MICCAI 2024 Papers — Page 5
International Conference on Medical Image Computing and Computer-Assisted Intervention · 856 papers
Image Distillation for Safe Data Sharing in Histopathology
Li, Zhe (Friedrich-Alexander-Universität Erlangen-Nürnberg), Kainz, Bernhard (Imperial College London)
ClassificationData SynthesisKnowledge DistillationConvolutional Neural NetworkDiffusion modelContrastive LearningImageBiomedical Data
🎯 What it does: This paper proposes InfoDist, which utilizes latent diffusion models to generate readable synthetic images. It selects the samples with the highest information content through Infomap community detection, constructing a small-scale shareable synthetic dataset for downstream classification.
IMG-GCN: Interpretable Modularity-Guided Structure-Function Interactions Learning for Brain Cognition and Disorder Analysis
Xia, Jing (Nanyang Technological University), Rajapakse, Jagath C. (Nanyang Technological University)
ClassificationExplainability and InterpretabilityGraph Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes an interpretable modular guided graph convolutional network (IMG-GCN) that utilizes high-order interaction information connecting structure (SC) and function (FC) for brain function prediction (fluid cognition) and Parkinson's disease (PD) classification.
Immune-guided AI for Reproducible Regions of Interest Selection in Multiplex Immunofluorescence Pathology Imaging
Gautam, Tanishq (University of Texas MD Anderson Cancer Center), Castillo, Simon P. (University of Texas MD Anderson Cancer Center)
Object DetectionSegmentationConvolutional Neural NetworkSupervised Fine-TuningImageBiomedical Data
🎯 What it does: A set of end-to-end AI pipelines based on immune guidance has been constructed for the automated selection of regions of interest (ROI) in whole slide images of multiplex immunofluorescence (mIF), aiming for reproducibility and interpretability.
Implicit Representation Embraces Challenging Attributes of Pulmonary Airway Tree Structures
Zhang, Minghui (Shanghai Jiao Tong University), Gu, Yun (Shanghai Jiao Tong University)
SegmentationGenerationFlow-based ModelPoint CloudBiomedical DataComputed Tomography
🎯 What it does: This paper proposes a Deep Geometric Correspondence Implicit Network (DGCI) that utilizes implicit neural networks for high-fidelity modeling of the lung airway tree structure in continuous space, achieving skeletonization and fracture repair.
Improved Classification Learning from Highly Imbalanced Multi-Label Datasets of Inflamed Joints in [99mTc]Maraciclatide Imaging of Arthritic Patients by Natural Image and Diffusion Model Augmentation
Cobb, Robert (King's College London), Reader, Andrew J. (King's College London)
ClassificationSegmentationData SynthesisConvolutional Neural NetworkDiffusion modelImageBiomedical DataPositron Emission Tomography
🎯 What it does: Utilizing diffusion models and Perlin noise for data augmentation of limited [99mTc] maraciclatide hand images, combined with natural hand images to enhance multi-label classification performance for RA inflammation.
Improved Esophageal Varices Assessment from Non-Contrast CT Scans
Li, Chunli (Shengjing Hospital of China Medical University), Shi, Yu (Shengjing Hospital of China Medical University)
ClassificationSegmentationConvolutional Neural NetworkTransformerContrastive LearningImageBiomedical DataComputed Tomography
🎯 What it does: This paper proposes a Multi-Organ Collaborative Network (MOON) that utilizes non-contrast CT scans to simultaneously analyze the imaging features of the esophagus, liver, and spleen, achieving graded assessment of esophageal varices.
