Medical Image Computing and Computer Assisted Intervention – MICCAI 2023: 26th International Conference, Vancouver, BC, Canada, October 8–12, 2023, ... IX (Lecture Notes in Computer Science, 14228)

The ten-volume set LNCS 14220, 14221, 14222, 14223, 14224, 14225, 14226, 14227, 14228, and 14229 constitutes the refereed proceedings of the 26th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2023, which was held in Vancouver, Canada, in October 2023. The 730 revised full papers presented were carefully reviewed and selected from a total of 2250 submissions. The papers are organized in the following topical sections: Part I: Machine learning with limited supervision and machine learning – transfer learning; Part II: Machine learning – learning strategies; machine learning – explainability, bias, and uncertainty; Part III: Machine learning – explainability, bias and uncertainty; image segmentation; Part IV: Image segmentation; Part V: Computer-aided diagnosis; Part VI: Computer-aided diagnosis; computational pathology; Part VII: Clinical applications – abdomen; clinicalapplications – breast; clinical applications – cardiac; clinical applications – dermatology; clinical applications – fetal imaging; clinical applications – lung; clinical applications – musculoskeletal; clinical applications – oncology; clinical applications – ophthalmology; clinical applications – vascular; Part VIII: Clinical applications – neuroimaging; microscopy; Part IX: Image-guided intervention, surgical planning, and data science; Part X: Image reconstruction and image registration.'Semantic Segmentation of Surgical Hyperspectral Images Under Geometric Domain Shifts'is available open access under a Creative Commons Attribution 4.0 International License via Springerlink. Preface Organization Contents – Part IX Image-Guided Intervention, Surgical Planning, and Data Science UXDiff: Synthesis of X-Ray Image from Ultrasound Coronal Image of Spine with Diffusion Probabilistic Network 1 Introduction 2 Method 2.1 Conditional Diffusion Models for U2X Translation 2.2 Attention Module with Angle Embedding 2.3 Consistent Conditional Loss 3 Experiments 3.1 Dataset 3.2 Implementation Details 3.3 Synthesis Performance 3.4 Comparison with Cobb Angle 4 Conclusion References EndoSurf: Neural Surface Reconstruction of Deformable Tissues with Stereo Endoscope Videos 1 Introduction 2 Method 2.1 Overview 2.2 EndoSurf: Representing Scenes as Deformable Neural Fields 2.3 Optimization 3 Experiments 3.1 Experiment Settings 3.2 Qualitative and Quantitative Results 4 Conclusion References Surgical Video Captioning with Mutual-Modal Concept Alignment 1 Introduction 2 Surgical Concept Alignment Network 2.1 Overview of SCA-Net 2.2 Surgical Concept Learning 2.3 Mutual-Modality Concept Alignment 2.4 Optimization 3 Experiment 3.1 Dataset and Implementation Details 3.2 Comparison on Neurosurgery Video Captioning 3.3 Comparison on EndoVis Image Captioning 4 Conclusion References SEDSkill: Surgical Events Driven Method for Skill Assessment from Thoracoscopic Surgical Videos 1 Introduction 2 Method 2.1 Basic Regression Module 2.2 Local-Global Difference Module 3 Experiments 3.1 Comparison with the State-of-the-Art Methods 3.2 Ablation Study 4 Conclusion References Neural LerPlane Representations for Fast 4D Reconstruction of Deformable Tissues 1 Introduction 2 Method 2.1 Overview 2.2 Preliminaries 2.3 Neural LerPlane Representations for Deformable Tissues 2.4 Spatiotemporal Importance Sampling 2.5 Coordinate-Time Encoding 2.6 Optimization 3 Experiments 3.1 Dataset and Evaluation Metrics 3.2 Implementation Details 3.3 Evaluation 3.4 Ablation Study 4 Conclusion and Future Work References SegmentOR: Obtaining Efficient Operating Room Semantics Through Temporal Propagation 1 Introduction 2 Related Work 3 Method 4 Experiments 5 Results and Discussion References Revisiting Distillation for Continual Learning on Visual Question Localized-Answering in Robotic