Dr. Christian Micheloni

Seminar - Danijel Skocaj - University of Ljubljana

Subspace methods for learning and recognition

We often face high-dimensional data when modelling real world phenomena, especially when modelling visual data. However, usually these high-dimensional data lie in a much lower-dimensional space. Since low-dimensional data is significantly more convenient and efficient to handle, many methods exist that map high-dimensional data into a low-dimensional subspace, preserving certain properties of the original data. We will discuss several standard subspace methods, such as Principal component analysis, Linear Discriminant Analysis, and Canonical Correlation Analysis. We will derive these methods, and learn when and how to use them. We will also discuss how to extend the standard methods into incremental and robust versions that are more appropriate to use in real world scenarios. More information->

Special ISSUE

I'm Guest Editing together with other four colleague a special issue on "Signal Processing for Situational Awareness from Networked Sensors and Social Media" in IEEE Journal of Selected Topics in Signal Processing. CFP

ICDSC 2014

ACM/IEEE International Conference on Distributed Smart Camera is an ideal venue for multidisciplinary approaches to the them of smart camera networks. In 2014, the conference will be held in Venezia ITALY.

AVSS 2013

AVSS is the premier annual international video and signal- based surveillance conference that brings together experts from academia, industry, and government to advance theories, methods, systems, and application related to surveillance. More information->

ICDSC 2012

RAWSNETS 2011

AVSS 2011

ICIAP 2011