LoginWelcome to the Vision and Graphics Lab homepage!
VGL is a new initiative to bring together researcher from Computer Vision and Graphics that englobes 5 research groups from the Technical University in Munich.
| 
| 
|
| 
|
VGL aims at building and strengthening the synergies between the five groups at TUM in a joint lab environment, pushing forward research in both the Vision and Graphics domains with a special interest in Machine Learning, Optimization and Visualization.
Research Groups
D. Cremers - Computer Vision Prof. Cremers has been working on the interface of computer vision, optimization, and machine learning. His research spans a spectrum from camera-based 3D reconstruction using both color and RGB-D cameras over convex relaxation methods for non-convex and non-smooth optimization problems in image and video analysis to 3D shape analysis and autonomous systems. He was involved in two excellency clusters, has pursued an Emmy Noether research group and three ERC funded projects, namely the Starting Grant Convex Vision, the Proof of Concept Grant CopyMe3D, and the Consolidator Grant 3D Reloaded. In 2016, he was awarded the Gottfried-Wilhelm Leibniz Award. One of the recurring themes in his reserach is the modeling and use of prior knowledge in computer vision.
L. Leal-Taixé - Dynamic Vision and Learning Prof. Leal-Taixé's research is focused on video and motion analysis, with special interest in recurrent deep architectures. The PI and her team focus on large-scale pedestrian motion analysis and behavior prediction as part of the project SocialMaps, winner of the Sofja Kovalevskaja Award by the Humboldt Foundation in 2017. Prof. Leal-Taixé's also has expertise on using social interactions to improve multiple object tracking, both by encoding physics-based models as well as more data-driven approaches. In addition her group exhibits a special focus in recurrent deep architectures, providing the necessary tools to analyze both the static as well as the dynamic world.
M. Niessner - Visual Computing Prof. Nie{\ss}ner works at the intersection of computer graphics, vision, and machine learning. His team focuses on the 3D reconstruction and understanding of real-world environments while leveraging modern discriminative and generative deep learning approaches. For instance, as part of the Google Faculty Award for Machine Perception, the team develops novel learning methods for semantic segmentation of 3D scans, and aims to turn incomplete and noisy RGB-D scans into high-quality computer graphics models.
N. Thuerey - Physics-based Simulation The focus of Prof. Thuerey's group is to develop numerical methods for physics simulations. A particular emphasis lies on simulating fluid flow and Navier-Stokes problems with deep learning techniques. Supported by an ERC Starting Grant, the goal of Prof. Thuerey's group is to combine neural networks, be it convolutional, recurrent, or adversarial, with physical knowledge and priors. Such methods have significant potential for efficient and robust simulations, e.g., in the context of visual effects and digital games, but also in engineering or medical areas.
R. Westermann - Computer Graphics and Visualization Prof. Westermann focuses on visual analytics solutions for the massive amounts of data. In the scope of the ERC AdG SaferVis (2012-2017), the Prof. Westermann and his team have developed a number of new approaches to support the visual exploration of the uncertainty that is present in measurements and numerical simulations. Much of the research has been devoted to the combination of statistical data analysis techniques and interactive visualization techniques for multi-faceted and ensemble data, and the design and realization of high-throughput computational schemes on parallel GPU.
