MICCAI 2024 Papers — Page 7
International Conference on Medical Image Computing and Computer-Assisted Intervention · 856 papers
PET Image Denoising Based on 3D Denoising Diffusion Probabilistic Model: Evaluations on Total-Body Datasets
Yu, Boxiao (University of Florida), Gong, Kuang (University of Florida)
RestorationConvolutional Neural NetworkDiffusion modelImageBiomedical DataPositron Emission Tomography
🎯 What it does: A whole-body PET low-dose image denoising method based on a 3D diffusion probabilistic model (DDPM) is proposed;
pFLFE: Cross-silo Personalized Federated Learning via Feature Enhancement on Medical Image Segmentation
Xie, Luyuan (Peking University), Wu, Zhonghai (Peking University)
SegmentationFederated LearningContrastive LearningImageBiomedical Data
🎯 What it does: A cross-data-center personalized federated learning framework pFLFE is proposed for medical image segmentation, which includes two-stage training of feature enhancement and supervised learning.
PG-MLIF: Multimodal Low-rank Interaction Fusion Framework Integrating Pathological Images and Genomic Data for Cancer Prognosis Prediction
Pan, Xipeng (Guilin University of Electronic Technology), Yang, Huihua (Guilin University of Electronic Technology)
ClassificationConvolutional Neural NetworkMultimodalityBiomedical Data
🎯 What it does: The PG-MLIF framework is proposed, which integrates pathological images and genomic data for cancer survival prediction.
PhenDiff: Revealing Subtle Phenotypes with Diffusion Models in Real Images
Bourou, Anis (ENS), Genovesio, Auguste (ENS)
Image TranslationGenerationData SynthesisDrug DiscoveryDiffusion modelGenerative Adversarial NetworkImage
🎯 What it does: A multi-class image-to-image translation method named PhenDiff has been developed, utilizing conditional diffusion models to identify subtle cellular phenotype differences in real microscope images.
Phy-Diff: Physics-guided Hourglass Diffusion Model for Diffusion MRI Synthesis
Zhang, Juanhua (University of Dundee), Li, Chao (University of Dundee)
GenerationData SynthesisTransformerDiffusion modelBiomedical DataMagnetic Resonance ImagingDiffusion Tensor Imaging
🎯 What it does: A physics-guided diffusion model, Phy-Diff, is designed to generate high-quality diffusion MRI images constrained by b-values, b-vectors, and white matter trajectories.
Physical-priors-guided Aortic Dissection Detection using Non-Contrast-Enhanced CT images
Ding, Zhengyao (Zhejiang University), Huang, Zhengxing (First Affiliated Hospital of Zhejiang University School of Medicine)
ClassificationSegmentationGenerationConvolutional Neural NetworkTransformerGenerative Adversarial NetworkImageBiomedical DataComputed Tomography
🎯 What it does: A multi-task model based on physical priors is proposed, utilizing non-contrast CT images to achieve segmentation of aortic dissection, synthetic contrast CT generation, and classification diagnosis.
Physics informed neural networks for estimation of tissue properties from multi-echo configuration state MRI
Adams-Tew, Samuel I. (University of Utah), Joshi, Sarang (University of Utah)
Biomedical DataMagnetic Resonance ImagingPhysics Related
🎯 What it does: Develop a simulation of imaging signals for configuration states based on physical modeling, and use a fully connected neural network (PINN) to estimate T1, T2, T2*, and flip angle from multi-echo data in one go; evaluate the performance of different normalization methods and network structures in simulations and salt pork experiments.
Physics-Informed Deep Learning for Motion-Corrected Reconstruction of Quantitative Brain MRI
Eichhorn, Hannah (Helmholtz Munich), Schnabel, Julia A. (Technical University of Munich)
RestorationOptimizationComputational EfficiencyConvolutional Neural NetworkReinforcement LearningBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A deep learning motion correction method based on physical information, PHIMO, was designed and implemented for the automatic detection and exclusion of motion-affected k-space lines in quantitative MRI (T2* estimation).
PitVQA: Image-grounded Text Embedding LLM for Visual Question Answering in Pituitary Surgery
He, Runlong (University College London), Islam, Mobarakol (University College London)
TransformerLarge Language ModelVision Language ModelImageVideoText
🎯 What it does: The PitVQA dataset and the PitVQA-Net model are proposed for the visual question answering task in endonasal transsphenoidal pituitary surgery.
Pixel2Mechanics: Automated biomechanical simulations of high-resolution intervertebral discs from anisotropic MRIs
Natarajan, Sai (Galgo Medical S.L.), González Ballester, Miguel A. (BCN MedTech)
Graph Neural NetworkImageMeshBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Developed the Pixel2Mechanics automation process to reconstruct high-resolution intervertebral disc surface meshes from low-resolution clinical MRI and perform personalized finite element mechanical simulations.
Poisson Ordinal Network for Gleason Group Estimation Using Bi-Parametric MRI
Xu, Yinsong (University College London), Hu, Yipeng (King's College London)
ClassificationContrastive LearningMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a method for estimating Gleason grading of prostate cancer based on multimodal MRI, with the core being the Poisson Ordinal Network (PON), which enables direct prediction of grading from images.
Polyp-Mamba: Polyp Segmentation with Visual Mamba
Xu, Zhongxing (Weill Cornell Medicine Cornell University), Su, Jionglong (Xi'an Jiaotong-Liverpool University)
SegmentationImageMagnetic Resonance Imaging
🎯 What it does: Proposes the Polyp-Mamba framework, which combines the Scale-Aware Semantic (SAS) module and the Global Semantic Injection (GSI) module to achieve precise colorectal polyp segmentation through multi-scale semantic analysis and global information injection.
Pose-GuideNet: Automatic Scanning Guidance for Fetal Head Ultrasound from Pose Estimation
Men, Qianhui (University of Oxford), Noble, J. Alison (University of Oxford)
SegmentationPose EstimationConvolutional Neural NetworkContrastive LearningImageBiomedical DataUltrasound
🎯 What it does: Estimate the 3D pose of the fetal head in freehand ultrasound and provide a path to guide to the standard plane (TVP/TCP).
