Medical Image Computing and Computer Assisted Intervention – MICCAI 2021: 24th International Conference, Strasbourg, France, September 27–October 1, 2021, Proceedings, Part I

Preface
Organization
Contents – Part I
Image Segmentation
Noisy Labels are Treasure: Mean-Teacher-Assisted Confident Learning for Hepatic Vessel Segmentation
	1 Introduction
	2 Methods
		2.1 Materials
		2.2 Hepatic CT Preprocessing
		2.3 Mean-Teacher-assisted Confident Learning Framework
		2.4 Loss Function
	3 Experiments and Results
	4 Conclusion
	References
TransFuse: Fusing Transformers and CNNs for Medical Image Segmentation
	1 Introduction
	2 Proposed Method
	3 Experiments and Results
	4 Conclusion
	References
Pancreas CT Segmentation by Predictive Phenotyping
	1 Introduction
	2 Method
		2.1 Problem Formulation
		2.2 Loss Functions
		2.3 Phenotype Embedding
	3 Experiments
		3.1 Dataset
		3.2 Implementation Details
		3.3 Comparison with State-of-the-Arts
		3.4 Results
	4 Discussion and Conclusion
	References
Medical Transformer: Gated Axial-Attention for Medical Image Segmentation
	1 Introduction
	2 Medical Transformer (MedT)
		2.1 Self-attention Overview
		2.2 Gated Axial-Attention
		2.3 Local-Global Training
	3 Experiments and Results
		3.1 Dataset Details
		3.2 Implementation Details
		3.3 Results
	4 Conclusion
	References
Anatomy-Constrained Contrastive Learning for Synthetic Segmentation Without Ground-Truth
	1 Introduction
	2 Methods
	3 Experimental Results
	4 Discussion and Conclusion
	References
Study Group Learning: Improving Retinal Vessel Segmentation Trained with Noisy Labels
	1 Introduction
	2 Methodology
		2.1 Model Structure
		2.2 Study Group Learning (SGL) Scheme
		2.3 Vessel Label Erasing
	3 Experiments
		3.1 Dataset and Implementation
		3.2 Learned Retinal Image Enhancement
		3.3 Study Group Learning
	4 Conclusions
	References
Multi-phase Liver Tumor Segmentation with Spatial Aggregation and Uncertain Region Inpainting
	1 Introduction
	2 Method
		2.1 Spatial Aggregation Module
		2.2 Uncertain Region Inpainting Module
		2.3 Loss Function and Training Strategy
	3 Materials and Experiments
	4 Discussion and Conclusions
	References
Convolution-Free Medical Image Segmentation Using Transformers
	1 Introduction
	2 Materials and Methods
		2.1 Proposed Network
		2.2 Implementation
	3 Results and Discussion
	4 Conclusions
	References
Consistent Segmentation of Longitudinal Brain MR Images with Spatio-Temporal Constrained Networks
	1 Introduction
	2 Method
		2.1 Dataset
		2.2 Spatio-Temporal Constrained Deep Learning Model
		2.3 Initial ROI Segmentation with Supervised Learning
		2.4 Consistent Segmentation with Semi-supervised Learning
		2.5 Implement Details
	3 Experiments and Results
	4 Conclusion
	References
A Multi-branch Hybrid Transformer Network for Corneal Endothelial Cell Segmentation
	1 Introduction
	2 Method
		2.1 Residual Transformer Block
		2.2 The Body, Edge, and Final Branches
		2.3 Loss Function
	3 Experiments
		3.1 Datasets and implementation Details
		3.2 Comparison with SOTA methods
		3.3 Ablation Study
	4 Conclusion
	References
TransBTS: Multimodal Brain Tumor Segmentation Using Transformer
	1 Introduction
	2 Method
		2.1 Overall Architecture of TransBTS
		2.2 Network Encoder
		2.3 Network Decoder
	3 Experiments
		3.1 Main Results
		3.2 Model Complexity
		3.3 Ablation Study
	4 Conclusion
	References
Automatic Polyp Segmentation via Multi-scale Subtraction Network
	1 Introduction
	2 Method
		2.1 Multi-scale Subtraction Module
		2.2 LossNet
	3 Experiments
		3.1 Datasets
		3.2 Evaluation Metrics
		3.3 Implementation Details
		3.4 Comparisons with State-of-the-art
		3.5 Ablation Study
	4 Discussion
	5 Conclusion
	References
