MICCAI 2025 Papers — Page 5
International Conference on Medical Image Computing and Computer-Assisted Intervention · 1027 papers
HASD: Hierarchical Adaption for Pathology Slide-Level Domain-Shift
Liu, Jingsong (Technical University of Munich), Schüffler, Peter J. (Technical University of Munich)
Domain AdaptationSupervised Fine-TuningImageBiomedical Data
🎯 What it does: A Hierarchical Adaption for Pathology Slide-Level Domain-Shift (HASD) framework is proposed to address the domain shift problem in pathology images at the whole slide level.
Heterogeneous Masked Attention-Guided Path Convolution for Functional Brain Network Analysis
Xu, Jiakun (South China University of Technology), Xu, Xiangmin (Southern Medical University)
Graph Neural NetworkBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: A heterogeneous mask attention-guided path convolution model (HM-AGPC) has been developed for diagnostic analysis of functional brain networks.
Hierarchical Anatomy-Aware Guidance for Brain Tissue Microstructure Reconstruction from T1-weighted MRI
Li, Yuxing (Beijing Institute of Technology), Ye, Chuyang (Beijing Institute of Technology)
RestorationGenerationTransformerGenerative Adversarial NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A hierarchical anatomy-aware guided framework based on Transformer is proposed to reconstruct brain tissue microstructure mapping using T1-weighted MRI.
Hierarchical Characterization of Brain Dynamics via State Space-based Vector Quantization
Yang, Yanwu (University of Tübingen), Wolfers, Thomas (University of Tübingen)
TransformerAuto EncoderTime SeriesBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Designed and implemented a Hierarchical State Space-based Brain Dynamics Quantification Network (HST), quantifying brain functional states and transitions into discrete tokens.
Hierarchical Corpus-View-Category Refinement for Carotid Plaque Risk Grading in Ultrasound
Zhu, Zhiyuan (Shenzhen University), Yang, Xin (Shenzhen University)
ClassificationConvolutional Neural NetworkMixture of ExpertsContrastive LearningImageUltrasound
🎯 What it does: This study proposes a hierarchical Corpus-View-Category Refinement Framework (CVC-RF) for risk stratification of carotid plaques under multi-view ultrasound.
Hierarchical Feature Learning for Medical Point Clouds via State Space Model
Zhang, Guoqing (Tsinghua University), Li, Yang (Tsinghua University)
ClassificationRestorationSegmentationPoint Cloud
🎯 What it does: A hierarchical point cloud feature learning framework based on the state space model (SSM) is proposed for classification, reconstruction, and segmentation tasks of medical point clouds.
Hierarchical Part-based Generative Model for Realistic 3D Blood Vessel
Chen, Siqi (Tsinghua University), Li, Yang (Tsinghua University)
GenerationData SynthesisTransformerAuto EncoderPoint CloudMesh
🎯 What it does: This paper proposes a hierarchical, part-based 3D vascular generation framework that separates the modeling of global tree topology from local geometric details, achieving more refined and coherent vascular network synthesis.
Hierarchical Self-Supervised Adversarial Training for Robust Vision Models in Histopathology
Malik, Hashmat Shadab (Mohamed Bin Zayed University of Artificial Intelligence), Khan, Salman (Mohamed Bin Zayed University of Artificial Intelligence)
SegmentationAdversarial AttackConvolutional Neural NetworkGenerative Adversarial NetworkContrastive LearningImageBiomedical Data
🎯 What it does: This paper proposes Hierarchical Self-Supervised Adversarial Training (HSAT), which generates adversarial samples using the patient-slice-patch three-level relationship and trains the model under a multi-level contrastive learning framework to enhance the robustness of pathological images.
Hierarchical Spatio-temporal Segmentation Network for Ejection Fraction Estimation in Echocardiography Videos
Wang, Dongfang (Nanjing University of Science and Technology), Zhou, Tao (Nanjing University of Science and Technology)
SegmentationConvolutional Neural NetworkVideoBiomedical DataUltrasound
🎯 What it does: This paper proposes a hierarchical spatio-temporal segmentation network HSS-Net for the segmentation of the left ventricular endocardium in cardiac ultrasound videos, and accurately estimates the ejection fraction based on the segmentation results.
Hierarchical Vision-Language Learning for Medical Out-of-Distribution Detection
Lai, Runhe (Sun Yat-sen Univerisity), Wang, Ruixuan (Sun Yat-sen Univerisity)
Anomaly DetectionTransformerVision Language ModelContrastive LearningImageMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a medical OOD detection framework based on a hierarchical vision-language model, capable of identifying unknown diseases within known disease categories.
HieraSurg: Hierarchy-Aware Diffusion Model for Surgical Video Generation
Biagini, Diego (Technische Universitaet Muenchen), Farshad, Azade (Technische Universitaet Muenchen)
GenerationData SynthesisDiffusion modelVideo
🎯 What it does: HieraSurg is proposed—a two-stage hierarchical perception diffusion model for synthesizing realistic laparoscopic surgery videos from initial frames and surgical stage/action information.
High-Fidelity Unified One-to-Many Medical Image Synthesis via Text-Conditioned Latent Diffusion
Zhang, Youjian (JancsiTech), Yu, Gang (Zhejiang University)
GenerationData SynthesisDiffusion modelImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A text-conditioned latent diffusion framework is proposed to achieve one-to-many synthesis of multimodal medical images using a single model.
High-Order Progressive Trajectory Matching for Medical Image Dataset Distillation
Dong, Le, Mou, Lichao (Xidian University)
ClassificationData SynthesisKnowledge DistillationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: A high-order progressive trajectory matching method is proposed for distilling medical imaging datasets, generating privacy-friendly and efficient synthetic datasets.
