My research spans Computer Vision, Machine Learning, and Graphics. I focus on computing and understanding motion in the world from video. In generic scenes I study optical flow (the 2D image motion) and how it relates to physical properties of the world including 3D shape, material, illumination, and motion. I also develop new methods to capture natural, complex, human and animal motion for applications in computer vision, animation, and neuroscience.
I am interested in computer vision and machine learning with a focus on 3D scene understanding, parsing and reconstruction. During my Ph.D. I have developed probabilistic models for 3D traffic scene understanding from movable platforms.
I am interested in the intersection between computer vision and machine learning with a focus on holistic visual scene understanding. In particular, I am interested in analyzing and modeling people in our complex visual scenes.
I am leading the Robot Perception Group at the Perceiving Systems Department. We have developed MPC-based formation-control methods to jointly perceive moving people from multiple flying robots, each equipped with a monocular camera. Real robot experimental results have not only validated our approach but have set a strong foundation for future research direction in this context. For further information, please visit my group page.
I work with computer vision researchers to coordinate, schedule and run human subjects trials involving body shape and motion analysis at the Perceiving Systems Department. To collect data we use several computer vision technologies, including our unique 3D and 4D body scanners and our new 4D face scanner.
My research focus and interest is in the area of 3D computer vision and computer graphics. I am especially interested in non-rigid shape analysis, statistical modelling of various kinds of shapes, and the analysis of motion data.
As a research engineer at the Perceiving Systems Department, I work alongside researchers in the field of computer graphics to develop basic tools that would help them with their projects. On a daily basis, I interact with large human motion datasets consisting of human motion capture and 3/4D scans, deep neural networks, and articulated human body models.
For my PhD I worked with Juergen Gall on Hand-Object Interaction. In particular we focused on capturing the motion of hands interacting with each other and/or with a rigid or an articulated object. We further studied the case of acquiring missing knowledge about the manipulated object, i.e. its shape or its kinematic model.