Biomedical Image Registration: 10th International Workshop, WBIR 2022, Munich, Germany, July 10–12, 2022, Proceedings (Lecture Notes in Computer Science, 13386)

This book constitutes the refereed proceedings of the 10th International Workshop on Biomedical Image Registration, WBIR 2020, which was supposed to be held in Munich, Germany, in July 2022. The 11 full and poster papers together with 17 short papers included in this volume were carefully reviewed and selected from 32 submitted papers. The papers are organized in the following topical sections: optimization, deep learning architectures, neuroimaging, diffeomorphisms, uncertainty, topology and metrics. Preface Organization Contents Atlases/Topology Unsupervised Non-correspondence Detection in Medical Images Using an Image Registration Convolutional Neural Network 1 Introduction 2 Materials and Methods 3 Results 4 Discussion References Weighted Metamorphosis for Registration of Images with Different Topologies 1 Introduction 2 Methods 3 Results and Perspectives References LDDMM Meets GANs: Generative Adversarial Networks for Diffeomorphic Registration 1 Introduction 2 Background on LDDMM 3 Generative Adversarial Networks for LDDMM 3.1 Proposed GAN Architecture 4 Experiments and Results 4.1 Results in the 3D NIREP Dataset 5 Conclusions References Towards a 4D Spatio-Temporal Atlas of the Embryonic and Fetal Brain Using a Deep Learning Approach for Groupwise Image Registration 1 Introduction 2 Method 3 Data and Experiments 4 Results 5 Discussion and Conclusion References Uncertainty DeepSTAPLE: Learning to Predict Multimodal Registration Quality for Unsupervised Domain Adaptation 1 Introduction 2 Method 3 Experiments 4 Results and Discussion 5 Conclusion and Outlook References A Method for Image Registration via Broken Geodesics 1 Introduction 1.1 Background 2 Method 2.1 Registration Metric Approximation 3 Results 3.1 Visual Assessment of Registration 3.2 Quantitative Assessment of Registration 3.3 Validation of Proposed Registration Metric 4 Discussion References Deformable Image Registration Uncertainty Quantification Using Deep Learning for Dose Accumulation in Adaptive Proton Therapy 1 Introduction 2 Methods 2.1 Probabilistic VoxelMorph 2.2 Combining Deep Learning with Existing DIR Software 2.3 Non-diagonal Covariance Matrix 2.4 Training 2.5 Validation 3 Results 3.1 Hyper Parameter Tuning 3.2 Dose Deformation 4 Discussion 5 Conclusion References Distinct Structural Patterns of the Human Brain: A Caveat for Registration 1 Introduction 2 Methods 3 Results 4 Conclusion References Architectures A Multi-organ Point Cloud Registration Algorithm for Abdominal CT Registration 1 Introduction 2 Method 3 Experiments 4 Conclusion References Voxelmorph++ 1 Introduction 2 Method 3 Experiments and Results 4 Discussion and Conclusions References Unsupervised Learning of Diffeomorphic Image Registration via TransMorph 1 Introduction 2 Background on LDDMM 3 Methods 4 Experiments and Results 5 Conclusion References SuperWarp: Supervised Learning and Warping on U-Net for Invariant Subvoxel-Precise Registration 1 Introduction 2 Related Work 2.1 Optical Flow Estimation 3 Mathematical Framework 3.1 Optical Flow Estimation and Duality Principle 3.2 Supervised Learning and Multi-scale Warping 3.3 Deep Supervision, Data Augmentation and Training 4 Experimental Evaluation 4.1 Invariant Registration of Brain MR Images 4.2 Optical Flow Estimation 5 Discussion 5.1 Future Work References Optimisation Learn to Fuse Input Features for Large-Deformation Registration with Differentiable Convex-Discrete Optimisation 1 Introduction and Related Work 2 Method 2.1 Differentiable Convex-Discrete Optimisation 2.2 Learning of Input Feature Fusion 3 Experiments and Results 4 Discussion and Conclusion References Multi-magnification Networks for Deformable Image Registration on Histopathology Images 1 Introduction 2 Methods 2.1 Network Architectures 2.2 Loss Function 3 Experiments 3.1 Experimental Settings 3.2 Evaluation Metrics 4 Results 5 Discussion and Conclusion References Realtime Optical Flow Estimation on Vein and Artery Ultrasound Sequences Based on Knowledge-Distillation 1 Introduction 2 Related Work 2.1 Dynamic Ultrasound Analysis 2.2 Optical Flow Estimation 2.3 Knowledge Distillation 2.4 Contributions 3 Results and Discussion 4 Conclusion References Metrics/Losses Motion Correction in Low SNR MRI Using an Approximate Rician Log-Likelihood 1 Introduction 2 Background: The Rice Distribution 3 Method 3.1 Priors 3.2 A Stable Approximation of the Rician Log-Likelihood 3.3 Inference 4 Experiments 4.1 Synthetic Data 4.2 Real Sodium MRI 5 Discussion 6 Conclusions References Cross-Sim-NGF: FFT-Based Global Rigid Multimodal Alignment of Image Volumes Using Normalized Gradient Fields 1 Introduction 2 Background 3 Method 3.1 Method for Global 3D Rigid Alignment 4 Performance Analysis 4.1 Similarity Landscape of the Average SNGF 4.2 Multimodal Brain Image Volume Alignment 4.3 Time Analysis 5 Conclusion References Identifying Partial Mouse Brain Microscopy Images from the Allen Reference Atlas Using a Contrastively Learned Semantic Space 1 Introduction 2 Methods 3 Data and Experiments 3.1 Data 3.2 Experiments 4 Discussions and Conclusions 5 Discussions and Conclusions 6 Conclusions References Transformed Grid Distance Loss for Supervised Image Registration 1 Introduction 2 Transformed Grid Distance 3 Experiments 4 Discussions and Conclusion References Efficiency Deformable Lung CT Registration by Decomposing Large Deformation 1 Introduction 2 Methodology 3 Experiments 4 Conclusion References You only Look at Patches: A Patch-wise Framework for 3D Unsupervised Medical Image Registration 1 Introduction 2 Proposed Framework 3 Experimental Results 4 Conclusion References Recent Developments of an Optimal Control Approach to Nonrigid Image Registration 1 Our Approach to Image Registration 1.1 Gradient with Respect to Control F-.4 1.2 Hessian Matrix with Respect to Control Function F-.4 1.3 Partial Gradients with Respect to Control Functions and g-.4 2 Numerical Examples 2.1 A Large Deformation Test and a MRI Registration Test 3 Discussion References 2D/3D Quasi-Intramodal Registration of Quantitative Magnetic Resonance Images 1 Introduction 1.1 Clinical Motivation 1.2 Clinical Data 2 Method 2.1 2D to 3D Reconstruction Using SFS 2.2 3D to 2D Projection Using Binary Search 3 Discussion References Deep Learning-Based Longitudinal Intra-subject Registration of Pediatric Brain MR Images 1 Introduction 2 Methodology 3 Experiments 4 Results 5 Discussion and Conclusion References Real-Time Alignment for Connectomics 1 Introduction 2 Methods 3 Results 4 Conclusion References Weak Bounding Box Supervision for Image Registration Networks 1 Introduction 2 Methods 3 Experiments 4 Results 5 Discussion and Conclusion References Author Index