Position-Guided Prompt Learning for Anomaly Detection in Chest X-Rays
Sun, Zhichao (Wuhan University), Xu, Yongchao (Wuhan University)
Anomaly DetectionPrompt EngineeringContrastive LearningImage
🎯 What it does: Using a frozen CLIP model combined with position-guided learnable text and image prompts for chest X-ray anomaly detection, and proposing a structure-preserving synthetic anomaly method.
Prediction of Disease-Related Femur Shape Changes Using Geometric Encoding and Clinical Context on a Hip Disease CT Database
Li, Ganping (Nara Institute of Science and Technology), Sato, Yoshinobu (Nara Institute of Science and Technology)
SegmentationGenerationTransformerImagePoint CloudBiomedical DataComputed Tomography
🎯 What it does: This paper proposes a multimodal deep learning pipeline that utilizes geometric encoding and clinical context to predict shape changes in the affected femur from the normal side of the patient's femur.
Prior Activation Map Guided Cervical OCT Image Classification
Wang, Qingbin (Wuhan University), Ma, Yutao (Central China Normal University)
ClassificationExplainability and InterpretabilityConvolutional Neural NetworkImage
🎯 What it does: By constructing a prior activation map (PAM) and aligning it with the model-generated CAM during training, high-risk lesion classification of cervical OCT images is achieved.
PRISM: A Promptable and Robust Interactive Segmentation Model with Visual Prompts
Li, Hao (Shanghai Jiaotong University), Oguz, Ipek (Vanderbilt University)
SegmentationConvolutional Neural NetworkTransformerImageBiomedical DataComputed Tomography
🎯 What it does: This paper presents PRISM, a 3D medical image interactive segmentation model that supports multiple visual prompts and can continuously improve segmentation results over multiple iterations.
Privacy Protection in MRI Scans Using 3D Masked Autoencoders
Van der Goten, Lennart A. (KTH Royal Institute of Technology), Smith, Kevin (KTH Royal Institute of Technology)
SegmentationGenerationSafty and PrivacyAuto EncoderImageBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: A CP-MAE model based on 3D VQ-VAE and MAE has been developed, capable of high-resolution (256³) facial and skull de-identification of MRI scans without removing brain structures.
Probabilistic Temporal Prediction of Continuous Disease Trajectories and Treatment Effects Using Neural SDEs
Durso-Finley, Joshua (McGill University), Arbel, Tal (McGill University)
Recurrent Neural NetworkImageTabularBiomedical DataMagnetic Resonance ImagingStochastic Differential Equation
🎯 What it does: A probabilistic time model based on Neural Stochastic Differential Equations (NSDE) has been developed to predict the continuous disease progression trajectories and individual treatment effects of multiple sclerosis patients under different treatment regimens.
Progressive Growing of Patch Size: Resource-Efficient Curriculum Learning for Dense Prediction Tasks
Fischer, Stefan M. (Technical University Munich), Schnabel, Julia A. (Technical University of Munich)
SegmentationComputational EfficiencyConvolutional Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A progressive growing of patch size implicit curriculum learning strategy is proposed and integrated into the nnU-Net framework for medical image segmentation training.
Progressive Knowledge Distillation for Automatic Perfusion Parameter Maps Generation from Low Temporal Resolution CT Perfusion Images
Son, Moo Hyun (Hong Kong University of Science and Technology), Chen, Hao (Harvard Medical School)
GenerationKnowledge DistillationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: Proposes a Progressive Knowledge Distillation (PKD) framework that automatically generates perfusion parameter maps (PPM) from low temporal resolution CT perfusion images, significantly reducing radiation dose without manual AIF selection.
Progressively Correcting Soft Labels via Teacher Team for Knowledge Distillation in Medical Image Segmentation
Wang, Yaqi (Northeastern University), Zaiane, Osmar R. (Amii)
SegmentationKnowledge DistillationBiomedical Data
🎯 What it does: This paper proposes an evolutionary soft label correction knowledge distillation framework, PLC-KD, which uses a teacher team to correct soft labels in medical image segmentation, enhancing the performance of the student model.
Prompt Your Brain: Scaffold Prompt Tuning for Efficient Adaptation of fMRI Pre-trained Model
Dong, Zijian (National University of Singapore), Zhou, Juan Helen (National University of Singapore)
ClassificationDomain AdaptationExplainability and InterpretabilityComputational EfficiencyTransformerPrompt EngineeringBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: The ScaPT framework is proposed, achieving parameter-efficient prompt tuning on fMRI pre-trained models, allowing for the update of only about 2% of parameters in low-resource tasks.
Prompt-based Segmentation Model of Anatomical Structures and Lesions in CT Images
Ouyang, Xi (Shanghai United Imaging Intelligence Co., Ltd.), Shen, Dinggang (Shanghai United Imaging Intelligence Co., Ltd.)
SegmentationTransformerSupervised Fine-TuningPrompt EngineeringImageBiomedical DataComputed Tomography
🎯 What it does: This paper proposes a prompt-based multi-task CT image segmentation foundation model, utilizing a Transformer encoder and an automatic path module to achieve segmentation of up to 83 anatomical structures and lesions, and supports rapid transfer to few-shot new tasks.
Prompting Segment Anything Model with Domain-Adaptive Prototype for Generalizable Medical Image Segmentation
Wei, Zhikai (Wuhan University), Xu, Yongchao (Wuhan University)
SegmentationDomain AdaptationTransformerPrompt EngineeringImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: For single-source domain generalization (SDG) in medical image segmentation, a domain-adaptive prompt framework DAPSAM based on the large model Segment Anything Model (SAM) is proposed, which combines a universal adapter and a prototype-based prompt generator.