High-Precision Mixed Feature Fusion Network Using Hypergraph Computation for Cervical Abnormal Cell Detection
Li, Jincheng (Nantong University), Zhao, Lili (Shanghai Jiao Tong University)
Object DetectionAnomaly DetectionGraph Neural NetworkImage
🎯 What it does: This paper proposes a hybrid feature fusion network based on hypergraph computation for high-precision detection of abnormal cervical cells in Thinprep cytology images.
High-Res Brain Source Imaging of MEG using a Vector Bayesian Beamformer with Noise learning
Gao, Tianyu (Beihang University), Ning, Xiaolin (Beihang University)
Time SeriesMagnetic Resonance Imaging
🎯 What it does: This paper proposes a brain imaging method based on vector Bayesian learning to improve the robustness of MEG vector beamforming in the presence of noise, limited samples, and correlated sources.
HiLa: Hierarchical Vision-Language Collaboration for Cancer Survival Prediction
Cui, Jiaqi (Sichuan University), Wang, Yan (East China Normal University)
ClassificationRecognitionTransformerLarge Language ModelVision Language ModelContrastive LearningMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A model based on hierarchical visual-language collaboration predicts the survival of cancer patients using multi-level visual features and diverse language prompts.
Historical Report Guided Bi-modal Concurrent Learning for Pathology Report Generation
Zhang, Ling (DAMO Academy, Alibaba Group), Wang, Yan (East China Normal University)
GenerationRetrievalTransformerImageTextMultimodalityRetrieval-Augmented Generation
🎯 What it does: A historical report-guided dual-modal parallel learning framework (BiGen) is proposed, which generates pathology reports for whole slide images (WSI) through visually and textually learnable joint encoding and knowledge retrieval.
Holistic White-light Polyp Classification via Alignment-free Dense Distillation of Auxiliary Optical Chromoendoscopy
Hu, Qiang (Huazhong University of Science and Technology), Wang, Zhiwei (Huazhong University of Science and Technology)
ClassificationKnowledge DistillationConvolutional Neural NetworkImage
🎯 What it does: A framework for classifying polyps in full-field white light endoscopy is proposed, utilizing alignment-free dense knowledge distillation to transfer diagnostic information from the NBI domain to the WLI domain.
HoloPointNet: A Deep Learning Framework for Efficient 3D Point Cloud Holography
Amrutkar, Ankit (RWTH Aachen University), Merhof, Dorit (Hannover Medical School)
GenerationComputational EfficiencyConvolutional Neural NetworkGenerative Adversarial NetworkPoint Cloud
🎯 What it does: HoloPointNet is proposed, a deep learning framework capable of directly processing 3D point clouds for holographic phase mapping, achieving rapid generation of multi-plane holographic images.
HRVVS: A High-resolution Video Vasculature Segmentation Network via Hierarchical Autoregressive Residual Priors
Yao, Xincheng (Hong Kong University of Science and Technology), Zhu, Lei (Hong Kong University of Science and Technology)
SegmentationTransformerOptical FlowVideo
🎯 What it does: The HRVVS model is proposed for liver vessel segmentation in high-resolution liver resection surgery videos, and the first Hepa-SEG dataset is constructed.
HWA-UNETR: Hierarchical Window Aggregate UNETR for 3D Multimodal Gastric Lesion Segmentation
Liang, Jiaming (South China University of Technology), Zhong, Xi (Guangzhou Medical University)
SegmentationConvolutional Neural NetworkTransformerMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A new framework for multimodal MRI segmentation of gastric cancer, HWA‑UNETR, has been proposed, and the first large-scale multimodal MRI dataset for gastric cancer, GCM 2025, has been publicly released.
Hybrid Graph Mamba: Unlocking Non-Euclidean Potential for Accurate Polyp Segmentation
Zhu, Yueyue (Northwestern Polytechnical University), Xia, Yong (Northwestern Polytechnical University)
SegmentationGraph Neural NetworkTransformerImageBiomedical Data
🎯 What it does: Proposes the Hybrid Graph Mamba (HGM) model for colon polyp segmentation, combining Mamba and GCN for multi-scale feature extraction and fusion.
Hybrid Local-Window-Attention–Assisted U-Net Model for Multimodal Medical-Image Segmentation
Kim, Jiwon (Sejong University), Noh, Wonjong (Hallym University)
SegmentationConvolutional Neural NetworkContrastive LearningMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A U-Net model based on Hybrid Local-Window Attention is proposed for multimodal brain tumor image segmentation.
Hybrid State-Space Models and Denoising Training for Unpaired Medical Image Synthesis
Zhang, Junming (Sun Yat-sen University), Jiang, Shancheng (Sun Yat-sen University)
GenerationData SynthesisConvolutional Neural NetworkGenerative Adversarial NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a hybrid network CRAViM and a cyclic denoising consistency training framework for unpaired medical image synthesis, addressing the issues of global structural distortion and local detail loss caused by traditional convolutional methods.
Hybrid-View Attention Network for Clinically Significant Prostate Cancer Classification in Transrectal Ultrasound
Feng, Zetian (Shenzhen University), Wang, Yi (Shenzhen University)
ClassificationConvolutional Neural NetworkTransformerImageBiomedical DataUltrasound
🎯 What it does: A hybrid perspective attention network is proposed for the classification of clinically significant prostate cancer in 3D TRUS images.
HybridMamba: A Dual-domain Mamba for 3D Medical Image Segmentation
Wu, Weitong (Hong Kong University of Science and Technology (Guangzhou)), Zhu, Lei (Hong Kong University of Science and Technology)
SegmentationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: The HybridMamba network is proposed for 3D medical image segmentation, combining local and global feature learning.
