Welcome to the Visual Learning Lab Heidelberg
The Visual Learning Lab Heidelberg works at the intersection of Computer Vision and Machine Learning. This means that we take images, videos or other dense measurements and process them, or extract useful information hidden in the data. We work on a broad range of applications, such as segmentation, pose estimation, matching, tracking, detection, and scene understanding. We are excited to talk to researches from other disciplines, such as computational biology, medicine, astronomy, or environmental physics - and see how we can help them. On the methodological side we focus on deep learning, explainable machine learning, scalable optimization of graphical models, as well as training data simulation.
- Learning to Think Outside the Box: Wide-Baseline Light Field Depth Estimation with EPI-Shift [pdf] [project page]
- Neural-Guided RANSAC: Learning Where to Sample Model Hypotheses [pre-print] [project page] Expert Sample Consensus Applied to Camera Re-Localization [pre-print] [project …
- Paper accepted to IPCAI 2019. Find arxiv paper here https://arxiv.org/abs/1903.03441 Uncertainty-aware performance assessment of optical imaging modalities with invertible neural …
- Panoptic Segmentation, Alexander Kirillov, Kaiming He, Ross Girshick, Carsten Rother, Piotr Dollar, CVPR 2019 arxiv version https://arxiv.org/abs/1801.00868
- Here is a gentle introduction to our Invertible Neural Network architecture to tackle ambiguous inverse problems. [Update: Accepted at ICLR …
The Rich Scene Model (ERC Consolidator Grant)
Given a sequence of images the goal is to recover a rich, detailed representation of the 3D world, ranging from physical to semantical aspects. To achieve this we investigate new ways to combine feature learning, modelling, physical laws, and optimization in large-scale discrete-continuous-valued probabilistic graphical model.
Collaborators & Industrial Partners
We have also been collaborating with various industrial research labs - such as Microsoft Research Cambridge, Bejing and Redmond, Daimler, and Facebook Artificial Intelligence Researchers (FAIR).
We are part of the Heidelberg HCI 3rd phase, where we collaborate closely with Bosch.