Prompting Vision-Language Models for Dental Notation Aware Abnormality Detection
Du, Chenlin (Tsinghua University), Lao, Qicheng (Ningbo Fregty Optoelectronics Technology Co., Ltd)
Object DetectionAnomaly DetectionTransformerSupervised Fine-TuningPrompt EngineeringVision Language ModelImage
🎯 What it does: By prompting a dental perspective-language model, tooth numbering is performed using the dental symbol system (FDI) to achieve multi-level tooth anomaly detection.
Prompting Whole Slide Image Based Genetic Biomarker Prediction
Zhang, Ling (DAMO Academy, Alibaba Group), Wang, Yan (East China Normal University)
ClassificationSegmentationExplainability and InterpretabilityTransformerLarge Language ModelPrompt EngineeringVision Language ModelImageBiomedical Data
🎯 What it does: The PromptBio framework is proposed, utilizing pathology text prompts generated by large language models and a visual language model to achieve foreground instance selection, fine-grained pathological component clustering, and interactive modeling, predicting genetic biomarkers such as MSI and BRAF from H&E slides.
PromptSmooth: Certifying Robustness of Medical Vision-Language Models via Prompt Learning
Hussein, Noor (Mohamed Bin Zayed University of Artificial Intelligence), Nandakumar, Karthik (Mohamed Bin Zayed University of Artificial Intelligence)
ClassificationPrompt EngineeringVision Language ModelImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Proposes the PromptSmooth method, which utilizes prompt learning to achieve certified robustness for pre-trained medical vision-language models without the need for retraining;
ProstNFound: Integrating Foundation Models with Ultrasound Domain Knowledge and Clinical Context for Robust Prostate Cancer Detection
Wilson, Paul F. R. (University of British Columbia), Mousavi, Parvin (Exact Imaging)
ClassificationObject DetectionConvolutional Neural NetworkPrompt EngineeringImageBiomedical DataUltrasound
🎯 What it does: A model named ProstNFound has been developed, which combines a medical foundation model with knowledge from the ultrasound field and clinical context for prostate cancer detection in microwave ultrasound images.
PULPo: Probabilistic Unsupervised Laplacian Pyramid Registration
Siegert, Leonard, Baumgartner, Christian F. (University of Tübingen)
Auto EncoderImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A probabilistic registration method PULPo is proposed, which can perform unsupervised registration through hierarchical variational inference and Laplacian pyramid model deformation fields.
PX2Tooth: Reconstructing the 3D Point Cloud Teeth from a Single Panoramic X-ray
Ma, Wen (Zhejiang University), Liu, Zuozhu (Zhejiang University)
SegmentationGenerationConvolutional Neural NetworkImagePoint CloudComputed Tomography
🎯 What it does: This paper proposes an end-to-end two-stage framework PX2Tooth, which directly reconstructs a single panoramic X-ray (PX) image into a three-dimensional dental point cloud;
Quality-Aware Fuzzy Min-Max Neural Networks for Dynamic Brain Network Analysis
Hou, Tao (Nantong University), Ding, Weiping (Nantong University)
ClassificationConvolutional Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A dynamic brain network analysis method based on Quality-Aware Fuzzy Min-Max Neural Network (QFMMNet) is proposed, which extracts dynamic functional connectivity features using multi-view learning, processes fuzzy information with multi-view fuzzy min-max neural networks, and considers the quality uncertainty of different windows through a quality-aware integration module to achieve classification of multi-window dynamic functional connectivity.
Quantitative Assessment of Thyroid Nodules through Ultrasound Imaging Analysis
Kim, Young-Min (Korea Advanced Institute of Science and Technology), Bae, Hyeon-Min (KAIST)
ClassificationRecognitionConvolutional Neural NetworkTransformerImageBiomedical DataUltrasound
🎯 What it does: This paper quantifies sound attenuation and sound speed using a single-probe ultrasound system to identify the malignancy of thyroid nodules.
QueryNet: A Unified Framework for Accurate Polyp Segmentation and Detection
Chai, Jiaxing, Li, Shaozi (Xiamen University)
Object DetectionSegmentationTransformerImageBiomedical Data
🎯 What it does: Proposes QueryNet, a unified framework that simultaneously performs polyp segmentation and detection.
Quest for Clone: Test-time Domain Adaptation for Medical Image Segmentation by Searching the Closest Clone in Latent Space
Basak, Hritam (Stony Brook University), Yin, Zhaozheng (Stony Brook University)
SegmentationDomain AdaptationAuto EncoderImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: This paper proposes a latent space search method based on Variational Autoencoders (VAE), which finds the most similar 'clones' in the source domain latent space for each target domain image during inference, thereby achieving Unsupervised Domain Adaptation (UDA) and Semi-Supervised Domain Adaptation (SSDA).
RadiomicsFill-Mammo: Synthetic Mammogram Mass Manipulation with Radiomics Features
Na, Inye (Sungkyunkwan University), Park, Hyunjin (VUNO Inc.)
Object DetectionGenerationData SynthesisPrompt EngineeringDiffusion modelImageTabularBiomedical Data
🎯 What it does: RadiomicsFill-Mammo has been developed, a synthetic tumor framework for mammography (Mammo) based on a stable diffusion model. This framework conditions the generation of realistic tumor images using mask images, ipsilateral breast images, clinical variables (BI-RADs, breast density), and low-dimensional radiomic features, and utilizes the synthetic images for data augmentation in detection models.
RDD-Net: Randomized Joint Data-Feature Augmentation and Deep-Shallow Feature Fusion Networks for Automated Diagnosis of Glaucoma
Tang, Yilin (Northwest University), Feng, Jun (Northwest University)
ClassificationDomain AdaptationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: A random joint data-feature enhancement and deep-shallow feature fusion network (RDD-Net) is proposed for automated glaucoma diagnosis.