Patch-Free 3D Medical Image Segmentation Driven by Super-Resolution Technique and Self-Supervised Guidance
	1 Introduction
	2 Methodology
		2.1 Multi-task Learning
		2.2 Self-Supervised Guidance Module
		2.3 Task-Fusion Module
		2.4 Overall Objective Function
	3 Experiments
		3.1 Datasets
		3.2 Implementation Details
		3.3 Ablation Study
		3.4 Experimental Results
	4 Conclusion
	References
Progressively Normalized Self-Attention Network for Video Polyp Segmentation
	1 Introduction
	2 Method
		2.1 Normalized Self-attention (NS)
		2.2 Progressive Learning Strategy
	3 Experiments
		3.1 Implementation Details
		3.2 Evaluation on Video Polyp Segmentation
		3.3 Ablation Study
	4 Conclusion
	References
SGNet: Structure-Aware Graph-Based Network for Airway Semantic Segmentation
	1 Introduction
	2 Method
		2.1 Multi-task U-Net Module
		2.2 Structure-Aware Graph Convolutional Network
		2.3 Loss Functions and Training Methodology
	3 Experiments and Results
		3.1 Datasets and Implementation Details
		3.2 Results
	4 Conclusion
	References
NucMM Dataset: 3D Neuronal Nuclei Instance Segmentation at Sub-Cubic Millimeter Scale
	1 Introduction
		1.1 Related Works
	2 NucMM Dataset
	3 Method
		3.1 Hybrid-Representation Learning
		3.2 Instance Decoding
		3.3 Implementation
	4 Experiments
		4.1 Methods in Comparison
		4.2 Benchmark Results on the NucMM Dataset
		4.3 Sensitivity of the Decoding Parameters
	5 Conclusion
	References
AxonEM Dataset: 3D Axon Instance Segmentation of Brain Cortical Regions
	1 Introduction
		1.1 Related Works
	2 AxonEM Dataset
		2.1 Dataset Description
		2.2 Dataset Annotation
		2.3 Dataset Analysis
	3 Methods
		3.1 Task and Evaluation Metric
		3.2 State-of-the-Art Methods
	4 Experiments
		4.1 Implementation Details
		4.2 Benchmark Results on SNEMI3D Dataset
		4.3 Benchmark Results on AxonEM Dataset
	5 Conclusion
	References
Improved Brain Lesion Segmentation with Anatomical Priors from Healthy Subjects
	1 Introduction
	2 Method
		2.1 Problem Formulation
		2.2 Intensity-Based Strategy
		2.3 Distribution-Based Strategy
		2.4 Integration with CNNs and Implementation Details
	3 Results
		3.1 Dataset Description
		3.2 Experimental Settings
		3.3 Segmentation Performance
	4 Conclusion
	References
CarveMix: A Simple Data Augmentation Method for Brain Lesion Segmentation
	1 Introduction
	2 Method
		2.1 Problem Formulation
		2.2 CarveMix
		2.3 Relationship with Mixup and CutMix
		2.4 Implementation Details
	3 Experiments
		3.1 Data Description
		3.2 Evaluation Results
	4 Conclusion
	References
Boundary-Aware Transformers for Skin Lesion Segmentation
	1 Introduction
	2 Method
		2.1 Basic Transformer for Segmentation
		2.2 Boundary-Aware Transformer
		2.3 Objective Function
	3 Experimental Results
		3.1 Datasets
		3.2 Implementation Details
		3.3 Comparison with State-of-the-Arts
		3.4 Ablation Study
	4 Conclusion
	References
A Topological-Attention ConvLSTM Network and Its Application to EM Images
	1 Introduction
	2 Related Works
	3 Method
		3.1 Spatial Topological-Attention (STA) Module
		3.2 Iterative Topological-Attention (ITA) Module
	4 Experiments
	5 Conclusion
	References
BiX-NAS: Searching Efficient Bi-directional Architecture for Medical Image Segmentation
	1 Introduction
	2 Methods
		2.1 BiO-Net++: A Multi-scale Upgrade of BiO-Net
		2.2 BiX-NAS: Hierarchical Search for Efficient BiO-Net++
		2.3 Analysis of Searching Fairness and Deficiency
	3 Experiments
		3.1 Datasets and Implementation Details
		3.2 Experimental Results
	4 Conclusion
	References
Multi-task, Multi-domain Deep Segmentation with Shared Representations and Contrastive Regularization for Sparse Pediatric Datasets
	1 Introduction
	2 Method
