DAGM MVTec Dissertation Award

Since 2015, the German Association for Pattern Recognition (DAGM e.V.) has granted the DAGM MVTec Dissertation Award. The Dissertation Award honors an outstanding dissertation successfully completed in the preceding year in the fields of pattern recognition, image processing, machine vision, and machine learning. An association with the DAGM is desirable. The Dissertation Award is donated by MVTec Software GmbH and consists of a financial award of 2500 euros.

Award Winner of the DAGM MVTec Dissertation Award 2019

In 2019, the DAGM MVTec Dissertation Award was granted to Umar Iqbal for his outstanding dissertation on

Articulated Human Pose Estimation in Unconstrained Images and Videos

Pictured left to right: Prof. Dr. Reinhard Koch (Chairman of the German Association for Pattern Recognition DAGM e.V.), Dr. Umar Iqbal, Prof. Dr. Carsten Steger (MVTec Software GmbH, TU Munich)

Dr. Umar Iqbal received his B.Sc. in Computer Engineering in 2010 from COMSATS Institute of Information Technology (CIIT), Pakistan, and his M.Sc. in Information Technology in 2013 from Tampere University of Technology, Finland. He then worked as a researcher at the University of Bonn from 2014 until 2018 under the supervision of Prof. Jürgen Gall and received his PhD degree from the University of Bonn in 2018. He currently is a research scientist at NVIDIA Research, USA.
In his thesis, entitled “Articulated Human Pose Estimation in Unconstrained Images and Videos,” Umar Iqbal has substantially advanced the field of 2D and 3D articulated pose estimation from RGB images or videos. The thesis has not only shown how the 3D hand pose can be efficiently and accurately estimated from a single image. It also has unified two approaches that were separate so far, human pose estimation and multi-person tracking, in a single framework. In addition to these outstanding achievements, the PoseTrack benchmark, proposed in the thesis, is an important contribution to the computer vision and pattern recognition community.

 

Award Winners of the DAGM MVTec Dissertation Award 2018

In 2018, the DAGM MVTec Dissertation Award was granted to Dr. Siyu Tang and Dr. Mateusz Malinowski.

Pictured left to right: Prof. Dr. Carsten Steger (MVTec Software GmbH, TU Munich), Dr.Siyu Tang, Dr. Mateusz Malinowski, Prof. Dr. Reinhard Koch (Chairman of the German Association for Pattern Recognition DAGM)

The DAGM MVTec Dissertation Award was granted to Dr. Siyu Tang for her outstanding dissertation on

People Detection and Tracking in Crowded Scenes

Dr. Siyu Tang graduated in media informatics at RWTH Aachen University and received her Master’s degree in 2011. She then worked as a researcher at the Max Planck Institute for Informatics in Saarbrücken from 2011 until 2017 under the supervision of Prof. Bernt Schiele and received her PhD degree from Saarland University in 2017. She currently is a Research Group Leader at the Max Planck Institute for Intelligent Systems in Tübingen. In her thesis, entitled “People Detection and Tracking in Crowded Scenes,” Siyu Tang made significant contributions in the area of multi-person tracking and detection. She proposed novel and rigorous mathematical formulations based on minimum cost multicut models that perform well for challenging real-word visual scenes and thus have defined the state of the art in this area, winning the Multi-Object Tracking Challenge at ECCV 2016 and CVPR 2017.

The DAGM MVTec Dissertation Award was granted to Dr. Mateusz Malinowski for his outstanding dissertation on

Towards Holistic Machines: From Visual Recognition To Question Answering About Real-World Images

Dr. Mateusz Malinowski graduated in computer science at Saarland University and received his Master’s degree in 2011. He then worked as a researcher at the Max Planck Institute for Informatics in Saarbrücken from 2011 until 2016 under the supervision of Dr. Mario Fritz and received his PhD degree from Saarland University in 2017. He currently is a Research Scientist at Google DeepMind. In his thesis, entitled “Towards Holistic Machines: From Visual Recognition To Question Answering About Real-World Images,” Mateusz Malinowski pioneered the work on visual question answering. He created the foundations of this task, built the first dataset with natural questions about real-world images, proposed evaluation methods, and developed two first approaches for answering questions about images: one connects with more traditional fields of AI, the other one builds upon the recent progress on deep learning. The task and methods have become very influential in the computer vision community, which even led to separate tracks for visual question answering at top-tier conferences such as NIPS and CVPR.

 

Award Winner of the DAGM MVTec Dissertation Award 2017

In 2017, the DAGM MVTec Dissertation Award was granted to Alexander Freytag for his outstanding dissertation on

Lifelong Learning for Visual Recognition Systems

Pictured left to right: Prof. Dr. Carsten Steger (MVTec Software GmbH, TU Munich), Dr. Alexander Freytag, Prof. Dr. Reinhard Koch (Chairman of the German Association for Pattern Recognition DAGM)

Dr. Alexander Freytag graduated in computer science at the University of Jena and received his Diploma degree in 2011. He then worked as a researcher at the University of Jena from 2011 until 2016 under the supervision of Prof. Joachim Denzler and received his PhD degree from the University of Jena in 2016. He currently is a Research Scientist for Machine Learning and Computer Vision at Carl-Zeiss AG, Corporate Research and Technology.
In his thesis entitled “Lifelong Learning for Visual Recognition Systems”, Alexander Freytag tackled the problem of lifelong learning in a very general manner, including not just the efficient incremental learning of classifiers over time, but also the detection of unknown objects, which is called novelty detection or open set learning. His main contribution arises from the EMOC (Expected Model Output Changes) method, an original and effective optimization criterion for active learning. He was able to close the loop in lifelong learning for the first time and demonstrated the effectiveness of his approach in several, diverse applications. His thesis also received an award for the best dissertation in 2016 from the Faculty of Mathematics and Computer Science of the University of Jena.
 

Award Winner of the DAGM MVTec Dissertation Award 2016

In 2016, the DAGM MVTec Dissertation Award was granted to Dr. Christoph Vogel for his outstanding dissertation on

Robust and accurate 3D motion estimation under adverse conditions

 

Pictured left to right: Prof. Dr. Carsten Steger (MVTec Software GmbH, TU Munich), Dr. Christoph Vogel, Prof. Dr. Reinhard Koch (Chairman of the German Association for Pattern Recognition DAGM)

 

Award Winner of the DAGM MVTec Dissertation Award 2015

In 2015, the DAGM MVTec Dissertation Award was granted to Dr. Marcus Rohrbach for his outstanding dissertation on

Linguistic Knowledge for Visual Recognition and Natural Language Descriptions of Visual Content

Pictured left to right: Prof. Joachim M. Buhmann (Chairman of the German Association for Pattern Recognition DAGM), Dr. Marcus Rohrbach, Prof. Dr. Carsten Steger (MVTec Software GmbH, TU Munich)