Simplifying Medical Ultrasound: Second International Workshop, ASMUS 2021, Held in Conjunction with MICCAI 2021, Strasbourg, France, September 27, ... Vision, Pattern Recognition, and Graphics)

This book constitutes the proceedings of the Second International Workshop on Advances in Simplifying Medical UltraSound, ASMUS 2021, held on September 27, 2021, in conjunction with MICCAI 2021, the 24th International Conference on Medical Image Computing and Computer-Assisted Intervention. The conference was planned to take place in Strasbourg, France, but changed to an online event due to the Coronavirus pandemic. The 22 papers presented in this book were carefully reviewed and selected from 30 submissions. They were organized in topical sections as follows: segmentation and detection; registration, guidance and robotics; classification and image synthesis; and quality assessment and quantitative imaging. Preface Organization Contents Segmentation and Detection Automatic Ultrasound Vessel Segmentation with Deep Spatiotemporal Context Learning 1 Introduction 2 Methodology 3 Results 4 Conclusion References Multimodal Continual Learning with Sonographer Eye-Tracking in Fetal Ultrasound 1 Introduction 2 Methods 2.1 Training 3 Results and Discussion 4 Conclusion References Development and Evaluation of Intraoperative Ultrasound Segmentation with Negative Image Frames and Multiple Observer Labels 1 Introduction 2 Methods 2.1 Segmentation Network 2.2 Frame Classification Network 2.3 Label Sampling 2.4 Pre-screening Strategy 2.5 Loss Functions for Segmentation 2.6 Evaluation Experiments 3 Results and Discussion 4 Conclusion References Automatic Tomographic Ultrasound Imaging Sequence Extraction of the Anal Sphincter 1 Introduction 2 Materials and Methods 2.1 Acquisition Protocol 2.2 The Proposed Pipeline 2.3 Data Collection 2.4 Evaluation Methodology 3 Experiments, Results and Discussion 3.1 Implementation Details 3.2 Results 3.3 Discussion 4 Conclusion References Lung Ultrasound Segmentation and Adaptation Between COVID-19 and Community-Acquired Pneumonia 1 Introduction 2 Methods 2.1 Supervised Segmentation with U-Net 2.2 Unsupervised Segmentation via Image and Feature Alignment 2.3 Implementation Details 2.4 Data 2.5 Experiments 3 Results 4 Discussion 5 Conclusion References An Efficient Tracker for Thyroid Nodule Detection and Tracking During Ultrasound Screening 1 Introduction 2 Methodology 2.1 Overall Architecture 2.2 Fusion Module 2.3 Advanced Post-processing Module 2.4 Minibatch Self-supervised Learning Module 3 Experiments 4 Conclusion References TransBridge: A Lightweight Transformer for Left Ventricle Segmentation in Echocardiography 1 Introduction 2 Methods 3 Experiments 4 Results 5 Discussion 6 Conclusion References Registration, Guidance and Robotics Adversarial Affine Registration for Real-Time Intraoperative Registration of 3-D US-US for Brain Shift Correction 1 Introduction 2 Methods 2.1 Affine Registration 2.2 Dataset and Experimental Setup 3 Results 4 Discussion and Conclusion References Robust Ultrasound-to-Ultrasound Registration for Intra-operative Brain Shift Correction with a Siamese Neural Network 1 Introduction 2 Methods 2.1 Siamese Network 2.2 Training 2.3 Preprocessing 2.4 The 2.5D Approach 2.5 Affine Transformation 2.6 Experimental Setup 3 Results 4 Discussion 5 Conclusion References Pose Estimation of 2D Ultrasound Probe from Ultrasound Image Sequences Using CNN and RNN 1 Introduction 2 Methods 2.1 Feature Extraction 2.2 Relative Pose Estimation 2.3 Feature Aggregation 2.4 Absolute Pose Estimation 