		2.1 Deep Segmentation Model with Domain-Specific Layers (DSL)
		2.2 Supervised Contrastive Regularization
	3 Experiments
		3.1 Imaging Datasets
		3.2 Implementation Details
		3.3 Evaluation of Predicted Segmentation
	4 Results and Discussion
		4.1 Segmentation Results
		4.2 Supervised Contrastive Regularization Visualization
	5 Conclusion
	References
TEDS-Net: Enforcing Diffeomorphisms in Spatial Transformers to Guarantee Topology Preservation in Segmentations
	1 Introduction
	2 Method
		2.1 Experimental Setup
	3 Results and Discussion
	4 Conclusion
	References
Learning Consistency- and Discrepancy-Context for 2D Organ Segmentation
	1 Introduction
	2 Methodology
		2.1 Consistency Context-Based Organ Segmentation
		2.2 Segmentation Refinement with Discrepancy Context Knowledge
	3 Datasets and Implementation Details
	4 Experimental Results and Analysis
		4.1 The Effectiveness of Our Proposed Method
		4.2 Portability with Different Segmentation Models
		4.3 Qualitative Results
	5 Conclusion
	References
Partially-Supervised Learning for Vessel Segmentation in Ocular Images
	1 Introduction
	2 Method
		2.1 Partially-Supervised Learning
		2.2 Active Learning Framework
		2.3 Latent MixUp
		2.4 Loss Function
	3 Experiment
		3.1 Experiment Setting
		3.2 Performance Comparison
		3.3 Ablation Studies
	4 Conclusion
	References
Unsupervised Network Learning for Cell Segmentation
	1 Introduction
	2 Related Work
	3 Methodology
		3.1 Formulation of the Unsupervised Segmentation Problem
		3.2 USAR for Unsupervised Segmentation Network Learning
		3.3 Avoid Trivial Solutions in Unsupervised Network Learning
		3.4 Objective Function for Learning the Segmentation Network
		3.5 Pseudo Labels to Continually Refine the Segmentation Network
	4 Experimental Results
		4.1 Data and Evaluation Metric
		4.2 Evaluation Results
	5 Conclusion
	References
MT-UDA: Towards Unsupervised Cross-modality Medical Image Segmentation with Limited Source Labels
	1 Introduction
	2 Methodology
		2.1 Dual Cycle Alignment Module
		2.2 Semantic Knowledge Transfer
		2.3 Structural Knowledge Transfer
		2.4 MT-UDA Framework
	3 Experiments and Results
	4 Conclusion
	References
Context-Aware Virtual Adversarial Training for Anatomically-Plausible Segmentation
	1 Introduction
	2 Proposed Method
		2.1 Context-Aware VAT Loss
		2.2 Local Connectivity Constraints
	3 Experimental Setup
	4 Experimental Results
	5 Conclusion
	References
Interactive Segmentation via Deep Learning and B-Spline Explicit Active Surfaces
	1 Introduction
	2 Materials and Methods
		2.1 Proposed Pipeline
		2.2 Data Collection
		2.3 Experimental Details
		2.4 Implementation Details
	3 Results
	4 Discussion
	5 Conclusion
	References
Multi-compound Transformer for Accurate Biomedical Image Segmentation
	1 Introduction
	2 Related Work
	3 Multi-compound Transformer Network
		3.1 Transformer-Self-attention
		3.2 Transformer-Cross-attention
	4 Experiments
		4.1 Datasets and Settings
		4.2 Ablation Studies
		4.3 Comparisons with State-of-the-Art Methods
	5 Conclusions
	References
kCBAC-Net: Deeply Supervised Complete Bipartite Networks with Asymmetric Convolutions for Medical Image Segmentation
	1 Introduction
	2 Method
		2.1 k-Complete-Bipartite Network (kCB-Net)
		2.2 kCBAC-Net: Leveraging Asymmetric Convolutions
		2.3 Deep Supervision (DS)
	3 Experiments and Results
	4 Conclusions
	References
Multi-frame Attention Network for Left Ventricle Segmentation in 3D Echocardiography
	1 Introduction
	2 Methods
		2.1 Multi-Frame Attention Network for Segmentation
	3 Experiments and Results
		3.1 Dataset
		3.2 Implementation Details
		3.3 Experimental Studies
		3.4 Results and Discussion
		3.5 Limitations and Future Works
	4 Conclusion
	References
