DAGM GCPR | 2021
DAGM German Conference on Pattern Recognition, Bonn
Talks in Sessions
Thursday, 30.09.
Computational Photography and Lighting
10:00-10:20
A Comparative Study of PnP and Learning Approaches to Super-Resolution in a Real-World Setting
Samim Zahoor Taray (K|Lens GmbH), Sunil Prasad Jaiswal (K|Lens GmbH), Shivam Sharma (K|Lens GmbH), Noshaba Cheema (Saarland Informatics Campus, DFKI); Klaus M Illgner-Fehns (K|Lens GmbH), Philipp Slusallek (DFKI, Saarland University), Ivo Ihrke (K|Lens GmbH, University of Siegen)
10:20-10:40
Merging-ISP: Multi-Exposure High Dynamic Range Image Signal Processing
Prashant Chaudhari (e.solutions GmbH), Franziska Schirrmacher (Friedrich-Alexander-Universität Erlangen-Nürnberg), Andreas Maier (Friedrich-Alexander-Universität Erlangen-Nürnberg), Christian Riess (Friedrich-Alexander-Universität Erlangen-Nürnberg), Thomas Köhler (e.solutions GmbH)
10:40-11:00
Spatiotemporal Outdoor Lighting Aggregation on Image Sequences
Haebom Lee (University of Heidelberg, Bosch), Robert Herzog (Bosch), Jan Rexilius (Fachhochschule Bielefeld), Carsten Rother (University of Heidelberg)
Vision Systems and Applications
11:00-11:12
Viewpoint-Tolerant Semantic Segmentation for Aerial Logistics
Shiming Wang (RWTH Aachen University), Fabiola Maffra (ETH Zurich), Ruben Mascaro (ETH Zurich), Lucas Teixeira (ETH Zurich), Margarita Chli (ETH Zurich)
11:12-11:24
T6D-Direct: Transformers for Multi-Object 6D Pose Direct Regression
Arash Amini (University of Bonn), Arul Selvam Periyasamy (University of Bonn), Sven Behnke (University of Bonn)
12:24-11:36
TetraPackNet: Four-Corner-Based Object Detection in Logistics Use-Cases
Laura Dörr (FZI Research Center for Information Technology), Felix Brandt (FZI Research Center for Information Technology), Alexander Naumann (FZI Research Center for Information Technology), Martin Pouls (FZI Research Center for Information Technology)
11:36-11:48
Detecting Slag Formations with Deep Convolutional Neural Networks
Christian Von Koch (Combient Mix AB), William Anzén (Combient Mix AB), Max Fischer (Combient Mix AB), Raazesh Sainudiin (Uppsala University)
11:48-12:00
Virtual Temporal Samples for Recurrent Neural Networks: applied to semantic segmentation in agriculture
Alireza Ahmadi (University of Bonn), Michael Halstead (University of Bonn), Chris McCool (University of Bonn)
Machine Learning and Optimization
13:00-13:20
Sublabel-Accurate Multilabeling Meets Product Label Spaces
Zhenzhang Ye (TU Munich), Bjoern Haefner (TU Munich), Yvain Quéau (CNRS), Thomas Möllenhoff (RIKEN Center for Advanced Intelligence Project), Daniel Cremers (TU Munich)
13:20-13:40
InfoSeg: Unsupervised Semantic Image Segmentation with Mutual Information Maximization
Robert Harb (Graz University of Technology), Patrick Knöbelreiter (Graz University of Technology)
13:40-14:00
Sampling-free Variational Inference for Neural Networks with Multiplicative Activation Noise
Jannik Schmitt (TU Darmstadt), Stefan Roth (TU Darmstadt)
14:00-14:12
Reintroducing Straight-Through Estimators as Principled Methods for Stochastic Binary Networks
Alexander Shekhovtsov (Czech Technical University in Prague), Viktor Yanush (HSE University, Moscow)
14:12-14:24
Bias-Variance Tradeoffs in Single-Sample Binary Gradient Estimators
Alexander Shekhovtsov (Czech Technical University in Prague)
14:24-14:36
End-to-end Learning of Fisher Vector Encodings for Part Features in Fine-grained Recognition
