Simulation and Synthesis in Medical Imaging: 6th International Workshop, SASHIMI 2021, Held in Conjunction with MICCAI 2021, Strasbourg, France, ... Vision, Pattern Recognition, and Graphics)

This book constitutes the refereed proceedings of the 6th International Workshop on Simulation and Synthesis in Medical Imaging, SASHIMI 2021, held in conjunction with MICCAI 2021, in Strasbourg, France, in September 2021.* The 14 full papers presented were carefully reviewed and selected from 18 submissions. The contributions span the following broad categories in alignment with the initial call-for-papers: methods based on generative models or adversarial learning for MRI/CT/ microscopy image synthesis, and several applications of image synthesis and simulation for data augmentation, image enhancement, or segmentation. *The workshop was held virtually. Preface Organization Contents Method-Oriented Papers Detail Matters: High-Frequency Content for Realistic Synthetic MRI Generation 1 Introduction 2 Methods 3 Experiments and Results 4 Conclusion References Joint Image and Label Self-super-Resolution 1 Introduction 1.1 Related Work 2 Methods 2.1 Formulation 2.2 Label Creation 2.3 Training Data Creation 2.4 Implementation 2.5 Comparison Methods 3 Results 3.1 Qualitative Results 3.2 Quantitative Results 4 Discussion and Conclusion References Super-Resolution by Latent Space Exploration: Training with Poorly-Aligned Clinical and Micro CT Image Dataset 1 Introduction 2 Methods 2.1 Overview 2.2 CML-SRNet 2.3 SR by Latent Space Exploration 3 Experiments and Results 3.1 Dataset and Parameter Settings 3.2 Results 4 Discussion and Conclusion References A Glimpse into the Future: Disease Progression Simulation for Breast Cancer in Mammograms 1 Introduction 2 Methods 2.1 Model Architecture 2.2 Semi-supervised Training 3 Experiments 3.1 Metrics and Results 4 Discussion, Limitations, and Future Work References Synth-by-Reg (SbR): Contrastive Learning for Synthesis-Based Registration of Paired Images 1 Introduction 2 Methods 2.1 Overview 2.2 Intra-modality Registration Network 2.3 Image-to-Image Translation Using a Registration Loss 3 Experiments and Results 3.1 Data 3.2 Experimental Setup 3.3 Results 4 Discussion and Conclusion References Learning-Based Template Synthesis for Groupwise Image Registration 1 Introduction 2 Method 2.1 Generative Adversarial Networks 2.2 Auxiliary Segmentor 3 Architecture 4 Experiments 4.1 Datasets and Implementation 4.2 Evaluation Metrics 4.3 Baselines 4.4 Results 5 Conclusion References The Role of MRI Physics in Brain Segmentation CNNs: Achieving Acquisition Invariance and Instructive Uncertainties 1 Introduction 2 Methods 2.1 Network Architecture 2.2 Stratification and Batch Homogeneity 2.3 Casting Simulation as an Augmentation Layer 2.4 Uncertainty Modelling 3 Experiments 3.1 Data 3.2 Simulation Sequence Details 4 Annealing Study: Robustness and Quality Analysis 4.1 Uncertainty Measures and Volumetric Bounds 5 Discussion and Conclusions References Transfer Learning in Optical Microscopy 1 Introduction 2 Data 2.1 Data Augmentation 3 Methods 3.1 Pix2Pix 3.2 Proposed Method 4 Experiments 4.1 Parameter Setup 4.2 Methods of Evaluation 4.3 Results 5 Conclusion References X-ray Synthesis Based on Triangular Mesh Models Using GPU-Accelerated Ray Tracing for Multi-modal Breast Image Registration 1 Background 2 Methods 2.1 Basic Principle 2.2 Implementation on Special Purpose Ray Tracing Units of GPUs 3 Results 3.1 Validation of Attenuation Calculation 3.2 Performance Evaluation 4 Discussion and Conclusion References Application-Oriented Papers Frozen-to-Paraffin: Categorization of Histological Frozen Sections by the Aid of Paraffin Sections and Generative Adversarial Networks 1 Motivation 2 Methods 2.1 Frozen-to-Paraffin Translation 2.2 WSI Feature Augmentation 2.3 Data Set 2.4 Experimental Details 3 Results 4 Discussion 5 Conclusion References SequenceGAN: Generating Fundus Fluorescence Angiography Sequences from Structure Fundus Image 1 Introduction 2 Method 2.1 Datasets 2.2 SequenceGAN for FA Sequences Generation 3 Experimental Results 3.1 Evaluation Metrics 3.2 Results Comparison on FA Sequences Generation 3.3 Ablation Study 4 Discussion and Conclusion References Cerebral Blood Volume Prediction Based on Multi-modality Magnetic Resonance Imaging 1 Introduction 2 Materials and Methods 2.1 Participants 2.2 MR Image Acquisition 2.3 Image Processing 2.4 Method 3 Experiments 3.1 Settings and Metrics 3.2 Result and Discussion 4 Conclusion References Cine-MRI Simulation to Evaluate Tumor Tracking 1 Introduction 2 Related Work 3 Methods 3.1 Data 3.2 Cine-MRI Simulation 3.3 Tracking Algorithms 3.4 Metrics 4 Experiments and Results 4.1 Patient Summary 4.2 Cine-MRI Image Quality 4.3 Tumor Tracking Performance 5 Discussion and Conclusion References GAN-Based Synthetic FDG PET Images from T1 Brain MRI Can Serve to Improve Performance of Deep Unsupervised Anomaly Detection Models 1 Introduction 2 Method 2.1 Synthesis of Realistic PET Data with GANs 2.2 Application to the Training of a Deep Epilepsy Lesion Detection Model 3 Experiments and Results 3.1 Data 3.2 Generation of Synthetic PET Images from T1 Healthy Controls 3.3 Application of Synthetic PET Data to the Training of a Brain Anomaly Detection Model for Epilepsy Patients Screening 4 Discussion and Conclusion References Correction to: Frozen-to-Paraffin: Categorization of Histological Frozen Sections by the Aid of Paraffin Sections and Generative Adversarial Networks Correction to: Chapter “Frozen-to-Paraffin: Categorization of Histological Frozen Sections by the Aid of Paraffin Sections and Generative Adversarial Networks” in: D. Svoboda et al. (Eds.): Simulation and Synthesis in Medical Imaging, LNCS 12965, https://doi.org/10.1007/978-3-030-87592-3_10 Author Index