Friends told me to get a new and more informative website. My old, dated web site is still available.

The topic of my PhD is writer identification and verification using online and offline data. The basis of my classifiers are Hidden Markov Models (HMMs) and Gaussian Mixture Models (GMMs). The HMMs are implemented using the HTK Toolkit, the GMMs by the Torch Library. Some of my Python scripts are briefly introduced in these tutorials. Furthermore, I am responsible for the IAM Handwriting Database which we use to train and test our systems.

Before starting a PhD I have written my master's thesis about using Java Byte Code to extend Piccola, a small, pure and general composition language based on the Pi-Calculus. My computer science project adds XMI/MOF support to Moose using Smalltalk.