Improving cone-beam CT Image Quality with Knowledge Distillation-Enhanced Diffusion Model in Imbalanced Data Settings
Hwang, Joonil (Korea Advanced Institute of Science and Technology), Kim, Jin Sung (Yonsei University)
Image TranslationGenerationData SynthesisKnowledge DistillationDiffusion modelGenerative Adversarial NetworkImageBiomedical DataComputed Tomography
🎯 What it does: A self-training framework based on knowledge distillation is proposed, utilizing a small amount of paired CBCT-CT data combined with a large amount of unilateral CBCT data, employing a Brownian Bridge diffusion model to generate high-quality pseudo-CT;
Improving cross-domain brain tissue segmentation in fetal MRI with synthetic data
Zalevskyi, Vladyslav (University of Lausanne), Bach Cuadra, Meritxell (University of Lausanne)
SegmentationData SynthesisDomain AdaptationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper implements cross-domain fetal brain tissue segmentation through a synthetic data generation method based on domain randomization called FetalSynthSeg.
Improving Neoadjuvant Therapy Response Prediction by Integrating Longitudinal Mammogram Generation with Cross-Modal Radiological Reports: A Vision-Language Alignment-guided Model
Gao, Yuan (University Health Network), Mann, Ritse (Nanjing University of Information Science and Technology)
ClassificationGenerationData SynthesisDiffusion modelImageMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper studies a cross-modal radiology report-guided method for generating longitudinal mammography images to fill in the missing intra-treatment mammograms in clinical settings, and uses the generated images to predict the pathological complete response (pCR) status of breast cancer patients.
In vivo deep learning estimation of diffusion coefficients of nanoparticles
Kirkegaard, Julius B. (University of Copenhagen), Lauze, François
Object TrackingOptimizationConvolutional Neural NetworkAuto EncoderVideo
🎯 What it does: A deep learning-based estimator has been developed that can directly estimate the diffusion coefficient of nanoparticles from in vivo two-photon microscopy videos without the need for particle localization or trajectory tracking.
Incorporating Clinical Guidelines through Adapting Multi-modal Large Language Model for Prostate Cancer PI-RADS Scoring
Zhang, Tiantian (Chinese University of Hong Kong), Dou, Qi (Huazhong University of Science and Technology)
ClassificationDomain AdaptationKnowledge DistillationTransformerLarge Language ModelSupervised Fine-TuningImageMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes a general method for integrating the Clinical PI-RADs Guidelines (PICG) into MRI PI-RADs scoring, utilizing a multi-modal large language model (MLLM) for two-stage fine-tuning and injecting PICG information into existing scoring networks through feature distillation to improve scoring accuracy.
Inject Backdoor in Measured Data to Jeopardize Full-Stack Medical Image Analysis System
Yang, Ziyuan (Sichuan University), Zhang, Yi (Sichuan University)
SegmentationAdversarial AttackConvolutional Neural NetworkSupervised Fine-TuningImageBiomedical DataComputed Tomography
🎯 What it does: A learnable trigger generation method for medical imaging measurement data is proposed, achieving a pre-imaging backdoor attack without compromising reconstruction quality.
Insight: A Multi-Modal Diagnostic Pipeline using LLMs for Ocular Surface Disease Diagnosis
Yeh, Chun-Hsiao (University of California, Berkeley), Lin, Meng C. (University of California, Berkeley)
ClassificationSegmentationTransformerLarge Language ModelSupervised Fine-TuningPrompt EngineeringImageTextMultimodality
🎯 What it does: A multimodal diagnostic pipeline has been constructed to convert meibomian gland images into quantifiable morphological data, which is then integrated with clinical metadata for diagnosis and report generation using LLM.
InstaSAM: Instance-aware Segment Any Nuclei Model with Point Annotations
Nam, Siwoo (Daegu Gyeongbuk Institute of Science and Technology), Park, Sang Hyun (DGIST)
SegmentationDomain AdaptationBiomedical Data
🎯 What it does: A weakly supervised nuclear instance segmentation method called InstaSAM is proposed, which utilizes the strong representation of SAM and achieves accurate nuclear instance segmentation with only point annotations through an adapter layer and high-confidence pseudo-labels.