Real-world Visual Navigation for Cardiac Ultrasound View Planning
Bao, Mingkun, Zhu, Haogang (Beihang University)
Convolutional Neural NetworkImageVideoBiomedical DataUltrasound
🎯 What it does: This paper proposes a real-time cardiac ultrasound visual navigation system based on view-invariant feature extractors and target consistency loss, aimed at helping novice ultrasound physicians quickly obtain standard views.
Realistic Surgical Image Dataset Generation Based On 3D Gaussian Splatting
Zeng, Tianle (University of Leeds), Jones, Dominic (University of Leeds)
Object DetectionSegmentationGenerationData SynthesisGaussian SplattingImageBiomedical Data
🎯 What it does: This paper designs a surgical image synthesis and annotation pipeline based on 3D Gaussian Splatting, which can synthesize surgical images of any pose from separately trained scene and surgical instrument models, and automatically generate pixel-level segmentation annotations.
Reciprocal Collaboration for Semi-supervised Medical Image Classification
Zeng, Qingjie (Northwestern Polytechnical University), Xia, Yong (Northwestern Polytechnical University)
ClassificationConvolutional Neural NetworkContrastive LearningImage
🎯 What it does: A semi-supervised medical image classification framework named ReCo is proposed, which utilizes complementary learning between a main network and an auxiliary network, and generates pseudo-labels and feature comparisons through opposite predictions.
Reducing Annotation Burden: Exploiting Image Knowledge for Few-Shot Medical Video Object Segmentation via Spatiotemporal Consistency Relearning
Zheng, Zixuan (MedAI Technology (Wuxi) Co. Ltd.), Mou, Lichao (Xidian University)
SegmentationImageVideoBiomedical Data
🎯 What it does: Under extremely low data conditions, few-shot medical video object segmentation is performed using existing medical image annotations and sparse video annotations;
Reference-free Axial Super-resolution of 3D Microscopy Images using Implicit Neural Representation with a 2D Diffusion Prior
Lee, Kyungryun (Korea University), Jeong, Won-Ki (Korea University)
RestorationSuper ResolutionDiffusion modelScore-based ModelImage
🎯 What it does: A no-reference axial super-resolution method based on implicit neural representation (INR) and two-dimensional diffusion model priors is proposed to restore isotropic resolution in 3D microscopy images.
Refining Intraocular Lens Power Calculation: A Multi-modal Framework Using Cross-layer Attention and Effective Channel Attention
Zhou, Qian, Wang, Yong (Wuhan University)
Convolutional Neural NetworkImageMultimodalityBiomedical Data
🎯 What it does: This study proposes a multimodal deep learning framework for predicting IOL optical power, integrating OCT images and biological features, and achieving accurate predictions through cross-layer attention and channel attention.
Region Attention Transformer for Medical Image Restoration
Yang, Zhiwen (Beihang University), Xu, Yan (Beihang University)
RestorationTransformerBiomedical DataComputed TomographyPositron Emission Tomography
🎯 What it does: A Transformer based on regional attention (RAT) is proposed for medical image restoration, utilizing SAM for dynamic segmentation of semantic regions and performing self-attention within those regions, combined with a focused region loss to enhance restoration quality.
Region-Specific Retrieval Augmentation for Longitudinal Visual Question Answering: A Mix-and-Match Paradigm
Yung, Ka-Wai (University College London), Mazomenos, Evangelos B. (University College London)
RecognitionRetrievalRecurrent Neural NetworkContrastive LearningImageMultimodalityBiomedical DataElectronic Health RecordsBenchmarkRetrieval-Augmented Generation
🎯 What it does: Designed and evaluated RegioMix, a region-level retrieval-enhanced longitudinal visual question answering framework that utilizes mixed matching to generate pseudo-difference descriptions and aligns the differences of input pairs with questions through dual alignment, thereby answering differential questions about chest X-ray images.
Reliable Multi-View Learning with Conformal Prediction for Aortic Stenosis Classification in Echocardiography
Gu, Ang Nan (University of British Columbia), Abolmaesumi, Purang (University of British Columbia)
ClassificationBiomedical DataUltrasound
🎯 What it does: A training method named RT4U is proposed, aimed at improving the classification of aortic stenosis in echocardiography by introducing uncertainty, especially when dealing with imperfect data.
Reliable Source Approximation: Source-Free Unsupervised Domain Adaptation for Vestibular Schwannoma MRI Segmentation
Zeng, Hongye (ShanghaiTech University), Fu, Huazhu (Agency for Science, Technology and Research (A*STAR))
SegmentationDomain AdaptationDiffusion modelImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a Reliable Source Approximation (RSA) method to achieve unsupervised adaptation of source domain models in the segmentation task of vestibular schwannoma MRI without source data.
Representation Learning with a Transformer-Based Detection Model for Localized Chest X-Ray Disease and Progression Detection
Eshraghi Dehaghani, Mehrdad (McMaster University), Moradi, Mehdi (University of Illinois Urbana-Champaign)
ClassificationObject DetectionRepresentation LearningConvolutional Neural NetworkTransformerImageBiomedical Data
🎯 What it does: Utilizing a Transformer-based detection model (DETR) to locate anatomical regions in chest X-ray images, and extracting feature vectors from this detection model, followed by local disease detection and monitoring of local disease progression in the same feature space.
Representing Functional Connectivity with Structural Detour: A New Perspective to Decipher Structure-Function Coupling Mechanism
Wei, Ziquan (University of North Carolina at Chapel Hill), Wu, Guorong (POSTECH)
ClassificationRecognitionGraph Neural NetworkTransformerBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: Proposed the 'Detour Connectivity' metric to capture the indirect relationship between functional connectivity (FC) and structural connectivity (SC), and based on this, designed the SC-FC Detour Network (SFDN) for graph neural network modeling.
Reprogramming Distillation for Medical Foundation Models
Zhou, Yuhang (Shanghai Jiao Tong University), Wang, Yanfeng (Shanghai Jiao Tong University)
Domain AdaptationKnowledge DistillationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: The Reprogramming Distillation (RD) framework is proposed, which maps the feature space of medical foundation models to the distribution of downstream tasks through training reprogrammable modules, and then co-trains with a lightweight student model while sharing classifiers to achieve efficient downstream transfer.