Surgery 1 Introduction 2 Methodology 2.1 Preliminaries 2.2 Continual Surgical VQLA (CS-VQLA) 3 Experiments 3.1 Dataset and Setup 3.2 Results 4 Conclusion References Intra-operative Forecasting of Standing Spine Shape with Articulated Neural Kernel Fields 1 Introduction 2 Methods 2.1 Intra-operative 3D Model from Multi-view X-Rays 2.2 Learnable Shape Deformation Kernel 2.3 Articulation SDF 3 Experiments 4 Conclusion References Rectifying Noisy Labels with Sequential Prior: Multi-scale Temporal Feature Affinity Learning for Robust Video Segmentation 1 Introduction 2 Method 2.1 Temporal Feature Affinity Learning 2.2 Multi-scale Supervision 3 Experiments 3.1 Dataset Description and Experiment Settings 3.2 Experiment Results on EndoVis 2018 Dataset 3.3 Experiment Results on Rat Colon Dataset 4 Discussion and Conclusion References Foundation Model for Endoscopy Video Analysis via Large-Scale Self-supervised Pre-train 1 Introduction 2 Method 2.1 Video Transformer for Spatial-Temporal Encoding 2.2 Self-supervised Pre-train via Spatial-Temporal Matching 2.3 Overall Optimization Objective and Pre-training Specifics 3 Experiment 3.1 Datasets and Downstream Setup 3.2 Comparison with State-of-the-Art Methods 3.3 Analytical Studies 4 Conclusion and Discussion References Point Cloud Diffusion Models for Automatic Implant Generation 1 Introduction 2 Methods 3 Experiments 4 Results and Discussion 5 Conclusion References A Novel Video-CTU Registration Method with Structural Point Similarity for FURS Navigation 1 Introduction 2 Video-CTU Registration 2.1 Ureteroscopic Image Characteristics 2.2 Ureteroscopic Structure Extraction 2.3 Virtual CTU Depth Map Generation and Thresholding 2.4 Structural Point Similarity and Optimization 3 Results and Discussion 4 Conclusion References Pelphix: Surgical Phase Recognition from X-Ray Images in Percutaneous Pelvic Fixation 1 Introduction 2 Related Work 3 Method 3.1 Image Sequence Simulation for Percutaneous Fixation 3.2 Transformer Architecture for X-ray-based SPR 4 Evaluation 5 Discussion and Conclusion References TCL: Triplet Consistent Learning for Odometry Estimation of Monocular Endoscope 1 Introduction 2 Methodology 2.1 Unsupervised SfM with Triplet Consistency Learning 2.2 Learning with Geometric Consistency 2.3 Learning with Appearance Inconsistency 2.4 Overall Loss 3 Experiments 4 Conclusion References Regressing Simulation to Real: Unsupervised Domain Adaptation for Automated Quality Assessment in Transoesophageal Echocardiography 1 Introduction 2 Methodology 2.1 Dataset Collection 2.2 Simulation-to-Real AQA Network (SR-AQA) 3 Experiments 3.1 Experimental Settings 3.2 Comparison with State-of-the-Arts 3.3 Ablation Study 4 Conclusion References Estimated Time to Surgical Procedure Completion: An Exploration of Video Analysis Methods 1 Introduction 2 Related Work 3 Methods 3.1 Evaluation Metrics 3.2 Datasets 3.3 Loss Functions 3.4 ETC Models 4 Results 4.1 Ablation Experiments 4.2 Appendectomy and RARP Results 5 Discussion References A Transfer Learning Approach to Localising a Deep Brain Stimulation Target 1 Introduction 2 Materials and Methods 3 Experiments and Discussions References Soft-Tissue Driven Craniomaxillofacial Surgical Planning 1 Introduction 2 Method 2.1 Bony Planner (BP) 2.2 Facial Simulator (FS) 2.3 Self-verified Virtual Planning 3 Experiments and Results 3.1 Dataset 3.2 Implementation and Evaluation Methods 3.3 Results 4 Discussions and Conclusions References ACT-Net: Anchor-Context Action Detection in Surgery Videos 1 Introduction 2 Methodology 2.1 Our ACD Module 2.2 CCD Module for Reliable Action Detection 3 Experimental Results 4 Conclusions References From Tissue to Sound: Model-Based Sonification of Medical Imaging 1 Introduction 2 State-of-the-Art of Sonification in the Medical Domain 3 Towards Model-Based Sonic Interaction with Multimodal Medical Imaging Data 4 Methods 4.1 Physical Model 4.2 Experiment and Results 5 Discussion and Conclusion