Integrating Clinical Knowledge into Concept Bottleneck Models
Pang, Winnie (Nanyang Technological University), Wen, Bihan (Nanyang Technological University)
ClassificationDomain AdaptationExplainability and InterpretabilityConvolutional Neural NetworkTransformerImageBiomedical Data
🎯 What it does: This paper incorporates the fine-grained knowledge of clinical experts regarding the importance of concepts into the Concept Bottleneck Model (CBM), guiding the model to prioritize clinically relevant concepts during prediction, thereby enhancing the model's interpretability and cross-domain robustness.
Integrative Graph-Transformer Framework for Histopathology Whole Slide Image Representation and Classification
Shi, Zhan (Stony Brook University), Wang, Fusheng (Stony Brook University)
ClassificationGraph Neural NetworkTransformerSupervised Fine-TuningImageBiomedical Data
🎯 What it does: This paper proposes an integrated Graph Neural Network and Global Self-Attention Graph-Transformer framework (IGT) for simultaneously extracting local spatial relationships and global contextual information from whole slide images (WSI) and performing binary/multi-class classification at the sliding window level.
Inter-Intra High-Order Brain Network for ASD Diagnosis via Functional MRIs
Han, Xiangmin (Tsinghua University), Gao, Yue (Tsinghua University)
ClassificationGraph Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A dual-layer high-order brain network (I²HBN) framework is proposed to simultaneously model high-order correlations within individuals (between brain regions) and between individuals (population level), thereby achieving functional MRI diagnosis of autism spectrum disorder (ASD).
Interpretable phenotypic profiling of 3D cellular morphodynamics
De Vries, Matt (Institute of Cancer Research), Bakal, Chris
ClassificationExplainability and InterpretabilityDrug DiscoveryGraph Neural NetworkTransformerContrastive LearningPoint CloudTime SeriesBiomedical Data
🎯 What it does: MorphoSense is proposed—a framework for interpretable 3D morphological dynamics time series classification based on multi-instance learning (MIL), capable of identifying key morphological changes during drug treatment or cell migration.
Interpretable Representation Learning of Cardiac MRI via Attribute Regularization
Di Folco, Maxime (Helmholtz Munich), Schnabel, Julia A. (Technical University of Munich)
Explainability and InterpretabilityRepresentation LearningAuto EncoderImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A model combining attribute regularization and Soft Introspective Variational Autoencoder (SIVAE) called AR-SIVAE is proposed for interpretable representation learning of cardiac MRI.
Interpretable Spatio-Temporal Embedding for Brain Structural-Effective Network with Ordinary Differential Equation
Tang, Haoteng (University of Texas Rio Grande Valley), Zhan, Liang (University of Southern California)
ClassificationExplainability and InterpretabilityGraph Neural NetworkBiomedical DataAlzheimer's DiseaseOrdinary Differential Equation
🎯 What it does: A brain effective network based on a dynamic causal model was constructed, and an interpretable spatiotemporal embedding ODE framework STE-ODE was proposed to learn the dynamic representation of the structure-effective network and perform clinical predictions.
Interpretable-by-design Deep Survival Analysis for Disease Progression Modeling
Gervelmeyer, Julius (University of Tübingen), Berens, Philipp (University of Tübingen)
Explainability and InterpretabilityConvolutional Neural NetworkImageAlzheimer's Disease
🎯 What it does: A explainable deep survival analysis model based on fundus images was designed and implemented to predict the progression time of age-related macular degeneration (AMD).
Intraoperative Registration by Cross-Modal Inverse Neural Rendering
Fehrentz, Maximilian (Harvard Medical School, Brigham and Women's Hospital), Haouchine, Nazim (Harvard Medical School, Brigham and Women's Hospital)
Image TranslationPose EstimationNeural Radiance FieldImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Using cross-modal inverse neural rendering technology to achieve single-view 3D/2D registration in neurosurgery. First, the preoperative MR grid learns the structure, and then the appearance of NeRF is adjusted in real-time on a single intraoperative image through a hypernetwork, utilizing differentiable rendering to solve for camera pose.