HyDA: Hypernetworks for Test Time Domain Adaptation in Medical Imaging Analysis
Serebro, Doron (Ben Gurion University of Negev), Riklin-Raviv, Tammy (Ben Gurion University of Negev)
Domain AdaptationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A hypernetwork-based test-time domain adaptation framework, HyDA, is proposed, which utilizes implicit domain representations to dynamically generate model parameters, enabling the model to adapt to different medical imaging domains without target domain training samples.
Hyperbolic Kernel GCN with Structure-Function Connectivity Coupling for Neurocognitive Impairment Analysis
Yang, Meimei (University of North Carolina at Chapel Hill), Liu, Mingxia (University of North Carolina at Chapel Hill)
Graph Neural NetworkMultimodalityBiomedical DataMagnetic Resonance ImagingDiffusion Tensor Imaging
🎯 What it does: A hyper-surface kernel graph convolutional network (HKC) that combines structure-function coupling is constructed to integrate DTI and fMRI data and predict HIV-related asymptomatic cognitive impairment (ANI).
Hypergraph Tversky-Aware Domain Incremental Learning for Brain Tumor Segmentation with Missing Modalities
Wang, Junze (Northeast Forestry University), Cong, Cong (Macquarie University)
SegmentationGraph Neural NetworkContrastive LearningMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A brain tumor segmentation framework based on domain incremental learning (ReHyDIL) is proposed, which can gradually learn new modalities in the absence of MRI modalities and maintain previous knowledge through a replay buffer.
Hypergraph-Guided Federated Distillation Learning for Efficient and Robust Multi-Center fMRI Data Analysis
Jin, Tao (Hangzhou Dianzi University), Gao, Yue (Tsinghua University)
Federated LearningSafty and PrivacyComputational EfficiencyKnowledge DistillationGraph Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A hypergraph-based federated distillation framework (HGFD) is proposed, which uses a hypergraph convolutional teacher network to extract high-order brain functional connectivity features and transfers these features to a lightweight MLP student network through knowledge distillation, achieving privacy-preserving sharing and efficient prediction of multi-center fMRI data.
HyperPath: Knowledge-Guided Hyperbolic Semantic Hierarchy Modeling for WSI Analysis
Huang, Peixiang (University of Hong Kong), Yu, Lequan (University of Hong Kong)
ClassificationVision Language ModelImageBiomedical Data
🎯 What it does: The HyperPath method is proposed, which utilizes text concept-guided hierarchical representation in hypercurvature space for the classification of whole slide images (WSI).
HyperSORT: Self-Organising Robust Training with hyper-networks
Joutard, Samuel (ImFusion), Prevost, Raphael (ImFusion)
RestorationSegmentationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: The HyperSORT framework is proposed, which uses a hypernetwork to predict UNet parameters based on the latent vectors of image and annotation variations, achieving self-organizing robust training.
ICE-PoGO: Improving Dynamic Panoramic Reconstruction of 4D ICE Imaging through Pose Graph Optimization
Herz, Sebastian (LUMA Vision), Wörz, Stefan (LUMA Vision)
OptimizationSimultaneous Localization and MappingImageBiomedical DataUltrasound
🎯 What it does: A 4D intracardiac echocardiography (ICE) dynamic panoramic reconstruction method based on pose graph optimization (PGO) has been implemented, which can simultaneously optimize the camera poses of multiple phases under the assumption of cyclic cardiac motion, thereby enhancing the field of view and accuracy of real-time navigation.
IFRFNet: Iterative Frequency Restoration-Fusion Network for Fast System Matrix Calibration on Magnetic Particle Image
Xu, Weixin (Beihang University), Mu, Wei (Beihang University)
RestorationSuper ResolutionConvolutional Neural NetworkImage
🎯 What it does: A frequency domain-based iterative recovery fusion network, IFRFNet, is proposed for rapid calibration of matrix (SM) in MPI systems.
IKAN: Interactive KAN with Modulation Fusion for Medical Image Segmentation
Liu, Sihan (Huazhong University of Science and Technology), Qiu, Wu (Huazhong University of Science and Technology)
SegmentationImageBiomedical DataComputed TomographyUltrasound
🎯 What it does: The IKAN framework is proposed, which achieves medical image segmentation through an interactive Kolmogorov-Arnold network.
IM-Fuse: A Mamba-based Fusion Block for Brain Tumor Segmentation with Incomplete Modalities
Pipoli, Vittorio (University of Modena and Reggio Emilia), Bolelli, Federico (University of Modena and Reggio Emilia)
SegmentationConvolutional Neural NetworkTransformerMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A multi-modal fusion block IM-Fuse based on Mamba has been constructed for brain tumor segmentation under missing modalities.
Implicit Deformable Medical Image Registration with Learnable Kernels
Fogarollo, Stefano (University of Innsbruck), Harders, Matthias (Medical University Innsbruck)
Image TranslationOptimizationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: This paper proposes an implicit medical image registration framework based on a learnable kernel, which reconstructs a global displacement field using sparse keypoint correspondences and supports interactive fine-tuning during testing.
Implicit U-KAN2.0: Dynamic, Efficient and Interpretable Medical Image Segmentation
Cheng, Chun-Wun (University of Cambridge), Aviles-Rivero, Angelica I. (Tsinghua University)
SegmentationExplainability and InterpretabilityComputational EfficiencyConvolutional Neural NetworkImageBiomedical DataUltrasoundOrdinary Differential Equation
🎯 What it does: An implicit U-KAN 2.0 network based on second-order neural ODE and MultiKAN is proposed for efficient and interpretable medical image segmentation.
Improved Baselines with Synchronized Encoding for Universal Medical Image Segmentation
Yang, Sihan, Sun, Jian (Xi'an Jiaotong University)
SegmentationConvolutional Neural NetworkTransformerImageBiomedical Data
🎯 What it does: A basic model for medical image segmentation based on SAM, called SyncSAM, is proposed, which uses a synchronized dual-branch encoder and a multi-scale dual-branch decoder, significantly improving the segmentation performance of medical images.
