Pia Bideau

Pia Bideau

Postdoctoral Researcher
Science of Intelligence
Technical University Berlin

Email: p.bideau AT tu-berlin.de
google scholar

About Me

I am a Postdoctoral Researcher at TU Berlin and part of the research cluster Science of Intelligence and the RBO lab. My research lies at the intersection of computer vision and robotics. Personally, I am interested in developing visual intelligence that allows agents to perceive and understand their environment. To this end, my work makes use of known physical information about the real world to enable learning systems to tackle changes and variation in the environment, while minimizing human supervision. Thinking of future, I am hoping agents will be able to safely interact with their environment and peers.

In December 2019 I graduated from University of Massachusetts, Amherst, where I was lucky to be advised by Prof. Erik Learned-Miller. In 2018 I did an internship at INRIA, Grenoble working with Karteek Alahari and Cordelia Schmid. I got my M.Sc degree from Ruhr-University Bochum.

"Man muss das Unmögliche versuchen, um das Mögliche zu erreichen." - (Hermann Hesse)
"You must try the impossible to reach the possible." - (Hermann Hesse)



"Action-based Contrastive Learning for Trajectory Prediction
Marah Halawa, Olaf Hellwich, Pia Bideau, in ECCV, 2022


"The Spatio-Temporal Poisson Point Process: A Simple Model for the Alignment of Event Camera Data
Cheng Gu, Erik Learned-Miller, Daniel Sheldon, Guillermo Gallego, Pia Bideau, in ICCV, 2021

project page code

"C-14: Assured Timestamps for Drone Videos
Zhipeng Tang, Fabien Delattre, Pia Bideau, Mark D. Corner, Erik Learned-Miller, in The 26th Annual International Conference on Mobile Computing and Networking (MobiCom), 2020


Best workshop paper
"MoA-Net: Self-Supervised Motion Segmentation
Pia Bideau, Rakesh Menon, Erik Learned-Miller, in ECCV workshop: what is optical flow for?, 2018

paper slides (pdf)

"The best of both worlds: Combining CNNs and geometric constraints for hierarchichal motion segmentation
Pia Bideau, Aruni RoyChowdhury, Rakesh Menon, Erik Learned-Miller, in CVPR, 2018

project page code evaluation code

"A Detailed Rubric for Motion Segmentation"
Pia Bideau, Erik Learned-Miller, ArXiv preprint, 2016

project page

"It's Moving! A Probabilistic Model for Causal Motion Segmentation in Moving Camera Videos"
Pia Bideau, Erik Learned-Miller, in ECCV, 2016

project page


X-Student Research Group

Active Perception - estimating depth from motion, WS22/23

course website NEWs



"Learning for and from motion" Scientific Symposium - Max Planck Institut for Intelligent Systems, Stuttgart
symposium schedule


"Different representations of motion information - and what can we learn from those?" Robotics Colloquium - Learning and Intelligent Systems Lab, TU Berlin


"MoA-Net: Self-Supervised Motion Segmentation" Pia Bideau, Rakesh R Menon, Erik Learned-Miller, Workshop: What is optical flow for?, ECCV2018
slides with example videos slides (pdf)


"Causal Motion Segmentation in Moving Camera Videos" Pia Bideau, Erik Learned-Miller, The Second International Workshop on Video Segmentation, ECCV2016
slides with example videos slides (pdf)