Resolving Variable Respiratory Motion From Unsorted 4D Computed Tomography
Huang, Yuliang (University College London), McClelland, Jamie R. (University College London)
OptimizationImageComputed Tomography
🎯 What it does: This paper proposes a 4DCT motion model based on hyper-gradient optimization that can estimate and compensate for motion errors caused by irregular breathing from unordered CT segments without external respiratory monitoring signals.
RET-CLIP: A Retinal Image Foundation Model Pre-trained with Clinical Diagnostic Reports
Du, Jiawei (Beijing Institute of Technology), Wang, Ningli (Beijing Institute of Ophthalmology)
ClassificationRepresentation LearningTransformerVision Language ModelContrastive LearningImageTextMultimodality
🎯 What it does: This paper proposes RET-CLIP, an audiovisual language foundation model based on CLIP, which is pre-trained using 193,865 clinical retinal images and diagnostic reports.
Rethinking Abdominal Organ Segmentation (RAOS) in the clinical scenario: A robustness evaluation benchmark with challenging cases
Luo, Xiangde (Stanford University), Wang, Guotai (University Of Electronic Science And Technology Of China)
SegmentationConvolutional Neural NetworkTransformerImageBiomedical DataComputed TomographyBenchmark
🎯 What it does: This paper constructs an abdominal multi-organ segmentation dataset named RAOS and establishes a robustness evaluation benchmark on this dataset.
Rethinking Autoencoders for Medical Anomaly Detection from A Theoretical Perspective
Cai, Yu (Hong Kong University of Science and Technology), Cheng, Kwang-Ting (Hong Kong University of Science and Technology)
Anomaly DetectionAuto EncoderBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Through information theory analysis, this paper clarifies the theoretical foundation of autoencoders in medical anomaly detection and proposes a strategy to achieve optimal performance by utilizing latent space entropy constraints.
Rethinking Cell Counting Methods: Decoupling Counting and Localization
Zheng, Zixuan (MedAI Technology (Wuxi) Co. Ltd.), Mou, Lichao (Xidian University)
Object DetectionSegmentationConvolutional Neural NetworkImageBiomedical DataBenchmark
🎯 What it does: A two-stage network architecture is proposed to decouple cell counting and localization. First, a counter network estimates the total count and generates a coarse density map based on intermediate features, and then a locator network generates a high-resolution cell position map. A global information propagation module is introduced to enhance counting accuracy.
Rethinking Histology Slide Digitization Workflows for Low-Resource Settings
Zehra, Talat (Jinnah Sindh Medical University), Nadeem, Saad (Memorial Sloan Kettering Cancer Center)
Image TranslationRestorationSuper ResolutionGenerative Adversarial NetworkOptical FlowImageVideo
🎯 What it does: This study proposes a cloud-based whole-slide digital workflow that utilizes low-cost microscope-captured 10X videos to automatically stitch, deblur, and reconstruct super-resolution images at 40X quality.
RetMIL: Retentive Multiple Instance Learning for Histopathological Whole Slide Image Classification
Chu, Hongbo (Tsinghua University), He, Yonghong (Tsinghua University)
ClassificationTransformerImage
🎯 What it does: Proposes RetMIL, a multi-instance learning framework based on a retention mechanism for the classification task of whole slide images (WSI).
Revisiting Deep Ensemble Uncertainty for Enhanced Medical Anomaly Detection
Gu, Yi (OMRON SINIC X Corporation), Chen, Hao (Harvard Medical School)
Anomaly DetectionAuto EncoderImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A medical anomaly detection framework D2UE based on multi-model uncertainty and feature space diversity is proposed.
Revisiting Self-Attention in Medical Transformers via Dependency Sparsification
Lin, Xian (Huazhong University of Science and Technology), Yu, Li (Huazhong University of Science and Technology)
SegmentationTransformerBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A Dependency Merging Attention (DMA) module is proposed, which dynamically aggregates similar feature tokens into prototypes, reducing redundant dependencies in ViT self-attention;
RIP-AV: Joint Representative Instance Pre-training with Context Aware Network for Retinal Artery/Vein Segmentation
Dai, Wei (Wenzhou Medical University), Su, Jianzhong (Wenzhou Medical University)
SegmentationConvolutional Neural NetworkGenerative Adversarial NetworkImage
🎯 What it does: A RIP-AV framework that combines Representative Instance Pre-training (RIP) with Context-Aware Networks (PCF, DA) is proposed for high-precision retinal artery-vein segmentation.
Robust Conformal Volume Estimation in 3D Medical Images
Lambert, Benjamin (University of Grenoble Alpes), Dojat, Michel (University of Grenoble Alpes)
SegmentationDomain AdaptationBiomedical DataMagnetic Resonance Imaging
🎯 What it does: The study uses Weighted Conformal Prediction (WCP) in three-dimensional medical image segmentation to estimate the prediction interval of lesion volume, addressing the issue of covariate shift between the calibration set and the test set.
Robust Semi-supervised Multimodal Medical Image Segmentation via Cross Modality Collaboration
Zhou, Xiaogen (Chinese University of Hong Kong), Dou, Qi (Huazhong University of Science and Technology)
SegmentationTransformerContrastive LearningMultimodalityBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A semi-supervised multimodal medical image segmentation framework is proposed, which enhances segmentation performance in scenarios with scarce annotations and modality mismatch by combining cross-modal collaboration and contrastive consistency learning.
Robustly Optimized Deep Feature Decoupling Network for Fatty Liver Diseases Detection
Huang, Peng (Fudan University), Wang, Xin (Chengdu University of Information Technology)
ClassificationRecognitionAdversarial AttackConvolutional Neural NetworkImageUltrasound
🎯 What it does: This paper proposes a deep learning framework that combines the Iterative Compressed Feature Decoupling Network (ICFDNet) with category accuracy adaptive adversarial training for fine-grained classification of fatty liver ultrasound images, significantly improving the model's recognition balance across different categories.