Coarse-To-Fine Segmentation of Organs at Risk in Nasopharyngeal Carcinoma Radiotherapy
	1 Introduction
	2 Methodology
		2.1 Architecture
		2.2 Objective Function
		2.3 Training Details
	3 Experiment and Analysis
		3.1 Dataset and Evaluation
		3.2 Comparison with State-Of-The-Art Methods
		3.3 Ablation Study
	4 Conclusion
	References
Joint Segmentation and Quantification of Main Coronary Vessels Using Dual-Branch Multi-scale Attention Network
	1 Introduction
	2 Dual-Branch Multi-scale Attention Network for Joint Segmentation and Quantification
		2.1 Problem Formulation
		2.2 Multi-scale Segmentation
		2.3 Attentive Quantification
	3 Experiments and Results
		3.1 Experiment Setup
		3.2 Results and Analysis
	4 Conclusion
	References
A Spatial Guided Self-supervised Clustering Network for Medical Image Segmentation
	1 Introduction
	2 Materials and Methods
		2.1 Materials
		2.2 Overview of the SGSCN
		2.3 Cross-Entropy Loss for Self-supervised Clustering
		2.4 Sparse Spatial Loss
		2.5 Context-Based Consistency Loss
		2.6 Training via Backpropagation
	3 Experimental Setup
		3.1 Evaluation
		3.2 Implementation Details
	4 Results
	5 Discussion
	6 Conclusion
	References
Comprehensive Importance-Based Selective Regularization for Continual Segmentation Across Multiple Sites
	1 Introduction
	2 Methods
		2.1 Shape-Aware Importance (SpAI)
		2.2 Uncertainty-Guided Semantics-Aware Importance (USmAI)
		2.3 Comprehensive Importance-Based Selective Regularization
	3 Experiments
	4 Conclusion
	References
ReSGAN: Intracranial Hemorrhage Segmentation with Residuals of Synthetic Brain CT Scans
	1 Introduction
	2 Method
		2.1 Learning to Synthesize Templates from CT Scans
		2.2 Segmentation Through Residuals
		2.3 Experiments
	3 Results
	4 Clinical Considerations
	5 Conclusion
	References
Refined Local-imbalance-based Weight for Airway Segmentation in CT
	1 Introduction
	2 Method
		2.1 Local-imbalance-based Weight
		2.2 BP-based Weight Enhancement
	3 Experiments and Results
	4 Conclusion
	References
Selective Learning from External Data for CT Image Segmentation
	1 Introduction
	2 Methodology
		2.1 Problem Setup
		2.2 Selective Learning
		2.3 Optimization
	3 Experiments
		3.1 Experimental Setup
		3.2 Experimental Results
		3.3 Analysis of Our Method
	4 Conclusion
	References
Projective Skip-Connections for Segmentation Along a Subset of Dimensions in Retinal OCT
	1 Introduction
		1.1 Clinical Background
		1.2 Related Work
		1.3 Contribution
	2 Method
		2.1 Projective skip-connections
	3 Experiments
		3.1 Data Sets
		3.2 Baseline Methods
		3.3 Training Details
		3.4 Evaluation Details
	4 Results and Discussion
	5 Conclusion
	References
MouseGAN: GAN-Based Multiple MRI Modalities Synthesis and Segmentation for Mouse Brain Structures
	1 Introduction
	2 Methods
		2.1 Modality Translation Model
		2.2 Segmentation Model
	3 Results and Discussion
		3.1 Dataset Description
		3.2 Experiment Details
		3.3 Evaluation on Image Synthesis
		3.4 Evaluation on Structural Segmentation
		3.5 Ablation Studies
	4 Conclusion
	References
Style Curriculum Learning for Robust Medical Image Segmentation
	1 Introduction
	2 Methodology
		2.1 Curriculum Learning for Robustness
		2.2 Style Transfer Based Sample Generation
		2.3 Gradient Manipulation Based Learning Strategy
		2.4 Local Gradient Smoothing for Stability
	3 Experimental Results
		3.1 Result Details
	4 Conclusion
	References
Towards Efficient Human-Machine Collaboration: Real-Time Correction Effort Prediction for Ultrasound Data Acquisition
	1 Introduction
	2 Method
		2.1 Framework
		2.2 Prediction Target
		2.3 Prediction Network
		2.4 Cost Model
	3 Experiments and Results
		3.1 Per-Frame Prediction
		3.2 Simulated Experiments