Improved Tumor Segmentation using Selective Synthetic Augmentation for Enhanced Surgical Planning in Breast MRI
Luna, Miguel (BeamWorks Inc.), Kim, Jaeil (Kyungpook National University)
SegmentationData SynthesisDiffusion modelImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper improves tumor segmentation for breast cancer surgical planning by inserting size-controllable synthetic tumors into real breast MRI.
Improving Autism Detection with Multimodal Behavioral Analysis
Saakyan, William (Bielefeld University), Drimalla, Hanna (Humboldt University of Berlin)
ClassificationAnomaly DetectionVideoMultimodalityAudio
🎯 What it does: Using standardized interactive videos with adult samples, this study extracts and models multimodal non-verbal features such as eye movement, facial expressions, vocal intonation, head movements, and heart rate variability, aiming to achieve computer-aided detection of Autism Spectrum Conditions (ASC).
Improving Medical Image Segmentation with Implicit Representation and Noisy Label Robustness
Kumari, Suruchi (Indian Institute of Technology Roorkee), Singh, Pravendra (Indian Institute of Technology Roorkee)
SegmentationSupervised Fine-TuningBiomedical DataComputed Tomography
🎯 What it does: A framework is proposed that combines a hierarchical channel attention encoder, a three-stage implicit decoder with noise-based index selection, and a high-frequency noise modulator to achieve high-resolution boundary refinement and noise label robustness in medical image segmentation using implicit neural representations.
Improving Motor Imagery EEG Signal Quality with Dynamic Visual Cues: An Innovative Paradigm and Dataset
Yue, Chenxi (Northwestern Polytechnical University), Zhang, Shu (Northwestern Polytechnical University)
ClassificationRecognitionConvolutional Neural NetworkTransformerTime SeriesBiomedical Data
🎯 What it does: A motion imagination EEG acquisition paradigm based on real human motion dynamic visual prompts is proposed, and a corresponding dataset is constructed.
Improving OCTA Imaging through Cross-Domain Adaptation: A Noise-Guided Framework Using Intralipid-Enhanced Rat Data
Yang, Bingyu (Beijing Institute of Technology), Cheng, Jun (Institute for Infocomm Research)
RestorationDomain AdaptationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: Through cross-domain self-supervised learning, high-quality OCTA data from mice after injection of Intralipid is transferred to monkey and human OCTA images, achieving B-scan denoising and face-to-face enhancement.
IMREPET: Implicit Neural Representation for Unsupervised Dynamic PET Reconstruction
Fan, Kailong, Wang, Linwei (Rochester Institute of Technology)
RestorationSuper ResolutionAuto EncoderBiomedical DataPositron Emission Tomography
🎯 What it does: This paper proposes a dynamic PET unsupervised reconstruction method IMREPET based on implicit neural representation (INR), which models space-time images directly on continuous functions, avoiding frame-by-frame reconstruction and complex regularization.
Incorporating Rather Than Eliminating: Achieving Fairness for Skin Disease Diagnosis Through Group-Specific Experts
Xu, Gelei (University of Notre Dame), Shi, Yiyu (University of Notre Dame)
ClassificationConvolutional Neural NetworkMixture of ExpertsImage
🎯 What it does: This paper presents FairMoE, a skin disease diagnosis model based on a hierarchical sparse expert (Mixture-of-Experts) structure, which utilizes sensitive attributes (such as skin color and age) for group learning, rather than completely eliminating sensitive information.
Indepth Integration of Multi-granularity Features from Dual-modal for Disease Classification
Wu, Yeli (Sun Yat-sen University), Zhang, Jianjia (Sun Yat-sen University)
ClassificationImageMultimodalityBiomedical Data
🎯 What it does: This paper proposes an IIMGF-Net model that achieves deep feature fusion of dual-modal medical images through cross-modal intra-granularity fusion and multi-granularity collaboration to improve disease classification accuracy.
Inferring Super-Resolved Gene Expression by Integrating Histology Images and Spatial Transcriptomics with HISTEX
Xue, Shuailin (Yunnan University), Min, Wenwen (Shenzhen Research Institute of Big Data)
GenerationSuper ResolutionTransformerSupervised Fine-TuningImageMultimodalityBiomedical Data
🎯 What it does: By integrating histological images with low-resolution spatial transcriptomics data, a super-resolution gene expression map is generated using bidirectional cross-attention and multi-instance learning.
Influence of Classification Task and Distribution Shift Type on OOD Detection in Fetal Ultrasound
Wong, Chun Kit (Technical University Of Denmark), Feragen, Aasa (DTU Compute)
ClassificationAnomaly DetectionConvolutional Neural NetworkImageUltrasound
🎯 What it does: This paper studies how the choice of classification task and the type of distribution shift affect out-of-distribution (OOD) detection and rejection prediction in fetal ultrasound images by comparing the performance of eight uncertainty quantification methods across four different classification tasks.
Information Bottleneck-based Causal Attention for Multi-label Medical Image Recognition
Cui, Xiaoxiao (Shandong University), Li, Shuo (Harbin Institute of Technology)
ClassificationRecognitionTransformerContrastive LearningImage
🎯 What it does: This paper proposes an Information Bottleneck-based Causal Attention (IBCA) model for multi-label medical image recognition, which effectively separates task-related and unrelated information, thereby improving diagnostic accuracy.
