Pia Bideau: Different representations of motion information - and what can we learn from those?

Speaker: Pia Bideau (TU Berlin) people.cs.umas...
Title: Different representations of motion information - and what can we learn from those?
Abstract: The ability to recognize motion is one of the most important functions of our visual system. Motion allows us both to recognize objects and to get a better understanding of the 3D world we are moving in. Because of its importance, motion is used to answer a wide variety of fundamental questions in computer vision such as: (1) Which objects are moving independently in the world? (2) Which objects are close and which objects are far away? (3) How is the camera moving?
During my PhD I have been working on moving object segmentation in unconstrained videos. In the first part of this talk I will talk about a probabilistic approach to recognize and segment independently moving objects from optical flow, connecting aspects of camera motion estimation, relative depth and flow statistics.
But what is actually motion? Here at science of intelligence (TU Berlin) I got into a new exciting research area, which is event-based vision. Other than earlier work in computer vision that estimates motion from video frames (e.g. optical flow), event cameras directly capture low-level motion information as brightness changes. In the second part I will talk about recent work on event alignment using a spatio-temporal Poisson-Point Process.
Bio: Pia Bideau is a postdoctoral researcher at TU Berlin and part of the excellence Cluster Science of Intelligence as well as the RBO Lab. Her major background is in computer vision. Pia did her PhD at the University of Massachusetts in Amherst (USA) working together with Prof. Erik Learned-Miller. During her PhD she also worked together with Cordelia Schmid and Karteek Alahari as part of an internship at Inria in Grenoble (France). She got her M.Sc. in Electrical Engineering and Information Technology from Ruhr-Universiy Bochum. Before that she successfully completed a dual education at Siemens combining a B.Sc in Electrical Engineering and an apprenticeship as a technician for automation technology.
Talk is jointly organized by the Intelligent Systems Lab (argmin.lis.tu-b...) and the Robotics and Biology Lab (www.robotics.tu...) at the TU Berlin.
Moderated by Andreas Orthey (aorthey.de/, / andreas_orthey ) and Marc Toussaint (www.marc-toussa....
Find us on twitter at / lis_tuberlin and / rbolab