References Deep Homography Prediction for Endoscopic Camera Motion Imitation Learning 1 Introduction 2 Materials and Methods 2.1 Theoretical Background 2.2 Data and Data Preparation 2.3 Proposed Pipeline 3 Experiments and Evaluation Methodology 3.1 Camera Motion Estimator 3.2 Camera Motion Predictor 4 Results 4.1 Camera Motion Estimator 4.2 Camera Motion Prediction 5 Conclusion and Outlook References Learning Expected Appearances for Intraoperative Registration During Neurosurgery 1 Introduction 2 Method 2.1 Problem Formulation 2.2 Expected Appearances Synthesis 2.3 Pose Regression Network 3 Results 4 Discussion and Conclusion References SENDD: Sparse Efficient Neural Depth and Deformation for Tissue Tracking 1 Introduction 2 Background and Related Work 3 Methods 4 Experiments 5 Conclusion References Cochlear Implant Fold Detection in Intra-operative CT Using Weakly Supervised Multi-task Deep Learning 1 Introduction 2 Methodology 2.1 Dataset 2.2 Synthetic CT Generation 2.3 Multi-task Deep Learning Network 3 Results 4 Discussion and Conclusion References Detecting the Sensing Area of a Laparoscopic Probe in Minimally Invasive Cancer Surgery 1 Introduction 2 Related Work 3 Dataset 4 Probe Axis-Surface Intersection Detection 4.1 Overview 4.2 Intersection Detection as Segmentation 4.3 Intersection Detection as Regression 4.4 Implementation 5 Results 6 Conclusion References High-Quality Virtual Single-Viewpoint Surgical Video: Geometric Autocalibration of Multiple Cameras in Surgical Lights 1 Introduction 2 Method 2.1 Camera Movement Detection 2.2 Detecting the Timing for Obtaining a Homography Matrix 3 Experiments and Results 3.1 Qualitative Evaluation 3.2 Quantitative Evaluation 4 Conclusion and Discussion References SurgicalGPT: End-to-End Language-Vision GPT for Visual Question Answering in Surgery 1 Introduction 2 Proposed Method 2.1 Preliminaries 2.2 LV-GPT: Language-Vision GPT 3 Experiment 3.1 Dataset 3.2 Implementation Details 4 Results 5 Conclusion References Intraoperative CT Augmentation for Needle-Based Liver Interventions 1 Introduction 2 Method 2.1 Vessel Map Extraction 2.2 Data Augmentation 2.3 Neural Network 2.4 Augmented CT 3 Results and Discussion 3.1 Dataset and Implementation Details 3.2 Results 3.3 Ablation Study and Additional Results 4 Conclusion References LABRAD-OR: Lightweight Memory Scene Graphs for Accurate Bimodal Reasoning in Dynamic Operating Rooms 1 Introduction 2 Methodology 2.1 Single Timepoint Scene Graph Generation 2.2 Scene Graphs as Memory Representations 2.3 Architecture Design 3 Experiments 4 Results and Discussion 5 Conclusion References Pelvic Fracture Segmentation Using a Multi-scale Distance-Weighted Neural Network 1 Introduction 2 Methods 2.1 The Overall Segmentation Framework 2.2 Fracture Segmentation Network 2.3 Post-processing 2.4 Data and Annotation 3 Experiments and Results 3.1 Implementation 3.2 Evaluation 3.3 Results 4 Discussion and Conclusion References Pelvic Fracture Reduction Planning Based on Morphable Models and Structural Constraints 1 Introduction 2 Method 2.1 Gaussian Process Morphable Models for Pelvic Bone 2.2 Symmetrical Complementary Registration 2.3 Articular Surface Detection and Matching 3 Experiments and Results 3.1 Experiments on Simulated Data 3.2 Experiments on Clinical Data 4 Discussion and Conclusion References High-Resolution Cranial Defect Reconstruction by Iterative, Low-Resolution, Point Cloud Completion Transformers 1 Introduction 2 Methods 2.1 Overview 2.2 Preprocessing 2.3 Network Architecture - Point Cloud Completion Transformer 2.4 Objective Function 2.5 Iterative Completion 2.6 Postprocessing 2.7 Dataset and Experimental Setup 3 Results 4 Discussion References Regularized Kelvinlet Functions to Model Linear Elasticity for Image-to-Physical Registration of the Breast 1 Introduction 2 Methods 2.1 Regularized Kelvinlet Functions 2.2 Registration Task 3 Experiments and Results 3.1 Hyperparameters Sensitivity Analysis 