InstructX2X: An Interpretable Local Editing Model for Counterfactual Medical Image Generation
Min, Hyungi (Seoul National University), Cho, Sungzoon (Seoul National University)
GenerationData SynthesisExplainability and InterpretabilityDiffusion modelImageBiomedical Data
🎯 What it does: InstructX2X is proposed, an interpretable local editing model for generating counterfactual images of chest X-rays;
Instrument-Splatting: Controllable Photorealistic Reconstruction of Surgical Instruments Using Gaussian Splatting
Yang, Shuojue (National University of Singapore), Jin, Yueming (University of Oxford)
GenerationPose EstimationGaussian SplattingVideo
🎯 What it does: This paper proposes the Instrument-Splatting framework, which combines 3D Gaussian splatting with CAD models and utilizes real monocular surgical videos to achieve controllable and highly realistic digital twin reconstruction of surgical instruments.
Integrate Semantic Radiomics as Prior Evidence into Evidential Deep Learning for Pelvic Lipomatosis Diagnosis
Zhang, Zheran (Shanghai University), Wei, Zhipeng (Shanghai University)
ClassificationSegmentationExplainability and InterpretabilityTransformerSupervised Fine-TuningImageBiomedical DataComputed Tomography
🎯 What it does: By converting various clinical semantic radiomics (bladder-rectal fat distance, rectal roundness, bladder-prostate angle, relative pelvic fat volume) into probabilistic priors and integrating them into the Evidence Deep Learning (EDL) framework, accurate diagnosis of pelvic fat syndrome (PL) has been achieved;
Integrating meta-analysis in multi-modal brain studies with graph-based attention transformer
Choi, Hyoungshin (Sungkyunkwan University), Park, Hyunjin (VUNO Inc.)
ClassificationGraph Neural NetworkTransformerMultimodalityBiomedical DataMagnetic Resonance ImagingPositron Emission TomographyAlzheimer's Disease
🎯 What it does: A multimodal brain network graph Transformer model (MEGATF) is proposed, which utilizes meta-analysis information for the integration of functional and structural data of brain regions and disease classification.
Intelligent Virtual Sonographer (IVS): Enhancing Physician-Robot-Patient Communication
Song, Tianyu (Technical University of Munich), Navab, Nassir (Technische Universität München)
SegmentationRobotic IntelligenceTransformerLarge Language ModelImageBiomedical DataUltrasound
🎯 What it does: Designed and implemented an embedded conversational virtual ultrasound examination assistant IVS based on dual-instance LLM to enhance real-time communication among doctors, robots, and patients, and to achieve real-time command translation and robot control in an XR environment.
Inter-class separability loss for weakly supervised mutually exclusive multiclass segmentation of brain tumor lesions
Dhamale, Vivek (Indian Institute of Science), Sundaresan, Vaanathi (Indian Institute of Science)
SegmentationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Achieve multi-class brain tumor segmentation using weakly supervised class activation mapping (CAM).
Interactive Segmentation and Report Generation for CT Images
Gu, Yannian (Shanghai Jiao Tong University), Zhang, Xiaofan (Shanghai Jiaotong University)
SegmentationGenerationTransformerImageBiomedical DataComputed Tomography
🎯 What it does: An interactive 3D CT image lesion morphology report generation framework is proposed, which combines point annotation for fine segmentation and simultaneously generates a structured report.
Interpretable fMRI Captioning via Contrastive Learning
Shen, Vyacheslav (KAIST), Kim, Daeshik (KAIST)
RetrievalExplainability and InterpretabilityTransformerLarge Language ModelContrastive LearningMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This study proposes a two-stage contrastive learning framework that utilizes the BLIP-2 model to map fMRI signals into a visual-language embedding space, enabling text descriptions of stimulus images (fMRI captioning) and multimodal retrieval.
Intra- and Cross-View Enhancement for OCTA Imaging
Zeng, Jingbo (University of Hong Kong), Cheng, Jun (Institute for Infocomm Research)
Image TranslationRestorationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a deep learning pipeline based on ICENet, which directly generates OCTA images from repeated OCT B-scans and achieves image reconstruction through learning without high-quality labels.
Investigating Voxel-level Brain Age Prediction as a Pretext Task for Brain MRI Segmentation
Nasser, Tasneem (University of Calgary), El-Sheimy, Naser (University of Calgary)
SegmentationTransformerSupervised Fine-TuningImageBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: This paper proposes using voxel-level brain age prediction as a self-supervised pre-training task to improve brain MRI segmentation models, particularly enhancing performance in low-sample scenarios.
IP-CRR: Information Pursuit for Interpretable Classification of Chest Radiology Reports
Ge, Yuyan (University of Pennsylvania), Vidal, René (University of Pennsylvania)
ClassificationExplainability and InterpretabilityTransformerLarge Language ModelSupervised Fine-TuningTextBiomedical DataElectronic Health Records
🎯 What it does: An interpretable chest radiology report classification framework, IP-CRR, is proposed, which uses an information tracking method to progressively question facts in the report and make diagnoses based on the answers.
Is Hyperbolic Space All You Need for Medical Anomaly Detection?
Gonzalez-Jimenez, Alvaro (Lucerne University of Applied Sciences and Arts), Navarini, Alexander A. (University of Basel)
Anomaly DetectionConvolutional Neural NetworkMultimodalityBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A framework for medical anomaly detection and localization in negatively curved hyperbolic space is proposed.
ISAC: Redefining the Vascular Segmentation Paradigm through Mask Completion for Cross-Domain Generalization
Zhao, Tianyu (Huazhong University of Science and Technology), Yang, Xin (Shenzhen University)
SegmentationDomain AdaptationTransformerImage
🎯 What it does: The ISAC model is proposed, transforming vascular segmentation into a mask completion task based on sparse annotations.
