This is my homepage at the Institute of Computer Graphics and Vision (ICG), Graz University of Technology, Austria. Since April 2006, I'm a PhD student and research assistant at ICG and my supervisor is Horst Bischof. My current research is focused mainly on pattern recognition, machine learning, semi-supervised and unsupervised learning methods, object recognition and categorization.

I have another website which I maintain and update it regularly, so for latest news please visit YMER.ORG.

News:

Publications:

Recent Publications:

• Isabelle Guyon, Amir Saffari, Gideon Dror, Gavin Cawley, "Model Selection: Beyond the Bayesian/Frequentist Divide", Journal of Machine Learning Research (JMLR), Vol. 11, Pages 61-87, 2010.
• Amir Saffari, Christian Leistner, Horst Bischof, "Regularized Multi-Class Semi-Supervised Boosting", Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Oral Presentation, 2009. Slides.
• Christian Leistner, Amir Saffari, Jakob Santner, Horst Bischof, "Semi-Supervised Random Forests", Proceedings of IEEE International Conference on Computer Vision (ICCV), 2009.
• Amir Saffari, Christian Leistner, Jakob Santner, Martin Godec, Horst Bischof, "On-line Random Forests", 3rd IEEE ICCV Workshop on On-line Learning for Computer Vision, 2009. Slides. Code.
• Christian Leistner, Amir Saffari, Peter Roth, Horst Bischof, "On Robustness of On-line Boosting - A Competitive Study", 3rd IEEE ICCV Workshop on On-line Learning for Computer Vision, 2009.
• Amir Saffari, Helmut Grabner, Horst Bischof, "SERBoost: Semi-supervised Boosting with Expectation Regularization", Proceedings of European Conference on Computer Vision (ECCV), 2008.
• Amir Saffari, Horst Bischof, "Boosting for Model-Based Data Clustering", Proc. of 30th Symposium of the German Association for Pattern Recognition (DAGM 2008), Oral Presentation, Pages 51-60, 2008.

Softwares:

CLOP: CLOP is a software package containing several ready-to-use machine learning algorithms.

BNN: A Matlab toolbox for small-scale biological neural networks simulations.

RSOM: This is package contains useful function to simulate and train Recurrent Self-Organizing Maps (RSOM) under MATLAB.

Robust Hue Histogram: C++ implementation of Robust Hue Histogram color descriptors.

Online Random Forests: This package is C++ implementation of the “Online Random Forests” (ORF) algorithm.