RoCoSDF: Row-Column Scanned Neural Signed Distance Fields for Freehand 3D Ultrasound Imaging Shape Reconstruction
Chen, Hongbo (ShanghaiTech University), Zheng, Rui (ShanghaiTech University)
RestorationImagePoint CloudUltrasound
🎯 What it does: RoCoSDF is proposed, a multi-view 3D ultrasound shape reconstruction framework based on row-column scanning neural SDF.
S-SAM: SVD-based Fine-Tuning of Segment Anything Model for Medical Image Segmentation
Paranjape, Jay N. (Johns Hopkins University), Patel, Vishal M. (Johns Hopkins University)
SegmentationSupervised Fine-TuningPrompt EngineeringImageBiomedical DataComputed TomographyUltrasound
🎯 What it does: A method named S-SAM is proposed, which achieves precise segmentation of medical images by performing singular value transformation only on the weight matrix of the image encoder of the Segment Anything Model (SAM), using category names as text prompts.
S-SYNTH: Knowledge-Based, Synthetic Generation of Skin Images
Kim, Andrea (U.S. Food and Drug Administration), Badano, Aldo (U.S. Food and Drug Administration)
SegmentationGenerationData SynthesisImage
🎯 What it does: A knowledge-based skin simulation framework S-SYNTH has been constructed to generate controllable parameter 3D skin models and synthetic images.
SALI: Short-term Alignment and Long-term Interaction Network for Colonoscopy Video Polyp Segmentation
Hu, Qiang (Huazhong University of Science and Technology), Wang, Zhiwei (Huazhong University of Science and Technology)
SegmentationConvolutional Neural NetworkVideo
🎯 What it does: A network called SALI, which combines short-term alignment and long-term interaction, is designed for real-time segmentation of polyps in endoscopic videos.
SAM Guided Task-Specific Enhanced Nuclei Segmentation in Digital Pathology
Swain, Bishal R. (Kumoh National Institute of Technology), Ko, Jaepil (Kumoh National Institute of Technology)
SegmentationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: Using the global features of the Segment Anything Model to guide the improved U-Net3+ (e-U-Net3+) for fine-grained nucleus segmentation, with the core being the X-Gated Fusion Block (GLU + Cross-Attention) to achieve dynamic fusion of global and local features.
SAM-Med3D-MoE: Towards a Non-Forgetting Segment Anything Model via Mixture of Experts for 3D Medical Image Segmentation
Wang, Guoan (Shanghai Artificial Intelligence Laboratory), Zhuang, Bohan (Monash University)
SegmentationTransformerMixture of ExpertsImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: In this work, the authors propose a Mixture of Experts framework based on SAM-Med3D—SAM-Med3D-MoE—which adds task-specific expert decoders and a lightweight gating network to the original base model, enabling the unification of general and specialized segmentation tasks without updating the original parameters.
SANGRE: a Shallow Attention Network Guided by Resolution Expansion for MR Image Segmentation
He, Ying (Queen Mary University of London), Zhang, Qianni (Queen Mary University of London)
SegmentationTransformerImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A MR image segmentation network named SANGRE is proposed, which combines a Transformer encoder with two new modules to achieve fine segmentation.
SaSaMIM: Synthetic Anatomical Semantics-Aware Masked Image Modeling for Colon Tumor Segmentation in Non-contrast Abdominal Computed Tomography
Dai, Pengyu (Institute of Integrated Research, Institute of Science Tokyo), Suzuki, Kenji (Institute of Integrated Research, Institute of Science Tokyo)
SegmentationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: This paper proposes SaSaMIM, a task-oriented self-supervised pre-training framework for colon tumor segmentation. It enhances early tumor segmentation performance on non-contrast CT by synthesizing intestinal and tumor semantics and performing masked image reconstruction under a dual-branch decoder in both frequency and spatial domains.
SBC-AL: Structure and Boundary Consistency-based Active Learning for Medical Image Segmentation
Zhou, Taimin (Beijing Institute of Technology), Chai, Senchun (Beijing Institute of Technology)
SegmentationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A proactive learning method based on structure and boundary consistency, SBC-AL, is proposed to enhance medical image segmentation results using Structure-aware Feature Prediction (SFP) and Attention Segmentation Refinement (ASR) modules, with sample queries conducted based on Structure Consistency Score (SCS) and Boundary Consistency Score (BCS).
SCMIL: Sparse Context-aware Multiple Instance Learning for Predicting Cancer Survival Probability Distribution in Whole Slide Images
Yang, Zekang (Chinese Academy of Sciences), Wang, Xiangdong (Chinese Academy of Sciences)
ClassificationTransformerImageBiomedical Data
🎯 What it does: A sparse context-aware multi-instance learning framework (SCMIL) is proposed to predict the cancer survival probability distribution of patients from whole slide images;
SDCL: Students Discrepancy-Informed Correction Learning for Semi-supervised Medical Image Segmentation
Song, Bentao (Southwest University of Science and Technology), Wang, Qingfeng (Southwest University of Science and Technology)
SegmentationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: In semi-supervised medical image segmentation, the authors propose the SDCL framework, which utilizes the segmentation differences of two student models to achieve self-correcting learning.
SDFPlane: Explicit Neural Surface Reconstruction of Deformable Tissues
Li, Hao (Shanghai Jiaotong University), Wang, Hesheng (Shanghai Jiaotong University)
RestorationSegmentationNeural Radiance FieldImageVideoMagnetic Resonance Imaging
🎯 What it does: A neural surface reconstruction method based on SDF, called SDFPlane, is designed for the rapid and high-precision reconstruction of deformable soft tissues from binocular endoscopic videos.