2.5 Loss Function 3 Materials 4 Experiments 4.1 Experimental Condition 4.2 Experimental Results 5 Conclusion References Evaluation of Low-Cost Hardware Alternatives for 3D Freehand Ultrasound Reconstruction in Image-Guided Neurosurgery 1 Introduction 2 Previous Work 3 Methodology 3.1 Hardware Setups 3.2 Phantoms and Marker Construction 3.3 Experiments 4 Results 4.1 Reconstruction Quality Results 4.2 Qualitative Results 5 Discussion 6 Conclusion and Future Work References Application Potential of Robot-Guided Ultrasound During CT-Guided Interventions 1 Introduction 2 Methods 2.1 Survey Design 2.2 Recruitment 2.3 Data Analysis 3 Results 3.1 Personal Professional Information 3.2 Current Clinical Practice 3.3 Potential Use Cases 3.4 HRI in the IR 4 Discussion 5 Conclusion References Classification and Image Synthesis Towards Scale and Position Invariant Task Classification Using Normalised Visual Scanpaths in Clinical Fetal Ultrasound 1 Introduction 2 Method 2.1 Time-Series Classification of Scanpaths 3 Data 4 Results 5 Discussion 6 Conclusion References Efficient Echocardiogram View Classification with Sampling-Free Uncertainty Estimation 1 Introduction 2 Related Works and Limitations 3 Proposed Method: Efficient-EvidNet 3.1 Background: Evidential Deep Learning 3.2 Methodology 4 Experiment 4.1 Dataset and Implementation Details 4.2 Results and Discussion 5 Conclusion References Contrastive Learning for View Classification of Echocardiograms 1 Introduction 2 Related Work 3 Methodology 3.1 Data 4 Experiments and Discussion 4.1 Experimental Setup 4.2 Classification Performance 4.3 Ablation Studies and Failure Cases 5 Conclusion References Imaging Biomarker Knowledge Transfer for Attention-Based Diagnosis of COVID-19 in Lung Ultrasound Videos 1 Introduction 2 Materials and Methods 3 Experiments 4 Discussion and Conclusion References Endoscopic Ultrasound Image Synthesis Using a Cycle-Consistent Adversarial Network 1 Introduction 2 Methods 2.1 Data 2.2 EUS/CT Image-To-Image Translation 2.3 Implementation Details 2.4 Evaluation 3 Results and Discussion 4 Conclusions References Realistic Ultrasound Image Synthesis for Improved Classification of Liver Disease 1 Introduction 2 Materials and Methods 2.1 Stage-I GAN 2.2 Stage-II GAN 3 Experiments and Results 3.1 Dataset 3.2 GAN-Based Network Evaluation 3.3 Classification Evaluation 4 Conclusion References Quality Assessment and Quantitative Imaging Adaptable Image Quality Assessment Using Meta-Reinforcement Learning of Task Amenability 1 Introduction 2 Methods 2.1 Image Quality Assessment by Task Amenable Data Selection 2.2 Meta-Reinforcement Learning with Different Labels 3 Experiments 4 Results 5 Discussion and Conclusion References Deep Video Networks for Automatic Assessment of Aortic Stenosis in Echocardiography 1 Introduction 1.1 Clinical Background: Aortic Stenosis 1.2 Our Contributions 1.3 Related Work 2 Materials and Method 2.1 AS Dataset 2.2 Modelling Prediction Uncertainty with Gaussian Priors 2.3 Implementation 3 Experiments and Results 3.1 Out-of-Distribution Detection 4 Conclusion References Pruning MobileNetV2 for Efficient Implementation of Minimum Variance Beamforming 1 Introduction 2 Beamforming Using MobileNetV2 3 The Proposed Pruning Method 4 Results and Discussions 5 Conclusions References Automatic Fetal Gestational Age Estimation from First Trimester Scans 1 Introduction 2 Materials and Methods 2.1 Dataset 2.2 Fetal Head and Body Segmentation 2.3 Data Augmentation 2.4 Evaluation Metrics 3 Results 4 Discussion and Conclusion References Author Index