Research

Current Projects

My research follows two somehow connected lines. One is correspondence based systems for invariant object recognition. We have built a neurally plausible system based on networks of cortical columns. The network achieves competitive results for recognition of faces in databases of more than 1000 faces.

The other direction is the modeling and analysis of information routing structures in the brain. Here we are interested in their role in automatic invariant recognition, optimal architecture design, and in the ontogenesis of such structures.

Other areas of interest

Structure of representations in the brain; active and goal-directed perception; understanding neural models in terms of Bayesian inference; computer vision & pattern recognition; system identification & dimensionality reduction; control theory; robotics; ubiquitous computing & sophisticated HMIs