[4eyes] Zygmunt Pizlo talk - Monday, November 30, 3:30pm

[Matthew Turk]

CS Colloquium


 


Speaker: Zygmunt Pizlo <http://www1.psych.purdue.edu/~zpizlo/> , Purdue University 


Monday, November 30, 2015, 3:30pm to 4:30pm

Harold Frank Hall 1132

 

Title: The role of symmetry in 3D vision: psychophysics and computational modeling

Host: Sven Dickinson

 


Abstract


All animals are mirror symmetrical: birds, fish, cats, elephants, humans, horses, pigs, and insects. Animal bodies are symmetrical because of the way the animals move. Plants are symmetrical because of the way they grow. Man-made objects are symmetrical because of the function they serve. Completely asymmetrical objects are dysfunctional. The ubiquity of symmetry in our natural environment makes it a perfect candidate for being an effective prior. Is it? How much will 3D shape recovery benefit from symmetry? A lot. Symmetry provides the solution for recovering accurately 3D shapes from 2D images, arguably the most difficult problem in vision. Does the human visual system use symmetry as a prior? Is the resulting percept veridical? Does the percept deteriorate when symmetry is removed? Is there an alternative way to produce 3D veridical percepts? The answers to these questions will be presented and discussed in the context of results of psychophysical experiments and computational modeling. Do humans update the symmetry prior through experience? Should they? Can a symmetry prior be overcome when the observer views 3D shapes with both eyes? I will present our psychophysical results on monocular and binocular 3D shape recovery and compare them with a Bayesian model in which the visual data is combined with priors. In this model, the priors supplement, rather than conflict with the visual information. I will conclude the talk by showing how symmetry of objects can be used to solve the figure-ground organization problem which refers to detecting objects in the 3D scene and in the retinal or camera images.

 


Bio


Zygmunt Pizlo is a Professor of Psychology at Purdue University. He received his PhD degree in electronic engineering in 1982 from the Institute of Electron Technology in Warsaw, Poland, and a PhD in psychology from the University of Maryland at College Park, in 1991. In 1991 he accepted the position of an Assistant Professor at Purdue where he has been ever since. Pizlo’s research has always revolved around 3D shape perception. He published a history of the subject in his 2008 book (MIT Press). This book was followed by the 2014 book (Oxford University Press) summarizing his work on 3D shape recovery based on symmetry. In 2008-2011 he co-organized, with Sven Dickinson, a series of workshops on shape perception in human and computer vision that resulted in an edited book (Springer, 2013). Pizlo also worked on a number of other visual mechanisms such as figure-ground organization, multiresolution processing, binocular vision, motion and size perception, eye movements, 3D scene perception, speed-accuracy tradeoff, as well as on cognitive mechanisms involved in solving combinatorial optimization problems. 

 

Everyone welcome!

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cs.ucsb.edu/pipermail/ilab-users/attachments/20151103/f3a68f64/attachment.html>