Per-Erik Forssén

Per-Erik Forssén

Senior Associate Professor

My research interests are in visual perception and perceptual learning for robots. Current work includes uncertainty representation, vision for action, and continuous-time modelling of 3D motion.

Projects

I am the principal investigator for the following projects:

Project funded by the ELLIIT Excellence network

Project funded by the Swedish Research Council (VR).

Scientific Output

My publications. Here is an

Other lists: the , and my .

Biography

PhD

2004 Linköping University

Postdoc

2006-2007, Laboratory for Computational Intelligence, University of British Columbia, VR fellowship

Docent

2009 Linköping University

Selected publications

Cover of publication 'Efficient Multi-frequency Phase Unwrapping Using Kernel Density Estimation'

Felix Järemo-Lawin, Per-Erik Forssén, Hannes Ovrén (2016)

Computer Vision 鈥� ECCV 2016 14th European Conference, Amsterdam, The Netherlands, October 11鈥�14, 2016, Proceedings, Part IV , s.170-185

Cover of publication 'Rolling Shutter Bundle Adjustment'

Johan Hedborg, Per-Erik Forssén, Michael Felsberg, Erik Ringaby (2012)

Cover of publication 'Efficient Video Rectification and Stabilisation for Cell-Phones'

Erik Ringaby, Per-Erik Forssén (2012)

International Journal of Computer Vision , Vol.96 , s.335-352

More publications

Ludvig Dillen, Per-Erik Forss茅n, Johan Edstedt, , IMAGE ANALYSIS, SCIA 2025, PT I, Lecture Notes in Computer Science, pp. 324-337, SPRINGER INTERNATIONAL PUBLISHING AG (2025)

Ludvig Dill茅n, Per-Erik Forss茅n, Johan Edstedt, FACT: Multinomial Misalignment Classification for Point Cloud Registration (2025)

Shipeng Liu, Ziliang Xiong, Bastian Wandt, Per-Erik Forss茅n, , IMAGE ANALYSIS, SCIA 2025, PT I, Lecture Notes in Computer Science, pp. 276-291, SPRINGER INTERNATIONAL PUBLISHING AG (2025)

Shipeng Liu, Ziliang Xiong, Bastian Wandt, Per-Erik Forss茅n, Continuous Normalizing Flows for Uncertainty-Aware Human Pose Estimation (2025)

Shipeng Liu, Ziliang Xiong, Khac-Hoang Ngo, Per-Erik Forss茅n, MATTER: Multiscale Attention for Registration Error Regression (2025)

News

3D models from cheap video cameras

Hannes Ovrén shows in his doctoral thesis in computer vision how 3D models can be created from video films recorded with simple cameras. The research opens new possibilities for both robots and humans, not least for the police and rescue services.

福利姬