CS&E Profile: Vassilios Morellas

Computer Vision, Robotics, Geometric and Statistical Methods in Image Processing, Machine Learning

Ph.D. 1995, Mechanical Engineering, University of Minnesota

M.S. 1988, Mechanical Engineering, Columbia University, NY

Diploma 1983, Mechanical Engineering, National Technical University of Athens, Greece

Research

My research lies in the areas of computer vision and robotics. I am interested in integrating sensors in general and video cameras in particular to various electromechanical systems in order to enhance their autonomy. Applications areas span a broad range from security and surveillance to access control to transportation to medical domains.

In the areas of computer vision and image processing I am interested in developing solutions that guarantee robust performance for long periods of time and under unpredictable and challenging realistic conditions for robots or flying drones. One such example is to separate interesting foreground video information from the uninteresting and stable background. This is a fundamental capability for every system with embedded video camera(s) such as a security system or a reconnaissance helicopter. I am also interested in developing representation models of objects such as vehicles and humans to track their presence across a network of video cameras.

In the areas of robotics I am interested in the design of new robots that exhibit different types of mobility particularly suitable for search and rescue applications. Miniature robots with high mobility-to-size ratio and embedded sensors can shine in new domains such as aquatic environments.

In the areas of sensor data processing I am interested in statistical machine learning, compressive sensing and clustering methodologies for fast searching of very large databases describing knowledge pertaining to a wide range of domains extracted again from embedded robotic sensors.