3.2 Registration Methods Comparison 4 Limitations and Conclusion References Synthesising Rare Cataract Surgery Samples with Guided Diffusion Models 1 Introduction 2 Method 2.1 Denoising Diffusion Probabilistic Models 2.2 Classifier-Free Guidance 2.3 Tool Usage Analysis and Sample Generation 3 Experiments and Results 3.1 Experimental Setup and Dataset 3.2 Quantitative Image Quality 3.3 Qualitative Results and User Study 3.4 Downstream Tool Classification 4 Conclusion References A Closed-Form Solution to Electromagnetic Sensor Based Intraoperative Limb Length Measurement in Total Hip Arthroplasty 1 Introduction 2 Methodology 2.1 Problem Formulation 2.2 Limb Length Measurement Without Femoral Repositioning 3 Experiments 3.1 Simulation and Robustness Assessment 3.2 Phantom Experiments 3.3 Cadaver Experiments 4 Conclusion References A Unified Deep-Learning-Based Framework for Cochlear Implant Electrode Array Localization 1 Introduction 2 Method 2.1 Data 2.2 Multi-task U-Net 2.3 Postprocessing Algorithms for EA Localization 3 Experiments and Results 3.1 Implementation Details 3.2 Evaluation and Results 4 Conclusions References GLSFormer: Gated - Long, Short Sequence Transformer for Step Recognition in Surgical Videos 1 Introduction 2 The GLSFormer Model 3 Experiments and Results 4 Conclusion References CAT-ViL: Co-attention Gated Vision-Language Embedding for Visual Question Localized-Answering in Robotic Surgery 1 Introduction 2 Methodology 2.1 Preliminaries 2.2 CAT-ViL DeiT 3 Experiments 3.1 Dataset 3.2 Implementation Details 3.3 Results 4 Conclusions References Semantic Virtual Shadows (SVS) for Improved Perception in 4D OCT Guided Surgery 1 Introduction 2 Related Work 3 Methodology 3.1 Shadow Augmentation on Surface Structures 3.2 Shadow Augmentation on Subsurface Structures 4 Experiments and Results 4.1 Implementation and Comparative Time Profiling 4.2 User Study 5 Discussion and Conclusion References Intelligent Virtual B-Scan Mirror (IVBM) 1 Introduction 2 Related Work 3 Methodology 3.1 Definition of an Intelligent Virtual B-Scan Mirror 3.2 Method Components 3.3 Composition of an Intelligent Virtual B-Scan Mirror 4 Experimental Setup and User Study 5 Discussion and Conclusion References Multi-view Guidance for Self-supervised Monocular Depth Estimation on Laparoscopic Images via Spatio-Temporal Correspondence 1 Introduction 2 Method 2.1 Self-supervised Monocular Depth Estimation 2.2 Improving Depth with Multi-view Guidance 2.3 Deformable Patch Matching and Cycled Prediction Learning 3 Experiments 3.1 Datasets and Evaluation Metrics 3.2 Implementation Details 3.3 Comparison Experiments 3.4 Ablation Study 4 Discussion and Conclusions References POV-Surgery: A Dataset for Egocentric Hand and Tool Pose Estimation During Surgical Activities 1 Introduction 2 Method 2.1 Multi-view Body Motion Capture 2.2 Hand-Object Manipulation Sequence Generation 2.3 Body and Hand Pose Fusion and Camera Pose Calculation 2.4 Rendering and POV-Surgery Dataset Statistics 3 Experiment 4 Conclusion References Surgical Activity Triplet Recognition via Triplet Disentanglement 1 Introduction 2 Methods 2.1 Sampling Video Clips 2.2 Generating Soft Labels 2.3 Disentanglement Framework 2.4 Hierarchical Training Schedule 2.5 Separation Processing 2.6 Loss Function 3 Experiments 3.1 Results and Discussion 4 Conclusion References Automatic Surgical Reconstruction for Orbital Blow-Out Fracture via Symmetric Prior Anatomical Knowledge-Guided Adversarial Generative Network 1 Introduction 2 Methods 2.1 Automatic Generation of SPAK Based on Spatial Transformation 2.2 SPAK-Guided GAN for Orbital Wall Reconstruction 2.3 Multiple Loss Function Supervision for GAN 3 Experiments and Results 3.1 Data Set and Settings 3.2 Ablation Experiment Results 3.3 Comparative Experiment Results 4 Conclusion References A Multi-task Network for Anatomy Identification in Endoscopic Pituitary Surgery 1 Introduction 2 Related Work 3 Methods 4 Experimental