ITAdaptor: Image-Tag Adapter Framework with Knowledge Enhancement for Radiology Report Generation
Ding, Shuaipeng (Chongqing Normal University), Wang, Chao (Chongqing Normal University)
GenerationKnowledge DistillationTransformerReinforcement LearningImageTextMultimodalityRetrieval-Augmented Generation
🎯 What it does: The ITAdaptor framework is proposed, which achieves radiology report generation through an image-label adapter and cross-modal knowledge enhancement.
Iterative Deployment Exposure for Unsupervised Out-of-Distribution Detection
Doorenbos, Lars (University of Bern), Márquez-Neila, Pablo (University of Bern)
Anomaly DetectionConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A novel unsupervised OOD detection framework called Iterative Deployment Exposure (IDE) is proposed, and a CSO method is designed to gradually utilize a small number of unlabeled OOD samples to improve the detector during deployment.
Iterative Foundation-Dedicated Learning: Optimized Key Frames, Prompts and Memories for Semi-Supervised Segmentation
Yin, Ziman (Beihang University), Tang, Zhenyu (Beihang University)
SegmentationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes an iterative semi-supervised medical image segmentation framework that combines the foundation model SAM2 with the specialized model UNet. It utilizes an uncertainty-aware module to adaptively select key frames and generate prompts, iteratively improving pseudo-labels and model performance.
ITMatch: Arch-Guided Semi-Supervised Tooth Arrangement via Iterative Confidence Evaluation
Wang, Chengyuan (Tsinghua University), Zhou, Yanheng (Tsinghua University)
Point Cloud
🎯 What it does: A semi-supervised regression framework called ITMatch is proposed for predicting tooth alignment.
KidneyDepth: A Synthetic Kidney Dataset for Metric Depth Estimation in Ureteroscopy
Oliva-Maza, Laura (German Aerospace Center), Triebel, Rudolph (German Aerospace Center)
Data SynthesisDepth EstimationTransformerSupervised Fine-TuningSimultaneous Localization and MappingImageComputed Tomography
🎯 What it does: This paper presents the KidneyDepth synthetic dataset and fine-tunes two existing MDE models (Depth Anything V2 and Zoedepth) based on it. The depth estimation accuracy is then evaluated on both synthetic and real ureteroscope images, and the inferred depth is integrated into an RGB-D SLAM system to enhance navigation performance.
KMUNet: A novel medical image segmentation model based on KAN and Mamba
Zhang, Hongsheng (Xiangtan University), Tang, Hongzhong (Xiangtan University)
SegmentationConvolutional Neural NetworkImageBiomedical DataUltrasound
🎯 What it does: This paper proposes a medical image segmentation network called KMUNet, which integrates KAN and Mamba. It uses a CNN encoder to extract local features in a U-shaped structure and introduces the Mamba module in the decoder to capture global dependencies, while also designing the KAN-PatchEmbed and KAN-SCA modules.
Knowing or Guessing? Robust Medical Visual Question Answering via Joint Consistency and Contrastive Learning
Jiang, Songtao (Zhejiang University), Liu, Zuozhu (Zhejiang University)
TransformerVision Language ModelContrastive LearningMultimodalityBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A multi-level perturbation RoMed medical VQA dataset was constructed, and a training framework combining Consistency and Contrastive Learning (CCL) was proposed to enhance the consistency of model responses under different phrasings.
Knowledge Bridges the Intent Gap: Contextual Fusion in Medical Fine-Grained Segmentation
Zhang, Hengyuan (National University of Defense Technology), Dou, Yong (National University of Defense Technology)
SegmentationTransformerSupervised Fine-TuningImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a Knowledge-Supported Segment Anything Model (KSAM) for fine-grained medical segmentation.
Knowledge Tree Driven Contextualized Instruction Tuning of Foundation Models for Epilepsy Drug Recommendation
Pham, Duy Khoa (Monash University), Ge, Zongyuan (Melbourne University)
Recommendation SystemDrug DiscoveryTransformerLarge Language ModelSupervised Fine-TuningImageTextBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Using a medical imaging-text-based foundational model, we predict the efficacy of antiepileptic drugs corresponding to MRI scans and reports of epilepsy patients through knowledge tree-driven contextual instruction tuning.
Knowledge-Enhanced Complementary Information Fusion with Temporal Heterogeneous Graph Learning for Disease Prediction
Yang, Zongbao (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences), Wang, Ruxin (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences)
Graph Neural NetworkTransformerContrastive LearningGraphTime SeriesBiomedical DataElectronic Health Records
🎯 What it does: This paper proposes the KCIF framework, which constructs a Temporal Heterogeneous Admission Graph (THAG) and integrates complementary information from laboratory tests and treatment events to achieve accurate predictions of ICU patient conditions.
Knowledge-guided Multi-scale Graph Mamba for Whole Slide Image Classification
Duan, Minghong (Fudan University), Song, Zhijian (Fudan University)
ClassificationGraph Neural NetworkImage
🎯 What it does: A knowledge-guided multi-scale graph Mamba framework is proposed for WSI classification.
Language of Stains: Tokenization Enhances Multiplex Immunofluorescence and Histology Image Synthesis
Sims, Zachary (Oregon Health and Science University), Chang, Young Hwan (Oregon Health and Science University)
GenerationData SynthesisTransformerVision Language ModelGenerative Adversarial NetworkImageMultimodality
🎯 What it does: A two-stage visual language framework is proposed, which utilizes VQGAN to quantize multi-channel immunofluorescence images into visual tokens, and then infers missing channels using a bidirectional Transformer masked language model, enhancing inference quality through H&E staining multimodal fusion.
Last Layer Laplacian Pseudocoresets for Robust Medical Image Analysis
Erick, Franciskus Xaverius (Friedrich-Alexander University Erlangen-Nürnberg), Kainz, Bernhard (Imperial College London)
ClassificationAnomaly DetectionImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A pseudo-core set framework based on the last layer Laplace approximation is designed and implemented to enhance robustness and uncertainty estimation in medical image classification.