Intrapartum Ultrasound Image Segmentation of Pubic Symphysis and Fetal Head Using Dual Student-Teacher Framework with CNN-ViT Collaborative Learning
Jiang, Jianmei (Jinan University), Lekadir, Karim (Institució Catalana de Recerca i Estudis Avançats)
SegmentationConvolutional Neural NetworkTransformerImageUltrasound
🎯 What it does: A dual student-teacher combination CNN and Transformer semi-supervised image segmentation framework (DSTCT) is proposed for the segmentation of the pubic symphysis and fetal head in obstetric ultrasound images.
IOSSAM: Label Efficient Multi-View Prompt-Driven Tooth Segmentation
Huang, Xinrui (Shanghai Jiao Tong University), Wang, Xudong (Shanghai Jiao Tong University)
SegmentationTransformerPrompt EngineeringImageBiomedical Data
🎯 What it does: A multi-view weakly supervised tooth segmentation method based on SAM, IOSSAM, is proposed, utilizing 2D bounding box training prompts to achieve 3D IOS tooth segmentation and FDI labeling.
IPLC: Iterative Pseudo Label Correction Guided by SAM for Source-Free Domain Adaptation in Medical Image Segmentation
Zhang, Guoning, Wang, Guotai (University Of Electronic Science And Technology Of China)
SegmentationDomain AdaptationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: An iterative pseudo-label correction framework based on SAM is proposed to address the domain adaptation problem in unsupervised medical image segmentation.
Is this hard for you? Personalized human difficulty estimation for skin lesion diagnosis
Kampen, Peter Johannes Tejlgaard (Technical University of Denmark), Hannemose, Morten Rieger (Technical University of Denmark)
TransformerImageText
🎯 What it does: This study proposes a Transformer-based neural network model to predict the diagnostic/answer accuracy of doctors (or learners) on new cases, thereby enabling the estimation of individual difficulty.
Iterative Online Image Synthesis via Diffusion Model for Imbalanced Classification
Li, Shuhan (Hong Kong University of Science and Technology), Cheng, Kwang-Ting (Hong Kong University of Science and Technology)
ClassificationGenerationData SynthesisConvolutional Neural NetworkDiffusion modelImageBiomedical Data
🎯 What it does: An iterative online image synthesis framework (IOIS) is proposed to address the class imbalance problem in medical image classification.
IterMask2: Iterative Unsupervised Anomaly Segmentation via Spatial and Frequency Masking for Brain Lesions in MRI
Liang, Ziyun (University of Oxford), Kamnitsas, Konstantinos (University of Birmingham)
SegmentationAnomaly DetectionConvolutional Neural NetworkAuto EncoderImageMagnetic Resonance Imaging
🎯 What it does: This paper proposes an unsupervised brain lesion segmentation method, IterMask 2, which enhances the accuracy of abnormal region detection through iterative spatial mask refinement and frequency domain high-frequency information guided reconstruction.
Joint EM Image Denoising and Segmentation with Instance-aware Interaction
Wang, Zhicheng (University of Science and Technology of China), Xiong, Zhiwei (University of Science and Technology of China)
RestorationSegmentationConvolutional Neural NetworkImage
🎯 What it does: An interactive framework for joint denoising of electron microscopy images and instance segmentation is proposed, and the feature layers of the two tasks are mutually enhanced through an instance-aware embedding module.
Joint multi-task learning improves weakly-supervised biomarker prediction in computational pathology
El Nahhas, Omar S. M. (TUD Dresden University of Technology), Kather, Jakob Nikolas (EKFZ for Digital Health TU Dresden)
ClassificationTransformerImageBiomedical Data
🎯 What it does: This paper proposes a weakly supervised joint multi-task Transformer for directly predicting two biomarkers, MSI and HRD, from whole slide images, and learns tumor microenvironment features through auxiliary regression tasks.
