§16§This book constitutes the refereed proceedings of the 42nd German Conference on Pattern Recognition, DAGM GCPR 2020, which took place during September 28 until October 1, 2020. The conference was planned to take place in Tübingen, Germany, but had to change to an online format due to the COVID-19 pandemic.§The 34 papers presented in this volume were carefully reviewed and selected from a total of 89 submissions. They were organized in topical sections named: Normalizing Flow, Semantics, Physics, Camera Calibration and Computer Vision, Pattern Recognition, Machine Learning.§ §04§§Normalizing Flow, Semantics, Physics, Camera Calibration.- Characterizing The Role of A Single Coupling Layer in Affine Normalizing Flows.- Semantic Bottlenecks: Quantifying 〓04§ Architecture Performance Prediction Using Graph Neural Networks.- Discovering Latent Classes for Semi-Supervised Semantic Segmentation.- Riemannian SOS-Polynomial Normalizing Flows.- Automated water segmentation and river level detection on camera images using transfer learning.- Does SGD Implicitly Optimize for Smoothness.- Looking outside the box: The role of context in Random Forest based semantic segmentation of PolSAR images.- Haar Wavelet based Block Autoregressive Flows for Trajectories.- Center3D: Center-based Monocular 3D Object Detection with Joint Depth Understanding.- Constellation Codebooks for Reliable Vehicle Localization.- Towards Bounding-Box Free Panoptic Segmentation.- Proposal-Free Volumetric Instance Segmentation from Latent Single-Instance Masks.- Unsupervised Part Discovery by Unsupervised Disentanglement.- On the Lifted Multicut Polytope for Trees.- Conditional Invertible Neural Networks for Diverse Image-to-Image Translation.- Image Inpainting with Learna §04§ble Feature Imputation.- 4Seasons: A Cross-Season Dataset for Multi-Weather SLAM in Autonomous Driving.- Inline Double Layer Depth Estimation with Transparent Materials.- A Differentiable Convolutional Distance Transform Layer for Improved Image Segmentation.- PET-guided Attention Network for Segmentation of Lung Tumors from PET/CT images.- Self-supervised Disentanglement of Modality-specific and Shared Factors Improves Multimodal Generative Models.- Multimodal semantic forecasting based on conditional generation of future features.§