See, Predict, Plan: Diffusion for Procedure Planning in Robotic Surgical Videos
Zhao, Ziyuan (Nanyang Technological University), Zhou, S. Kevin (Hohai University)
Robotic IntelligenceTransformerDiffusion modelVideo
🎯 What it does: A multi-scale phase condition diffusion (MS-PCD) framework is proposed to predict action sequences that meet the target visual state in robot-assisted surgical videos, thereby achieving surgical procedure planning.
Seeing the Invisible: On Aortic Valve Reconstruction in Non-Contrast CT
Bujny, Mariusz (Graylight Imaging), Kostur, Marcin (University of Silesia)
SegmentationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: By registering high-quality contrast CT aortic valve segmentation results with corresponding non-contrast CT, semi-automatic Ground Truth is generated, and the nnU-Net model is trained using weakly supervised learning to achieve aortic valve segmentation in non-contrast CT.
SegMamba: Long-range Sequential Modeling Mamba For 3D Medical Image Segmentation
Xing, Zhaohu (Hong Kong University of Science and Technology (Guangzhou)), Zhu, Lei (Hong Kong University of Science and Technology)
SegmentationConvolutional Neural NetworkTransformerImageBiomedical DataComputed Tomography
🎯 What it does: SegMamba is proposed, a 3D medical image segmentation framework based on Mamba, utilizing three-directional Mamba (ToM) and gated spatial convolution (GSC) to achieve full-scale long-range modeling.
SegNeuron: 3D Neuron Instance Segmentation in Any EM Volume with a Generalist Model
Zhang, Yanchao (Chinese Academy of Sciences), Han, Hua (Chinese Academy of Sciences)
SegmentationTransformerSupervised Fine-TuningMultimodalityBiomedical Data
🎯 What it does: Construct a large-scale multi-resolution multimodal multi-species EM database EMNeuron, and based on this, train a general 3D neuron instance segmentation model SegNeuron to achieve zero-shot segmentation of any EM volume;
Self-guided Knowledge-injected Graph Neural Network for Alzheimer’s Diseases
Wang, Zhepeng, Zhang, Yanfu (University of Pittsburgh)
Explainability and InterpretabilityDrug DiscoveryGraph Neural NetworkLarge Language ModelMultimodalityGraphBiomedical DataDiffusion Tensor ImagingAlzheimer's Disease
🎯 What it does: A self-guided multimodal graph neural network is proposed, which automatically injects unstructured AD literature knowledge into brain networks, learns the importance of graphs and knowledge through masked learning, and performs graph enhancement to improve prediction performance and interpretability.
Self-Paced Sample Selection for Barely-Supervised Medical Image Segmentation
Su, Junming (Northeastern University), Zaiane, Osmar R. (Amii)
SegmentationContrastive LearningImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A framework called SPSS based on self-accelerated sample selection is proposed for extremely low-supervision medical image segmentation, aiming to alleviate the label scarcity problem by improving the quality of pseudo-labels generated by registration.
Self-supervised 3D Skeleton Completion for Vascular Structures
Ren, Jiaxiang (Stony Brook University), Ling, Haibin (Stony Brook University)
SegmentationGenerationConvolutional Neural NetworkContrastive LearningBiomedical DataComputed Tomography
🎯 What it does: A self-supervised 3D vascular skeleton completion method is proposed, which uses a synthetic skeleton breakpoint to train the network to reconnect broken skeletons.
Self-Supervised Contrastive Graph Views for Learning Neuron-level Circuit Network
Li, Junchi (Wuhan University), Du, Bo (Wuhan University)
ClassificationRepresentation LearningGraph Neural NetworkContrastive LearningGraphBiomedical Data
🎯 What it does: This paper proposes a self-supervised contrastive learning framework, FlyGCL, for learning topological features from neuron-level circuit networks, and utilizes these features for neuron classification and connectivity prediction.
Self-supervised Denoising and Bulk Motion Artifact Removal of 3D Optical Coherence Tomography Angiography of Awake Brain
Li, Zhenghong (Stony Brook University), Ling, Haibin (Stony Brook University)
RestorationConvolutional Neural NetworkAuto EncoderImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A blind-slice self-supervised learning framework called SOAD is proposed for the one-time removal of noise and bulk motion artifacts (BMA) from awake OCTA volumes while preserving capillary details.
Self-Supervised k-Space Regularization for Motion-Resolved Abdominal MRI Using Neural Implicit k-Space Representations
Spieker, Veronika (Helmholtz Munich), Schnabel, Julia A. (Technical University of Munich)
RestorationOptimizationImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a self-supervised k-space regularization method called PISCO, which is combined with neural implicit k-space representation (NIK) to achieve motion-resolved reconstruction of dynamic abdominal MRI.
Self-supervised Learning with Adaptive Graph Structure and Function Representation For Cross-Dataset Brain Disorder Diagnosis
Chen, Dongdong (Shanghai Jiao Tong University), Zhang, Lichi (Shanghai Jiao Tong University)
ClassificationDomain AdaptationRepresentation LearningGraph Neural NetworkContrastive LearningBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: A self-supervised graph contrastive learning framework is proposed for diagnosing brain diseases across datasets using learnable graph structures and multi-state encoders;
Self-supervised Vision Transformer are Scalable Generative Models for Domain Generalization
Doerrich, Sebastian (University of Bamberg), Ledig, Christian (University of Bamberg)
GenerationData SynthesisDomain AdaptationTransformerContrastive LearningImageBiomedical Data
🎯 What it does: A self-supervised ViT generative method is proposed to generate diverse synthetic pathological images through feature orthogonalization to enhance domain generalization performance.
SelfReg-UNet: Self-Regularized UNet for Medical Image Segmentation
Zhu, Wenhui (Arizona State University), Wang, Yalin (Clemson University)
SegmentationKnowledge DistillationConvolutional Neural NetworkImageMagnetic Resonance ImagingComputed Tomography
🎯 What it does: This paper proposes a self-regularized UNet that combines semantic consistency regularization and internal feature distillation losses to address the issues of supervision imbalance and feature redundancy in medical image segmentation.