Jumpstarting Surgical Computer Vision
Alapatt, Deepak (University of Strasbourg), Padoy, Nicolas (University of Strasbourg, CNRS, INSERM, ICube, UMR7357)
ClassificationRecognitionConvolutional Neural NetworkContrastive LearningVideo
🎯 What it does: This study investigates the impact of the composition of pre-training data in self-supervised learning on the performance of downstream tasks in surgical computer vision, and proposes a pre-training strategy based on multi-center and multi-surgery type data.
k-t Self-Consistency Diffusion: A Physics-Informed Model for Dynamic MR Imaging
Liu, Ye (ShanghaiTech University), Liang, Dong (Chinese Academy of Sciences)
RestorationOptimizationDiffusion modelScore-based ModelBiomedical DataMagnetic Resonance ImagingStochastic Differential Equation
🎯 What it does: A physical information diffusion model based on k-t Self-Consistency is proposed to achieve accelerated reconstruction of dynamic MRI (dMRI);
KARGEN: Knowledge-enhanced Automated Radiology Report Generation Using Large Language Models
Li, Yingshu (University of Sydney), Zhou, Luping (University of Sydney)
GenerationTransformerLarge Language ModelImageText
🎯 What it does: This paper proposes the KARGEN framework, which combines medical knowledge graphs with LLM for generating lung X-ray reports.
Keypoint Matching for Instrument-Free 3D Registration in Video-based Surgical Navigation
Baptista, Tânia (University of Coimbra), Barreto, Joao P. (University of Coimbra)
Object DetectionSegmentationPose EstimationVideo
🎯 What it does: The study utilizes only arthroscopic video and bone markers to achieve tool-free 3D registration and systematically evaluates the performance of various keypoint matching methods in arthroscopic images.
Knowledge-driven Subspace Fusion and Gradient Coordination for Multi-modal Learning
Zhang, Yupei (University of Hong Kong), Li, Chao (University of Dundee)
ClassificationOptimizationSpiking Neural NetworkMultimodalityBiomedical Data
🎯 What it does: A knowledge-driven subspace fusion and gradient coordination multimodal learning framework is proposed, which splits pathological images and genomic data into tumor subspaces and microenvironment subspaces for joint modeling, aiming to better capture the interaction information between tumors and the microenvironment.
Knowledge-grounded Adaptation Strategy for Vision-language Models: Building a Unique Case-set for Screening Mammograms for Residents Training
Urooj Khan, Aisha (Mayo Clinic), Banerjee, Imon (Arizona State University)
RetrievalDomain AdaptationTransformerSupervised Fine-TuningVision Language ModelContrastive LearningImageTextMagnetic Resonance Imaging
🎯 What it does: This paper proposes a knowledge-driven selective sampling strategy for adapting visual-language models (VLM) to the medical field (mammography) and enhances image-text retrieval performance through this strategy.
Knowledge-Guided Prompt Learning for Lifespan Brain MR Image Segmentation
Teng, Lin (ShanghaiTech University), Shen, Dinggang (Shanghai United Imaging Intelligence Co., Ltd.)
SegmentationTransformerPrompt EngineeringImageBiomedical DataMagnetic Resonance ImagingAlzheimer's Disease
🎯 What it does: This study proposes a two-stage brain MRI full lifecycle segmentation framework that achieves high-precision segmentation using Knowledge-Guided Prompt Learning (KGPL) based on a pre-trained model;
LaB-GATr: geometric algebra transformers for large biomedical surface and volume meshes
Suk, Julian (University of Twente), Wolterink, Jelmer M. (University of Twente)
Graph Neural NetworkTransformerMeshBiomedical Data
🎯 What it does: We propose LaB-GATr, a geometric algebra Transformer model capable of handling large-scale (bio)medical surface and volumetric meshes;
Label merge-and-split: A graph-colouring approach for memory-efficient brain parcellation
Kujawa, Aaron (King's College London), Vercauteren, Tom (King's College London)
SegmentationComputational EfficiencyConvolutional Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A label merge-and-split method is proposed, which first automatically merges spatially separated and similarly sized brain region labels through graph coloring, then trains a segmentation network for the merged labels using 3D U-Net, and finally restores the original labels using an influence region map after inference, significantly reducing the number of labels and lowering memory and time costs.