Setup 5 Dataset Description 6 Results and Discussion 7 Conclusion References Robust Vertebra Identification Using Simultaneous Node and Edge Predicting Graph Neural Networks 1 Introduction 2 Related Work 3 Method 4 Experiments 4.1 Dataset 4.2 Training 4.3 Evaluation 4.4 Results 5 Conclusion References Imitation Learning from Expert Video Data for Dissection Trajectory Prediction in Endoscopic Surgical Procedure 1 Introduction 2 Method 2.1 Implicit Modeling for Surgical Dissection Decision-Making 2.2 Training Implicit Policy as Diffusion Models 2.3 Conditional Sampling with Forward-Diffusion Guidance 3 Experiments 3.1 Experimental Dataset and Evaluation Metrics 3.2 Comparison with State-of-the-Art Methods 3.3 Ablation Study 4 Conclusion References Surgical Action Triplet Detection by Mixed Supervised Learning of Instrument-Tissue Interactions 1 Introduction 2 Methodology 3 Experimental Results and Discussion 3.1 Dataset and Evaluation Metrics 3.2 Implementation Details 3.3 Results 4 Conclusion References A Patient-Specific Self-supervised Model for Automatic X-Ray/CT Registration 1 Introduction 2 Method 2.1 Pose Regression Model 2.2 Refinement Model 3 Experiments 3.1 Dataset 3.2 Experimental Design and Evaluation Metrics 4 Results 4.1 The Choice of Backbone and Ablation Study 4.2 Experimental Results on DRR Datasets and X-Ray Dataset 5 Conclusion and Discussion References Dose Guidance for Radiotherapy-Oriented Deep Learning Segmentation 1 Introduction 2 Methodology 2.1 Deep Learning-Based Dose Prediction 2.2 Dose Segmentation Loss (DOSELO) 3 Experiments and Results 3.1 Data and Model Training 3.2 Evaluation 3.3 Results 4 Discussion and Conclusion References Realistic Endoscopic Illumination Modeling for NeRF-Based Data Generation 1 Introduction 2 Method 2.1 Neural Radiance Fields 2.2 Extending NeRF for Endoscopy 3 Experiments and Results 3.1 Dataset 3.2 Implementation Details 3.3 Model Ablation Study 3.4 Data Generation 4 Conclusion References Multi-task Joint Prediction of Infant Cortical Morphological and Cognitive Development 1 Introduction 2 Methods 2.1 Network Architecture 2.2 Loss Design 3 Experiments 3.1 Dataset and Experimental Settings 3.2 Evaluation and Comparison 4 Conclusions References From Mesh Completion to AI Designed Crown 1 Introduction 2 Related Work 3 Method 3.1 Network Architecture 4 Experimental Results 4.1 Dataset and Preprocessing 4.2 Implementation Details 4.3 Evaluation and Metrics 4.4 Ablation Study 5 Conclusion References Spatiotemporal Incremental Mechanics Modeling of Facial Tissue Change 1 Introduction 2 Method 2.1 Data Representation 2.2 Network 3 Experiments and Results 3.1 Dataset 3.2 Implementation Details and Evaluation Metrics 3.3 Comparison with Baselines 3.4 Ablation Studies 4 Discussions and Conclusions References Analysis of Suture Force Simulations for Optimal Orientation of Rhomboid Skin Flaps 1 Introduction 2 Related Works 3 Methods 3.1 Constitutive Model of Skin 3.2 Cutting and Suturing 3.3 Triangulation 3.4 Validation 3.5 Suture Force Analysis 4 Results 4.1 Validation with UWG Simulator 4.2 Suture Force Analysis 4.3 Database Creation 5 Discussion and Conclusion References FLIm-Based in Vivo Classification of Residual Cancer in the Surgical Cavity During Transoral Robotic Surgery 1 Introduction 2 Method 2.1 FLIm Hardware and Data Acquisition 2.2 Patient Cohort and FLIm Data Labeling 2.3 FLIm Preprocessing 2.4 Novelty Detection Model 2.5 Classifier Training and Evaluation 2.6 Classifier Augmented Display 3 Results 4 Discussion 5 Conclusion References Spinal Nerve Segmentation Method and Dataset Construction in Endoscopic Surgical Scenarios 1 Introduction 2 Method 2.1 Details of Dataset 2.2 Network Architecture 2.3 IFA Module 2.4 CSA Module 3 Experiments 3.1 Implementation Details 3.2 Experimental Results 3.3 Ablation Experiments 4 Conclusion References Optical Coherence Elastography Needle for Biomechanical Characterization of Deep Tissue 1 