Normalizing Flows for Anomaly Detection and Probabilistic Human Pose Estimation
In this seminar prof. Bodo Rosenhahn will speak about invertible neural networks (also known as Normalizing Flows) and its applications. The presentation will start with a brief overview about the Institute for Information Processing (tnt) in Hannover, recent research challenges and projects. It continues with the core concepts of invertible neural networks (INNs) and its usage for representation learning. Due to its inherent property of invertibility, he will show its potential on the applications of defect detection and probabilistic 3D human pose lifting from 2D image data.
The presentation is based on the former works
- Fully Convolutional Cross-Scale-Flows for Image-based Defect Detection by Marco Rudolph, Tom Wehrbein, Bodo Rosenhahn and Bastian Wandt,
Winter Conference on Applications of Computer Vision (WACV), IEEE, Hawaii, USA, January 2022 - Probabilistic Monocular 3D Human Pose Estimation with Normalizing Flows by Tom Wehrbein, Marco Rudolph, Bodo Rosenhahn and Bastian Wandt,
International Conference on Computer Vision (ICCV), IEEE, October 2021 - CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild by Bastian Wandt, Marco Rudolph, Petrissa Zell, Helge Rhodin and Bodo Rosenhahn, Computer Vision and Pattern Recognition (CVPR), IEEE, Virtual, June 2021
- Same Same But DifferNet: Semi-Supervised Defect Detection with Normalizing Flows by Marco Rudolph, Bastian Wandt and Bodo Rosenhahn,
Winter Conference on Applications of Computer Vision (WACV), IEEE, Online, January 2021
About the Speaker
Prof. Bodo Rosenhahn studied Computer Science (minor subject Medicine) at the University of Kiel. He received the Dipl.-Inf. and Dr.-Ing. degrees from the University of Kiel in 1999 and 2003, respectively. From 10/2003 till 10/2005, he worked as post doc at the University of Auckland (New Zealand), funded with a scholarship from the German Research Foundation (DFG). In 11/2005-08/2008 he worked as senior researcher at the Max-Planck Insitute for Informatics in Saarbruecken. Since 09/2008 he is Full Professor at the Leibniz University Hannover, heading a group on automated image interpretation.