Pablo Sala


Department of Computer Science

University of Toronto





About me


I graduated in 2002 with a Licentiate in Computer Science degree from the Departamento de Computación of the Universidad de Buenos Aires, Argentina, and a M.Sc. in Computer Science degree in 2004 from the University of Toronto.

I am currently a PhD student in computer vision in the University of Toronto under the supervision of Sven Dickinson.

In 2006, I did an internship in Microsoft Research working for Gavin Jancke. I developed robust methods to read two-dimensional barcodes under challenging imaging conditions, which is the foundation of what eventually became Microsoft Tag.



Research Interests


During my masters, I worked on the interesting problem of visual robot navigation. In my PhD, I am concentrating on the challenging problem of generic object recognition. Rather than assuming an object-level shape prior, as in the currently popular object detection paradigm, I am revisiting the classic formulation of this problem, which assumes that objects can be formed by the composition of a small number of intermediate-level shape parts recoverable from a single image through perceptual mechanisms. Armed with new segmentation techniques, new shape representations, and new machine learning algorithms, and focusing not only on the contour grouping problem but also on the contour abstraction problem, I hope to make new progress on the important problem of recovering from a 2-D image a set of qualitative 3-D volumetric parts making up the scene’s objects.




6 King’s College Rd.

Rm 265A, Pratt Building

Toronto, Ontario

Canada M5S 3G4



(416) 827-7980