Anitha Kannan

Ph.D. candidate
Department of Computer Science
Probabilistic and Statistical Inference Group
University of Toronto
anitha at psi.toronto.edu

I have moved on from wonderful graduate life to an equally exciting place!

----

I am a final year PhD student working under the supervision of Prof. Brendan Frey. I also collaborate with Dr. Nebojsa Jojic at Microsoft Research, USA and with Dr. Andrew Emili at Banting and Best Institute, University of Toronto.   

My research interests are primarily in machine learning with emphasis in probabilistic models.  My research focuses on developing efficient models and algorithms for solving problems in computer vision and computational biology.

 

Research

 

      Computer vision:

Layers of appearance and motion

Epitomic analysis of appearance and shapes in images

Fast transformation invariant component analysis

    Computational biology:

Understanding relationships between gene products – mRNA and protein

Publications

 

¨      Fast transformation-invariant component analysis [pdf] [webpage]

A.Kannan, N. Jojic & B.J. Frey

Submitted to:  Intl. Journal of Computer Vision special issue: Learning for Vision and Vision for Learning

 

¨      Global survey of organ and organelle protein expression in mouse: Combined proteomic and transcriptomic Profiling [link]

       T.Kislinger(*), B.Cox(*), A.Kannan(*), C.Chung, A.Ignatchenko , M.S. Scott, A. Gramolini, Q.Morris, T.Hughes, J.Rossant, B.Frey, & A.Emili

          [(*) : joint   first authors]

       Cell. Apr 7, 2006

 

¨      Layers of appearance and deformation [webpage]

A. Kannan, N. Jojic & B.J. Frey

Tenth International Workshop on Artificial Intelligence and Statistics (AISTATS), Barbados, 2005 [pdf]

 

¨      Comprehensive survey of tissue and organelle selective protein expression in mouse: Integrated global proteomic, bioinformatic and genomic analysis

T.Kislinger(*), B.Cox(*), A.Kannan(*), C.Chung, A.Ignatchenko , M.S. Scott, A. Gramolini, Q.Morris, T.Hughes, J.Rossant, B.Frey, & A.Emili [(*) : joint first authors]

Keystone Symposia: Proteomics and Bioinformatics, Keystone Colorado, USA, 2005

 

¨      A generative model for dense optical flow in layers

A. Kannan, B.J. Frey & N. Jojic

Workshop on Spatial Coherence for Visual Motion Analysis, In conjunction with ECCV 2004 [pdf]

¨      Epitomic analysis of appearance and shape [webpage]

N.Jojic, B.Frey & A.Kannan

Proc. International Conference on Computer Vision (ICCV).  Nice, France. Oct 2003. [pdf] [Bibtex]

 

¨      Layered density models and unsupervised video analysis

B.J. Frey, N. Jojic & A. Kannan

Proc. of IEEE conference on Computer Vision and Pattern Recognition, 2003. [pdf] [Bibtex]

 

¨      Fast transformation-invariant factor analysis [webpage]

A. Kannan, N. Jojic & B.J. Frey

Advances in Neural Information Processing Systems (NIPS) 15; MIT press, MA [pdf]

 

¨      Product analysis - Learning observations as products of hidden variables

B.J. Frey, A. Kannan & N. Jojic

Advances in Neural Information Processing Systems 14; MIT press, MA [ps]

¨      Accumulator networks: Suitors of local probability propagation

B.J. Frey, & A. Kannan

Advances in Neural Information Processing Systems 13; MIT press, MA [ps]

 

Thesis:

          Product Analysis - Explaining observations as products of hidden variables
          A. Kannan
          Master's Thesis, University of Waterloo, Canada, 2001