Introduction 2 Methods 2.1 OCE Needle for Deep Tissue Indentation 2.2 OCE Measurement 2.3 Experimental Setup 2.4 Indentation Experiments 3 Results 4 Discussion and Conclusion References Semantic Segmentation of Surgical Hyperspectral Images Under Geometric Domain Shifts 1 Introduction 2 Materials and Methods 2.1 Deep Learning-Based Surgical Scene Segmentation 2.2 Experiments 3 Results 4 Discussion References Ultrasonic Tracking of a Rapid-Exchange Microcatheter with Simultaneous Pressure Sensing for Cardiovascular Interventions 1 Introduction 2 Materials and Methods 2.1 Ultrasonic Tracking and Concurrent Pressure Sensing System 2.2 Sensor and Microcatheter 2.3 Relative Tracking Accuracy 2.4 Impact of Microcatheter Orientation 2.5 In Vivo Validation 3 Results and Discussion References Self-distillation for Surgical Action Recognition 1 Introduction 2 Methods 2.1 Task Description and Dataset 2.2 Concept Overview 2.3 Implementation Details 3 Experiments and Results 4 Discussion References Encoding Surgical Videos as Latent Spatiotemporal Graphs for Object and Anatomy-Driven Reasoning 1 Introduction 2 Methods 2.1 Graph Construction 2.2 Graph Editing Module 2.3 Downstream Task Decoder 2.4 Training 3 Experiments and Results 3.1 Evaluation Tasks and Datasets 3.2 Baseline Methods 3.3 Main Experiments 3.4 Ablation Studies 4 Conclusion References Deep Reinforcement Learning Based System for Intraoperative Hyperspectral Video Autofocusing 1 Introduction 1.1 Background 1.2 Related Autofocusing Works 1.3 Contributions 2 Materials and Methods 2.1 Optical System 2.2 Datasets 2.3 Autofocus Policies 3 Results 4 Conclusion References Towards Multi-modal Anatomical Landmark Detection for Ultrasound-Guided Brain Tumor Resection with Contrastive Learning 1 Introduction 2 Related Work 3 Methods and Materials 3.1 Data and Landmark Annotation 3.2 Contrastive Learning Framework 3.3 Landmark Matching with a 2.5D Approach 3.4 Landmark Matching with 3D SIFT 4 Experimental Setup 4.1 Data Preprocessing 4.2 Loss Function 4.3 Implementation Details and Evaluation 5 Results 6 Discussion 7 Conclusions References ConTrack: Contextual Transformer for Device Tracking in X-Ray 1 Introduction 2 Methodology 2.1 Target Localization with Multi-template Feature Fusion 2.2 Localization Refinement with Context Flow 3 Experiments and Results 4 Conclusion References FocalErrorNet: Uncertainty-Aware Focal Modulation Network for Inter-modal Registration Error Estimation in Ultrasound-Guided Neurosurgery 1 Introduction 2 Methods and Materials 2.1 Dataset and Preprocessing 2.2 Network Architecture 2.3 Uncertainty Quantification 2.4 Experimental Setup and Implementation 3 Results 3.1 Error Regression Accuracy 3.2 Validation of the Uncertainty Evaluation 3.3 Robustness of the Proposed Model 4 Discussion 5 Conclusion References Optical Ultrasound Imaging for Endovascular Repair of Abdominal Aortic Aneurysms: A Pilot Study 1 Introduction 2 Methods 2.1 Device Design 2.2 Benchtop Imaging and Fenestration 2.3 In Vivo Imaging and Fenestration 3 Results 3.1 Benchtop Imaging and Fenestration 3.2 In Vivo Imaging and Fenestration 4 Discussion References Self-supervised Sim-to-Real Kinematics Reconstruction for Video-Based Assessment of Intraoperative Suturing Skills 1 Introduction 2 Methodology 2.1 LiveMAE 3 Experiments and Results 3.1 Understanding the Benefits of Kinematics 3.2 Evaluation of LiveMAE on Live Videos 4 Conclusion References Cascade Transformer Encoded Boundary-Aware Multibranch Fusion Networks for Real-Time and Accurate Colonoscopic Lesion Segmentation 1 Introduction 2 Approaches 2.1 Transformer Cascade Encoding 2.2 Boundary-Aware Multibranch Fusion Decoding 2.3 Hybrid Spatial-Frequency Loss 3 Experiments 4 Results and Discussion 5 Conclusion References Learnable Query Initialization for Surgical Instrument Instance Segmentation 1 Introduction 2 Related Work 3 Proposed Methodology 4 Results and Discussion 5 Conclusion References Author Index