My Profile

My photo
He graduated from the Dept. of Mathematics of the Aristotle University of Thessaloniki in 2001. He continued his studies at the School of Medicine of the same University until 2003, where he obtained the M.Sc. in Medical Informatics. In 2008, he obtained the Ph.D. in Informatics entitled as "Digital Processing Techniques in Speech Emotion Recognition" at the Computer Science faculty of the same University. He has been awarded the ERCIM fellowship for 2009-2011. In 2009, he was with VTT Technical Research Center of Finland working on Alzheimer's disease and Neuraly Adjusted Ventilation Assist (NAVA). In 2010-2011, he was with IAIS Fraunhofer Institute in Bonn working on Speech Analysis. From 2012 until now he is a researcher and software developer in Centre for Research and Technology Hellas (CERTH). In the 15 years of his professional career, he has experience in signal processing and statistical pattern recognition with Python and Matlab, Android development, Javascript-PHP development for WordPress, Joomla, Three.js frameworks, Augmented Reality with Layar-Wikitude frameworks, Virtual Reality with Unity3D, dance recognition with Kinect, and gesture recognition with Myo.

Saturday, October 10, 2009

Gaussian Mixture Modeling for Matlab

The Expectation-Maximization algorithm (EM) is widely used to find the parameters of a mixture of Gaussian probability density functions (pdfs) or briefly Gaussian components that fits the sample measurement vectors in maximum likelihood sense. In our work, the expectation-maximization (EM) algorithm for Gaussian mixture modeling is improved via three statistical tests: a) A multivariate normality test, b) a central tendency (kurtosis) criterion, and c) a test based on marginal cdf to find a discriminant to split a non-Gaussian component. Details of the method can be found in journal paper [3]. The method can be downloaded from: Download from Matlab file exchange.

Video: dimitriosververidis.blospot.com