Latent Motion Profiling for Annotation-free Cardiac Phase Detection in Adult and Fetal Echocardiography Videos
Yang, Yingyu (University of Oxford), Noble, J. Alison (University of Oxford)
Auto EncoderVideoBiomedical DataUltrasound
🎯 What it does: An unsupervised latent motion trajectory learning framework is proposed, utilizing an autoencoder for structural-motion decomposition of adult and fetal echocardiograms, automatically extracting ED/ES frames without manual annotation.
LDDMEm: Large Deformation Diffeomorphic Metric Embedding: Decoupling Shape Analysis from Image Registration
Fleishman, Greg M. (HHMI, Janelia Research Campus), Fletcher, P. Thomas (University of Virginia)
Biomedical DataAlzheimer's Disease
🎯 What it does: The study proposes the LDDMEm method, which embeds arbitrary deformation vector fields into the LDDMM framework, decoupling shape analysis and image registration.
LEAF: Latent Diffusion with Efficient Encoder Distillation for Aligned Features in Medical Image Segmentation
Huang, Qilin (Sun Yat-sen University), Zheng, Fudan (Sun Yat-sen University)
SegmentationKnowledge DistillationConvolutional Neural NetworkTransformerDiffusion modelImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A medical image segmentation framework called LEAF based on latent diffusion models is proposed, which replaces traditional noise prediction with direct segmentation map prediction during the fine-tuning phase, while introducing a feature distillation module aligned with the Transformer visual encoder.
Learning 3D Medical Image Models From Brain Functional Connectivity Network Supervision For Mental Disorder Diagnosis
Hu, Xingcan (University of Science and Technology of China), Xiao, Li (University of Science and Technology of China)
ClassificationRepresentation LearningTransformerContrastive LearningImageBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: A CINP framework based on contrastive learning is designed, utilizing 3D T1w structural MRI and functional connectivity networks (FCN) for image-network alignment pre-training, and proposes network prompting to achieve low-sample psychiatric diagnosis using only structural MRI.
Learning Concept-Driven Logical Rules for Interpretable and Generalizable Medical Image Classification
Gao, Yibo (Fudan University), Zhuang, Xiahai (Fudan University)
ClassificationExplainability and InterpretabilityImageBiomedical Data
🎯 What it does: A framework named Concept Rule Learner (CRL) is proposed, which uses binary visual concept learning Boolean logic rules to achieve interpretability and generalizability in medical image classification.
Learning Contrastive Multimodal Fusion with Improved Modality Dropout for Disease Detection and Prediction
Gu, Yi (OMRON SINIC X Corporation), Ma, Jiaxin (OMRON SINIC X Corporation)
ClassificationAnomaly DetectionTransformerContrastive LearningMultimodalityTabularBiomedical DataComputed TomographyElectronic Health Records
🎯 What it does: This study investigates a framework that combines improved modality loss (learnable modality tokens + simultaneous modality dropout) with multimodal contrastive learning for disease detection and prediction using CT images and electronic health records.
Learning Disease State from Noisy Ordinal Disease Progression Labels
Schmidt, Gustav (University of Tübingen), Müller, Sarah (University of Tübingen)
ClassificationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: This paper utilizes noisy ordinal disease progression labels (better, stable, worse) to train a Siamese network, learning continuous representations of disease states, and applies this representation for few-shot activity classification on an external OCT dataset.
Learning Explainable Imaging-Genetics Associations Related to a Neurological Disorder
Wang, Jueqi (Boston University), Venkataraman, Archana (Boston University)
Explainability and InterpretabilityTransformerBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: This paper presents NeuroPathX, an interpretable deep learning framework based on early fusion, designed to jointly analyze MRI structural images and genetic polymorphism (SNP) data to explore the interactions between neuroimaging and genetic pathways.
Learning Foundation Models from Multi-Organ Medical Images by Capturing Consistency and Diversity of Anatomical Structures
Hosseinzadeh Taher, Mohammad Reza (Arizona State University), Avinash, Gopal (GE HealthCare)
ClassificationSegmentationConvolutional Neural NetworkContrastive LearningImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: The Coda framework is proposed, utilizing self-supervised learning to simultaneously capture the cross-organ diversity and intra-organ consistency of multi-organ medical images, thereby constructing a high-quality foundational model.
Learning from Sparse Point Labels for Dense Carcinosis Localization in Advanced Ovarian Cancer Assessment
Zarin, Farahdiba (University of Strasbourg), Padoy, Nicolas (University of Strasbourg, CNRS, INSERM, ICube, UMR7357)
Object DetectionSegmentationConvolutional Neural NetworkImageVideoMagnetic Resonance Imaging
🎯 What it does: This study investigates how to train a model using only sparse point annotations to achieve full pixel localization of cancer nodules in laparoscopic images.
Learning Segmentation from Radiology Reports
Bassi, Pedro R. A. S. (John Hopkins University), Zhou, Zongwei (University of California, San Francisco)
SegmentationConvolutional Neural NetworkTransformerLarge Language ModelImageTextBiomedical DataComputed Tomography
🎯 What it does: A new framework called R-Super is proposed, which utilizes radiology reports for direct supervision of CT tumor segmentation. It maps tumor counts, sizes, and locations from the reports to voxel-level supervision through two new loss functions (volume loss and spherical loss) and is trained and evaluated on multiple datasets.
Learning with Explicit Topological Priors for Chest X-ray Rib Segmentation
Zhao, Xiaowei (Anhui University), Li, Chuanfu (Anhui University of Chinese Medicine)
SegmentationConvolutional Neural NetworkImage
🎯 What it does: In chest X-ray images, a framework is proposed to segment each rib by explicitly embedding the topological priors (connectivity and interaction) of the ribs into the network.
