[Colloquium] Talk at TTIC: Iasonas Kokkinos (Ecole Centrale Paris, France)

[George Papandreou]

 When:     Friday, June 20, at 11 am

Where:    TTIC, 6045 S Kenwood Avenue, 5th Floor, Room #5
​26​

Who:       Iasonas Kokkinos (Ecole Centrale Paris, France)

Title:       Fast and Exact Algorithms for Object Detection with Deformable
Models

Abstract:

Deformable Part Models (DPMs) are at the core of current object recognition
research, but can come at an increased  computational cost. In this talk we
will see how optimization techniques can accelerate detection with DPMs.

In the first part of the talk we will present how Branch-and-Bound (BB)
 applies to detection with star-shaped DPMs. We will first present how to
compute bounds that accommodate part deformations and then present a
technique that uses shared mid-level parts to compute quick probabilistic
bounds to the individual part scores. We show that the combination of these
techniques yields substantial acceleration while delivering virtually
identical results.

In the second part of the talk we will see how the Alternating Direction
Method of Multipliers (ADDM) can be used to accommodate loops in the
model's graph structure. For this we decompose a loopy graph into a set of
loop-free `slaves' and use a master-slave scheme to ensure that the slave
solutions are consistent. We demonstrate substantial acceleration over Dual
Decomposition, which often failed to converge.  On a challenging medical
shape segmentation benchmark our ADMM-based technique yields substantially
better results than the previous state-of-the-art, while converging in
typically less than 10 iterations.

References:
H. Boussaid and I. Kokkinos, Fast and Exact: ADMM-Based Discriminative
Shape Segmentation with Loopy Part Models, CVPR 2014

I. Kokkinos, Shufflets: Shared Mid-level Parts for Fast Object Detection,
ICCV 2013

I. Kokkinos, Rapid Deformable Object Detection using Dual-Tree
Branch-and-Bound, NIPS 2011.


Bio:
Iasonas Kokkinos obtained the Diploma of Engineering in 2001 and the Ph.D.
Degree in 2006 from the School of Electrical and Computer Engineering of
the National Technical University of Athens in Greece, and the Habilitation
in 2013 from Universite Paris-Est.

In 2006 he joined the Center for Image and Vision Sciences in the
University of California at Los Angeles as a postdoctoral scholar. As of
2008 he is an Assistant Professor at the Department of Applied Mathematics
of Ecole Centrale Paris and INRIA-Saclay in Paris.
His research activity is focused on efficient algorithms for object
detection, shape-based object recognition and learning-based approaches to
feature detection.

He has been awarded a young researcher grant by the French National
Research Agency, and serves regularly as a reviewer for all major computer
vision conferences and journals; he has served as an area chair for CVPR
2012, co-organized POCV 2012 and is an associate editor for the Image and
Vision Computing Journal.

http://cvn.ecp.fr/personnel/iasonas/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20140618/f683265c/attachment.htm