	4 Conclusion
	References
Residual Feedback Network for Breast Lesion Segmentation in Ultrasound Image
	1 Introduction
	2 Method
		2.1 Overview
		2.2 Encoder-Decoder Baseline
		2.3 Residual Representation Module
		2.4 Residual Feedback Transmission and Loss Function
	3 Experiments
		3.1 Segmentation Performance
		3.2 Ablation Study
	4 Discussion and Conclusion
	References
Learning to Address Intra-segment Misclassification in Retinal Imaging
	1 Introduction
	2 Methods
		2.1 Adversarial Segmentation Network
		2.2 Binary-to-multi-class Fusion Network
	3 Experiments
		3.1 Experiment Setting
		3.2 Experiment Results
	4 Discussion
	References
Flip Learning: Erase to Segment
	1 Introduction
	2 Method
	3 Experimental Result
	4 Conclusion
	References
DC-Net: Dual Context Network for 2D Medical Image Segmentation
	1 Introduction
	2 Related Works
	3 Our Proposed Method
		3.1 Global Context Transformer Encoder
		3.2 Decoder with Adaptive Context Fusion Module
	4 Experiments and Discussion
		4.1 Experimental Settings
		4.2 Evaluation on ISIC 2018 and ISBI 2012
		4.3 Ablation Study
	5 Conclusion
	References
LIFE: A Generalizable Autodidactic Pipeline for 3D OCT-A Vessel Segmentation
	1 Introduction
	2 Methods
		2.1 Local Intensity Fusion: LIF
		2.2 Cross-Modality Feature Extraction: LIFE
		2.3 Experimental Details
	3 Results
	4 Discussion and Conclusion
	References
Superpixel-Guided Iterative Learning from Noisy Labels for Medical Image Segmentation
	1 Introduction
	2 Method
		2.1 Superpixel Representation
		2.2 Iterative Model Learning
	3 Experiment
		3.1 Dataset
		3.2 Experiment Setup
		3.3 Experiments on ISIC Dataset
		3.4 Ablation Study
		3.5 Experiments on JSRT Dataset
	4 Conclusion
	References
A Hybrid Attention Ensemble Framework for Zonal Prostate Segmentation
	1 Introduction and Related Work
	2 Methods
		2.1 Hybrid Attention Ensemble Framework
		2.2 Attention Bridge Network
		2.3 Targeted Segmentation Network
		2.4 Post-processing
	3 Experiments and Results
		3.1 Data and Implementation
		3.2 Results
	4 Discussion and Conclusions
	References
3D-UCaps: 3D Capsules Unet for Volumetric Image Segmentation
	1 Introduction
	2 Background
	3 Our Proposed 3D-UCaps Network
	4 Experimental Results
	5 Conclusion
	References
HRENet: A Hard Region Enhancement Network for Polyp Segmentation
	1 Introduction
	2 Methodology
		2.1 Informative Context Enhancement (ICE)
		2.2 Adaptive Feature Aggregation (AFA)
		2.3 Edge and Structure Consistency Aware Loss (ESCLoss)
	3 Experiments
		3.1 Experimental Settings
		3.2 Comparison with State-of-the-arts
		3.3 Ablation Study
	4 Conclusion
	References
A Novel Hybrid Convolutional Neural Network for Accurate Organ Segmentation in 3D Head and Neck CT Images
	1 Introduction
	2 Method
		2.1 Dataset
		2.2 Network Architecture
		2.3 Loss Function
		2.4 Implementation Details
	3 Experimental Results
		3.1 Results on the Collected Public Dataset
		3.2 Results on MICCAI 2015 Challenge Dataset
	4 Conclusion
	References
TumorCP: A Simple but Effective Object-Level Data Augmentation for Tumor Segmentation
	1 Introduction
	2 Method
		2.1 TumorCP's augmentation
		2.2 Intuitions on TumorCP's Effectiveness
	3 Experiments and Discussion
		3.1 Experiment Settings
		3.2 Ablation Study
		3.3 Towards Extremely Low-Data Regime
	4 Conclusion and Future Works
	References
Modality-Aware Mutual Learning for Multi-modal Medical Image Segmentation
	1 Introduction
	2 Method
		2.1 Modality-Specific Model
		2.2 Modality-Aware Module
		2.3 Mutual Learning Strategy
	3 Experiments and Results
	4 Conclusion
	References
Hybrid Graph Convolutional Neural Networks for Landmark-Based Anatomical Segmentation
	1 Introduction
	2 Hybrid Graph Convolutional Neural Networks
