio:
Stefano Soatto received his Ph.D. in Control and Dynamical Systems from the Cali fornia Institute of Technology in 1996; he joined UCLA in 2000 after being Assis tant and then Associate Professor of Electrical and Biomedical Engineering at Wa shington University, and Research Associate in Applied Sciences at Harvard Unive rsity. Between 1995 and 1998 he was also Ricercatore in the Department of Mathem atics and Computer Science at the University of Udine - Italy. He received his D .Ing. degree (highest honors) from the University of Padova- Italy in 1992. His general research interests are in Computer Vision and Nonlinear Estimation and C ontrol Theory. In particular, he is interested in ways for computers to use sens ory information (e.g. vision, sound, touch) to interact with humans and the envi ronment. Dr. Soatto is the recipient of the David Marr Prize for work on Euclide an reconstruction and reprojection up to subgroups (with Y. Ma, J. Kosecka and S . Sastry). He also received the Siemens Prize with the Outstanding Paper Award f rom the IEEE Computer Society for his work on optimal structure from motion (wit h R. Brockett). He received the National Science Foundation Career Award and the Okawa Foundation Grant. He is a Member of the Editorial Board of the Internatio nal Journal of Computer Vision (IJCV) and Foundations and Trends in Computer Gra phics and Vision. He is the founder and director of the UCLA Vision Lab; more in formation is available at http://vision.ucla. edu. ~
Bio:
Stefano Soatto received his Ph.D. in Control and Dynamical Systems from the Cali fornia Institute of Technology in 1996; he joined UCLA in 2000 after being Assis tant and then Associate Professor of Electrical and Biomedical Engineering at Wa shington University, and Research Associate in Applied Sciences at Harvard Unive rsity. Between 1995 and 1998 he was also Ricercatore in the Department of Mathem atics and Computer Science at the University of Udine - Italy. He received his D .Ing. degree (highest honors) from the University of Padova- Italy in 1992. His general research interests are in Computer Vision and Nonlinear Estimation and C ontrol Theory. In particular, he is interested in ways for computers to use sens ory information (e.g. vision, sound, touch) to interact with humans and the envi ronment. Dr. Soatto is the recipient of the David Marr Prize for work on Euclide an reconstruction and reprojection up to subgroups (with Y. Ma, J. Kosecka and S . Sastry). He also received the Siemens Prize with the Outstanding Paper Award f rom the IEEE Computer Society for his work on optimal structure from motion (wit h R. Brockett). He received the National Science Foundation Career Award and the Okawa Foundation Grant. He is a Member of the Editorial Board of the Internatio nal Journal of Computer Vision (IJCV) and Foundations and Trends in Computer Gra phics and Vision. He is the founder and director of the UCLA Vision Lab; more in formation is available at http://vision.ucla. edu. ~
UC Berkeley Networking
Kristen Woyach and Pulkit Grover Last Modification Date: Wednesday, August 19, 2009
Kristen Woyach and Pulkit Grover Last Modification Date: Wednesday, August 19, 2009