Dimitri Korsch (Friedrich-Schiller-University Jena), Paul Bodesheim (Friedrich-Schiller-University Jena), Joachim Denzler (Friedrich-Schiller-University Jena, DLR)
14:36-14:48
Investigating the Consistency of Uncertainty Sampling in Deep Active Learning
Niklas Penzel (Friedrich-Schiller-University Jena), Christian Reimers (Friedrich-Schiller-University Jena), Clemens A Brust (Friedrich-Schiller-University Jena), Joachim Denzler (Friedrich-Schiller-University Jena, DLR)
14:48-15:00
ScaleNet: An Unsupervised Representation Learning Method for Limited Information
Huili Huang (Georgia Institute of Technology), M. Mahdi Roozbahani (Georgia Institute of Technology)
Action Recognition and Video Understanding
15:15-15:35
A New Split for Evaluating True Zero-Shot Action Recognition
Shreyank N Gowda (University of Edinburgh), Laura Sevilla-Lara (Facebook AI Research), Kiyoon Kim (University of Edinburgh), Frank Keller (University of Edinburgh), Marcus Rohrbach (Facebook AI Research)
15:35-15:55
Video Instance Segmentation with Recurrent Graph Neural Networks
Joakim Johnander (Linköping University), Emil Brissman (Linköping University), Martin Danelljan (ETH Zurich), Michael Felsberg (Linköping University)
15:55-16:15
Distractor-Aware Video Object Segmentation
Andreas Robinson (Linköping University), Abdelrahman Eldesokey (Linköping University), Michael Felsberg (Linköping University)
Actions, Events, and Segmentation
16:15-16:27
(SP)²Net for Generalized Zero-Label Semantic Segmentation
Anurag Das (MPI for Informatics, Saarland Informatics Campus), Yongqin Xian (ETH Zurich), Yang He (Amazon), Bernt Schiele (MPI for Informatics, Saarland Informatics Campus), Zeynep Akata (MPI for Intelligent Systems, University of Tübingen)
16:27-16:39
Contrastive Representation Learning for Hand Shape Estimation
Christian Zimmermann (University of Freiburg), Max Argus (University of Freiburg), Thomas Brox (University of Freiburg)
16:39-16:51
Fusion-GCN: Multimodal Action Recognition using Graph Convolutional Networks
Michael Duhme (University of Koblenz-Landau), Raphael Memmesheimer (University of Koblenz-Landau), Dietrich Paulus (University of Koblenz-Landau)
16:51-17:03
FIFA: Fast Inference Approximation for Action Segmentation
Yaser Souri (University of Bonn), Yazan Abu Farha (University of Bonn), Fabien Despinoy (Toyota Motor Europe), Gianpiero Francesca (Toyota Motor Europe), Juergen Gall (University of Bonn)
17:03-17:15
Hybrid SNN-ANN: Energy-Efficient Classification and Object Detection for Event-Based Vision
Alexander Kugele (Bosch Center for Artificial Intelligence), Thomas Pfeil (Bosch Center for Artificial Intelligence), Michael Pfeiffer (Bosch Center for Artificial Intelligence), Elisabetta Chicca (University of Groningen)
Pattern Recognition in the Life- and Natural Sciences
17:15-17:35
Weakly Supervised Segmentation Pretraining for Plant Cover Prediction
Matthias Körschens (Friedrich Schiller University Jena), Paul Bodesheim (Friedrich Schiller University Jena), Christine Römermann (Friedrich Schiller University Jena), Solveig Franziska Bucher (Friedrich Schiller University Jena), Mirco Migliavacca (Max Planck Institute for Biogeochemistry Jena), Josephine Ulrich (Friedrich Schiller University Jena), Joachim Denzler (Friedrich Schiller University Jena)
17:35-17:55
How Reliable Are Out-of-Distribution Generalization Methods for Medical Image Segmentation?
Antoine Sanner (TU Darmstadt), Camila Gonzalez (TU Darmstadt), Anirban Mukhopadhyay (TU Darmstadt)
Friday, 1.10.