Semantics-Aware Attention Guidance for Diagnosing Whole Slide Images
Liu, Kechun (University of Washington), Shapiro, Linda G. (UCLA)
ClassificationSegmentationExplainability and InterpretabilityConvolutional Neural NetworkTransformerContrastive LearningImage
🎯 What it does: Proposes the Semantics-Aware Attention Guidance (SAG) framework, which converts diagnostic entities into attention signals and uses flexible attention loss to guide whole slide image (WSI) diagnostic models to better focus on pathologically important areas.
Semi-Supervised Contrastive VAE for Disentanglement of Digital Pathology Images
Hasan, Mahmudul (Stony Brook University), Chen, Chao (Harvard Medical School)
ClassificationSegmentationGenerationAuto EncoderGenerative Adversarial NetworkContrastive LearningImage
🎯 What it does: This paper proposes a semi-supervised contrastive VAE (SS-cVAE) for digital pathology images to achieve separable interpretable latent spaces for detecting tumor-infiltrating lymphocytes (TIL).
Semi-Supervised Learning for Deep Causal Generative Models
Ibrahim, Yasin (University of Oxford), Kamnitsas, Konstantinos (University of Birmingham)
GenerationData SynthesisExplainability and InterpretabilityAuto EncoderImageBiomedical DataMagnetic Resonance ImagingElectronic Health Records
🎯 What it does: A semi-supervised deep causal generative model is proposed, capable of training with fully labeled, partially labeled, and unlabeled data, and generating realistic counterfactual images for missing labeled samples.
Semi-supervised Lymph Node Metastasis Classification with Pathology-guided Label Sharpening and Two-streamed Multi-scale Fusion
Li, Haoshen (Peking University), Jin, Dakai (Alibaba Group)
ClassificationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: A semi-supervised learning framework based on prior pathological reports is proposed, utilizing a two-stream 2.5D multi-scale feature fusion network to classify lymph node metastasis in esophageal cancer CT images.
Semi-supervised Segmentation through Rival Networks Collaboration with Saliency Map in Diabetic Retinopathy
Kim, Eunjin (VUNO Inc.), Park, Hyunjin (VUNO Inc.)
SegmentationConvolutional Neural NetworkImage
🎯 What it does: This paper proposes a semi-supervised multi-lesion segmentation method named RiCo, which achieves pixel-level segmentation of diabetic retinopathy through a competitive network and saliency maps.
Semi-supervised Tubular Structure Segmentation with Cross Geometry and Hausdorff Distance Consistency
Zhu, Ruiyun (Nagoya University), Mori, Kensaku (Aichi Institute of Technology)
SegmentationConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: A semi-supervised 3D tubular structure segmentation method based on cross geometric consistency and Hausdorff distance consistency is proposed, utilizing a dual-task of segmentation and distance transformation to enhance geometric feature learning.
SHAN: Shape Guided Network for Thyroid Nodule Ultrasound Cross-Domain Segmentation
Zhang, Ruixuan (Tianjin University), Li, Xuewei (Tianjin University)
SegmentationDomain AdaptationConvolutional Neural NetworkImageBiomedical DataUltrasound
🎯 What it does: This paper proposes a shape prior-based network (SHAN) that learns the elliptical distribution of thyroid nodules through global and neighborhood affine modules, constructing a domain-invariant latent feature space to achieve cross-center ultrasound image nodule segmentation.
ShapeMamba-EM: Fine-Tuning Foundation Model with Local Shape Descriptors and Mamba Blocks for 3D EM Image Segmentation
Shi, Ruohua (Peking University), Jiang, Tingting (Peking University)
SegmentationConvolutional Neural NetworkSupervised Fine-TuningImageBiomedical Data
🎯 What it does: We propose ShapeMamba-EM, aimed at fine-grained segmentation tasks for 3D electron microscopy images, achieving efficient fine-tuning by incorporating the Mamba Adapter and Local Shape Descriptors (LSD) into a 3D medical foundation model.
Shortcut Learning in Medical Image Segmentation
Lin, Manxi (Technical University of Denmark), Feragen, Aasa (DTU Compute)
SegmentationConvolutional Neural NetworkImageBiomedical DataUltrasound
🎯 What it does: This study investigates the shortcut learning phenomenon in medical image segmentation and demonstrates two common shortcuts: annotation markers in ultrasound images and center cropping + zero padding in skin lesion segmentation.
SiFT: A Serial Framework with Textual Guidance for Federated Learning
Li, Xuyang (Sun Yat-sen University), Wang, Ruixuan (Sun Yat-sen Univerisity)
Federated LearningRecurrent Neural NetworkLarge Language ModelImageBiomedical Data
🎯 What it does: A serial federated learning framework SiFT is proposed, combining continuous learning strategies with text prior knowledge to achieve model training for multiple clients without a central server.
SimBrainNet: Evaluating Brain Network Similarity for Attention Disorders
Das Chakladar, Debashis (Luleå University of Technology), Saini, Rajkumar (Luleå University of Technology)
Time SeriesBiomedical Data
🎯 What it does: For patients with Attention Deficit Hyperactivity Disorder (ADHD), a brain functional network based on multivariate transfer entropy (MTE) was constructed, and a new similarity measurement method, SimBrainNet, was proposed to compare the brain network similarities of different age groups under the same cognitive task.
Simplify Implant Depth Prediction as Video Grounding: A Texture Perceive Implant Depth Prediction Network
Yang, Xinquan (Shenzhen University), Deng, Yongqiang (Shenzhen University General Hospital)
Object DetectionDepth EstimationConvolutional Neural NetworkImageVideoComputed Tomography
🎯 What it does: This paper simplifies the prediction of dental implant depth into a video localization task and designs the Texture Perceive Implant Depth Prediction Network (TPNet);