	3 Experiments and Discussion
	4 Conclusions
	References
RibSeg Dataset and Strong Point Cloud Baselines for Rib Segmentation from CT Scans
	1 Introduction
	2 Materials and Methods
		2.1 RibSeg Dataset
		2.2 Rib Segmentation from a Viewpoint of Point Clouds
	3 Results
		3.1 Quantitative Analysis
		3.2 Qualitative Analysis
	4 Conclusion and Further Work
	References
Hierarchical Self-supervised Learning for Medical Image Segmentation Based on Multi-domain Data Aggregation
	1 Introduction
	2 Methodology
		2.1 Multi-Domain Data Aggregation
		2.2 Hierarchical Self-supervised Learning (HSSL)
	3 Experiments and Results
	4 Conclusions
	References
CCBANet: Cascading Context and Balancing Attention for Polyp Segmentation
	1 Introduction
	2 Method
	3 Experiments
		3.1 Experiment Results
		3.2 Ablation Study
	4 Conclusion
	References
Point-Unet: A Context-Aware Point-Based Neural Network for Volumetric Segmentation
	1 Introduction
	2 Related Work
	3 Proposed Point-Unet
		3.1 Saliency Attention
		3.2 Context-Aware Sampling
		3.3 Point-Based Segmentation
	4 Experimental Results
	5 Conclusion
	References
TUN-Det: A Novel Network for Thyroid Ultrasound Nodule Detection
	1 Introduction
	2 Proposed Method
		2.1 TUN-Det Architecture
		2.2 Multi-head Classification and Regression Module
		2.3 Supervision
	3 Experimental Results
		3.1 Datasets and Evaluation Metrics
		3.2 Implementation Details
		3.3 Ablation Study
		3.4 Comparisons Against State-of-the-Arts
	4 Conclusion and Discussion
	References
Distilling Effective Supervision for Robust Medical Image Segmentation with Noisy Labels
	1 Introduction
	2 Method
		2.1 Pixel-Wise Robust Learning
		2.2 Image-Level Robust Learning
	3 Experiments and Results
		3.1 Datasets and Implementation Details
		3.2 Results
	4 Conclusion
	References
On the Relationship Between Calibrated Predictors and Unbiased Volume Estimation
	1 Introduction
	2 The Relationship Between Calibration and Volume Bias
	3 Empirical Setup
	4 Results and Discussion
	5 Conclusions
	References
High-Resolution Segmentation of Lumbar Vertebrae from Conventional Thick Slice MRI
	1 Introduction
	2 Methods
		2.1 Segmentation
		2.2 Reconstruction
		2.3 Implementation Details
	3 Results
		3.1 Thick Slice MRI Segmentation
		3.2 Reconstruction
		3.3 Registration
	4 Conclusion
	References
Shallow Attention Network for Polyp Segmentation
	1 Introduction
	2 Related Work
	3 Method
		3.1 Color Exchange
		3.2 Shallow Attention Module
		3.3 Probability Correction Strategy
		3.4 Loss Function
	4 Experiments
		4.1 Datasets and Training Settings
		4.2 Quantitative Comparison
		4.3 Visual Comparison
		4.4 Ablation Study
	5 Conclusion
	References
A Line to Align: Deep Dynamic Time Warping for Retinal OCT Segmentation
	1 Introduction
	2 State of the Art
		2.1 Automated Retinal Layer Segmentation
		2.2 Dynamic Time Warping and Relaxed Formulation
	3 Proposed Approach
		3.1 Segmentation as Sequence Alignment
		3.2 First Approach: Pretrained CNN Features
		3.3 Second Approach: End-to-End Learning of CNN Features
		3.4 The Cumulative Neighborhood Matrix
	4 Experiments
	5 Discussion and Conclusion
	References
Learnable Oriented-Derivative Network for Polyp Segmentation
	1 Introduction
	2 Methods
		2.1 Overview
		2.2 Learnable Oriented-Derivative Representation
		2.3 Border Region Searching
	3 Experiments
		3.1 Datasets and Evaluation Metrics
		3.2 Implementation Details
		3.3 Experiment Results
		3.4 Ablation Study
	4 Conclusion
	References
LambdaUNet: 2.5D Stroke Lesion Segmentation of Diffusion-Weighted MR Images
	1 Introduction
	2 Methods
		2.1 Lambda+ Layers
	3 Experiments
		3.1 Results
		3.2 Discussion
	4 Conclusion
	References
Author Index