LEAVS: An LLM-based Labeler for Abdominal CT Supervision
Bigolin Lanfredi, Ricardo (National Institutes of Health), Summers, Ronald M. (Icahn School of Medicine at Mount Sinai)
ClassificationAnomaly DetectionTransformerLarge Language ModelPrompt EngineeringImageTextComputed TomographyChain-of-Thought
🎯 What it does: Utilizing a large language model through the zero-shot prompting system LEAVS to extract abnormal labels and urgency levels for various organs from abdominal CT reports, aimed at training imaging models.
Lesion-Aware Post-Training of Latent Diffusion Models for Synthesizing Diffusion MRI from CT Perfusion
Lee, Junhyeok (Seoul National University), Choi, Kyu Sung (Seoul National University)
Image TranslationGenerationData SynthesisDiffusion modelImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: Translate CT perfusion images of patients with acute ischemic stroke from CT to MRI, and incorporate lesion-aware image spatial loss during the post-training phase of the latent diffusion model.
Lesion-centered vision transformer for stroke outcome prediction from image and clinical data
Liu, Mingtian, Frindel, Carole (Universite Lyon 1)
ClassificationExplainability and InterpretabilityTransformerImageMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A lesion-centered Vision Transformer (LC-ViT) is proposed, which combines DWI MRI and 62 clinical variables to predict the modified Rankin Scale (mRS) outcomes 3 months after stroke.
LesionDiffusion: Towards Text-controlled General Lesion Synthesis
Lei, Wenhui (Shanghai Jiaotong University), Zhang, Xiaofan (Shanghai Jiaotong University)
SegmentationGenerationData SynthesisDiffusion modelImageBiomedical DataComputed Tomography
🎯 What it does: A text-controlled 3D CT lesion synthesis framework called LesionDiffusion is proposed, which can generate various lesions and their masks, and achieve fine-grained control of attributes through structured lesion reports.
Leveraging Diffusion Models for Continual Test-Time Adaptation in Fundus Image Classification
Liu, Mingsi (University of Electronic Science and Technology of China), Xu, Yanwu (South China University of Technology)
ClassificationDomain AdaptationDiffusion modelImageBiomedical Data
🎯 What it does: A continuous testing adaptation framework based on diffusion models, DiffCTA, is proposed for fundus image classification without modifying the source model.
Leveraging Semantic Asymmetry for Accurate Gross Tumor Volume Segmentation of Nasopharyngeal Carcinoma in Planning CT
Li, Zi (DAMO Academy, Alibaba Group), Jin, Dakai (Alibaba Group)
SegmentationConvolutional Neural NetworkContrastive LearningImageBiomedical DataComputed Tomography
🎯 What it does: A three-dimensional SATS method based on semantic asymmetry is proposed, which directly segments the total tumor volume of nasopharyngeal carcinoma on non-contrast-enhanced planning CT.
Leveraging Visual Prompt with Diffusion Adversarial Network for Radiotherapy Dose Prediction
Feng, Zhenghao (Sichuan University), Wang, Yan (East China Normal University)
SegmentationGenerationOptimizationConvolutional Neural NetworkPrompt EngineeringDiffusion modelGenerative Adversarial NetworkImageBiomedical DataComputed Tomography
🎯 What it does: ViPDose is proposed, a visual prompt-guided radiotherapy dose distribution prediction model that first compresses the actual dose map into a prompt vector using a prompt encoder, and then generates prompts based solely on structural images during inference through an improved diffusion adversarial network (DAN), achieving high-quality dose predictions.
LG-DBGL: Lateralization-Guided Dissociative Brain Graph Learning for Alzheimer’s Disease Identification
Ye, Jiazhen (Inner Mongolia University), Li, Jiapei (Inner Mongolia University)
ClassificationRecognitionGraph Neural NetworkGraphBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: A decentralized brain graph learning framework LG-DBGL based on hemispheric decoupling is proposed for Alzheimer's disease recognition.
LHU-Net: a Lean Hybrid U-Net for Cost-efficient, High-performance Volumetric Segmentation
Sadegheih, Yousef (University of Regensburg), Merhof, Dorit (Hannover Medical School)
SegmentationConvolutional Neural NetworkTransformerBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A lightweight hybrid U-Net called LHU-Net is proposed, which utilizes a hybrid attention module combining convolution and Transformer to achieve voxel-level medical image segmentation.
Lifespan Cortical Surface Reconstruction from Thick-Slice Clinical MRI
Dong, Xiuyu (University of North Carolina), Li, Gang (University of North Carolina at Chapel Hill)
RestorationSegmentationSuper ResolutionConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance ImagingAlzheimer's DiseaseOrdinary Differential Equation
🎯 What it does: A two-stage coarse-to-fine resolution iterative framework based on deep learning (DATAN) is proposed, capable of simultaneously performing super-resolution reconstruction and cortical surface reconstruction from thick-slice clinical MRI, covering the entire human lifespan.
Lightweight Data-Free Denoising for Detail-Preserving Biomedical Image Restoration
Chobola, Tomáš (Technical University of Munich), Peng, Tingying (Technical University of Munich)
RestorationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: A lightweight unsupervised denoising framework called Noise2Detail is proposed, which achieves high-quality denoising on single medical images through a three-stage process.
LIP-CAR: a learned inverse problem approach for medical imaging with contrast agent reduction
Evangelista, Davide (University of Bologna), Bianchi, Davide (Sun Yat-sen University)
OptimizationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A method is proposed to transform the problem of reducing contrast agent dosage into a constrained optimization inverse problem by learning a forward operator from high dose to low dose, thereby generating simulated images at high dose levels.