10:00-10:20
Conditional Adversarial Debiasing: Towards Learning Unbiased Classifiers from Biased Data
Christian Reimers (Friedrich-Schiller-University Jena, DLR), Paul Bodesheim (Friedrich-Schiller-University Jena), Jakob Runge (DLR, TU Berlin); Joachim Denzler (Friedrich-Schiller-University Jena, DLR)
10:20-10:40
Revisiting Consistency Regularization for Semi-Supervised Learning
Yue Fan (Max Planck Institute for Informatics), Anna Kukleva (Max Planck Institute for Informatics), Bernt Schiele (Max Planck Institute for Informatics)
10:40-11:00
Learning Robust Models Using the Principle of Independent Causal Mechanisms
Jens Müller (University of Heidelberg), Robert Schmier (University of Heidelberg, Bosch Center for Artificial Intelligence), Lynton Ardizzone (University of Heidelberg), Carsten Rother (University of Heidelberg), Ullrich Köthe (University of Heidelberg)
Generative Models and Multimodal Data
11:00-11:12
AttrLostGAN: Attribute Controlled Image Synthesis from Reconfigurable Layout and Style
Stanislav Frolov (Technical University of Kaiserslautern, DFKI), Avneesh Sharma (Technical University of Kaiserslautern), Jörn Hees (DFKI), Tushar Karayil (DFKI), Federico Raue (DFKI), Andreas Dengel (DFKI)
11:12-11:24
Learning Conditional Invariance through Cycle Consistency
Maxim Samarin (University of Basel), Vitali Nesterov (University of Basel), Mario Wieser (University of Basel), Aleksander Wieczorek (University of Basel), Sonali Parbhoo (Harvard University), Volker Roth (University of Basel)
12:24-11:36
CAGAN: Text-To-Image Generation with Combined Attention Generative Adversarial Networks
Henning Schulze (Karlsruhe Institute of Technology), Dogucan Yaman (Karlsruhe Institute of Technology), Alexander Waibel (Karlsruhe Institute of Technology)
11:36-11:48
TxT: Crossmodal End-to-End Learning with Transformers
Jan-Martin O. Steitz (TU Darmstadt), Jonas Pfeiffer (TU Darmstadt), Iryna Gurevych (TU Darmstadt), Stefan Roth (TU Darmstadt)
11:48-12:00
Diverse Image Captioning with Grounded Style
Franz Klein (TU Darmstadt), Shweta Mahajan (TU Darmstadt), Stefan Roth (TU Darmstadt)
Photogrammetry and Remote Sensing
13:00-13:20
Clustering Persistent Scatterer Points Based on a Hybrid Distance Metric
Philipp J Schneider (University of Stuttgart), Uwe Sörgel (University of Stuttgart)
13:20-13:40
CATEGORISE: An Automated Framework for Utilizing the Workforce of the Crowd for Semantic Segmentation of 3D Point Clouds
Michael Kölle (University of Stuttgart), Volker Walter (University of Stuttgart), Ivan Shiller (University of Stuttgart), Uwe Sörgel (University of Stuttgart)
13:40-14:00
Zero-Shot remote sensing image super resolution based on image continuity and self tessellations
Rupak Bose (IIT Bombay), Vikrant Rangnekar (IIT Bombay), Biplab Banerjee (IIT Bombay), Subhasis Chaudhuri (IIT Bombay)
3D Modeling and Reconstruction
14:00-14:12
A Comparative Survey of Geometric Light Source Calibration Methods
Mariya Kaisheva (Bauhaus-Universität Weimar), Volker Rodehorst (Bauhaus-Universität Weimar)
14:12-14:24
Quantifying point cloud realism through adversarially learned latent representations
Larissa T Triess (Mercedes-Benz AG), David Peter (Mercedes Benz AG), Stefan A Baur (Mercedes Benz AG), J. Marius Zöllner (Karlsruhe Institute of Technology)
14:24-14:36
Full-Glow: Fully conditional Glow for more realistic image generation
Moein Sorkhei (KTH Royal Institute of Technology), Gustav Eje Henter (KTH Royal Institute of Technology), Hedvig Kjellström (KTH Royal Institute of Technology)
14:36-14:48
Multidirectional Conjugate Gradients for Scalable Bundle Adjustment
Simon Weber (Technical University of Munich), Nikolaus Demmel (Technical University of Munich), Daniel Cremers (Technical University of Munich)
Labeling and Self-Supervised Learning
15:15-15:27
Leveraging Group Annotations in Object Detection Using Graph-Based Pseudo-Labeling
Daniel Pototzky (Robert Bosch GmbH), Matthias Kirschner (Robert Bosch GmbH), Lars Schmidt-Thieme (Universität Hildesheim)
15:27-15:39
Quantifying Uncertainty of Image Labelings Using Assignment Flows
Daniel Gonzalez (Heidelberg University), Alexander Zeilmann (Heidelberg University), Christoph Schnörr (Heidelberg University)
15:39-15:51
Implicit and Explicit Attention for Zero-Shot Learning
Faisal Alamri (University of Exeter), Anjan Dutta (University of Exeter)
15:51-16:03
Self-Supervised Learning for Object Detection in Autonomous Driving
Daniel Pototzky (Robert Bosch GmbH), Azhar Sultan (Robert Bosch GmbH), Matthias Kirschner (Robert Bosch GmbH), Lars Schmidt-Thieme (Universität Hildesheim)
16:03-16:15
Assignment Flows and Nonlocal PDEs on Graphs
Dmitrij Sitenko (Heidelberg University), Bastian Boll (Heidelberg University), Christoph Schnörr (Heidelberg University)