Label-guided Teacher for Surgical Phase Recognition via Knowledge Distillation
Guan, Jiale (Shanghai Jiao Tong University), Zheng, Guoyan (Shanghai Jiao Tong University)
RecognitionKnowledge DistillationTransformerContrastive LearningVideo
🎯 What it does: This paper proposes a label-guided teacher network based on knowledge distillation for surgical phase recognition.
Language-Enhanced Local-Global Aggregation Network for Multi-Organ Trauma Detection
Yu, Jianxun (Xidian University), Dou, Qi (Huazhong University of Science and Technology)
ClassificationObject DetectionTransformerLarge Language ModelPrompt EngineeringImageBiomedical DataComputed Tomography
🎯 What it does: A language-enhanced local-global aggregation network is proposed for multi-organ trauma detection, integrating local (organ-level) and global features of images, and introducing large language model text embeddings to enhance anatomical and trauma knowledge.
Laplacian Segmentation Networks Improve Epistemic Uncertainty Quantification
Zepf, Kilian (Technical University of Denmark), Feragen, Aasa (DTU Compute)
SegmentationAnomaly DetectionConvolutional Neural NetworkGaussian SplattingImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes Laplacian Segmentation Networks (LSN), which achieves simultaneous quantification of model (epistemic) and data (aleatoric) uncertainty by performing a Laplace approximation on the posterior of the weights of medical image segmentation models, thereby detecting out-of-distribution (OOD) anomalies in segmentation results.
Large-Scale 3D Infant Face Model
Schnabel, Till N. (ETH Zurich), Solenthaler, Barbara (ETH Zurich)
RestorationGenerationData SynthesisAuto EncoderPoint CloudMesh
🎯 What it does: A multi-nonlinear separable shape model INFACE based on large-scale unconstrained 3D infant facial scans is proposed and trained, supporting shape and appearance completion, single image reconstruction, and expression neutralization.
Latent Spaces Enable Transformer-Based Dose Prediction in Complex Radiotherapy Plans
Wang, Edward (Western University), Mattonen, Sarah A. (Western University)
GenerationOptimizationTransformerGenerative Adversarial NetworkImageBiomedical DataComputed Tomography
🎯 What it does: This study investigates the real-time prediction of multi-focal lung SABR dose distribution using a two-stage latent space framework (LDFormer) based on Transformer and VQ-VAE.
LaTiM: Longitudinal representation learning in continuous-time models to predict disease progression
Zeghlache, Rachid (LaTIM UMR 1101, Inserm), Lamard, Mathieu (LaTIM UMR 1101, Inserm)
Representation LearningRecurrent Neural NetworkContrastive LearningImageBiomedical DataOrdinary Differential Equation
🎯 What it does: This paper designs and pre-trains a time-aware neural ODE (NODE) model, utilizing self-supervised learning (SimCLR, BYOL) for representation learning of continuous-time medical images, and applies this model to predict the progression of diabetic retinopathy (DR).
LB-UNet: A Lightweight Boundary-assisted UNet for Skin Lesion Segmentation
Xu, Jiahao (Wuhan University), Tong, Lyuyang (Wuhan University)
SegmentationConvolutional Neural NetworkImage
🎯 What it does: A lightweight boundary-assisted UNet (LB-UNet) is proposed to specifically address the issues of unclear boundaries and resource constraints in skin lesion segmentation.
Learnable Skeleton-Based Medical Landmark Estimation with Graph Sparsity and Fiedler Regularizations
Wang, Yao (United-Imaging Research Institute of Intelligent Imaging), Qian, Zhen (United-Imaging Research Institute of Intelligent Imaging)
Pose EstimationGraph Neural NetworkImageBiomedical Data
🎯 What it does: A learnable skeletal structure based on Graph Convolutional Networks (FRGCN) has been developed, which automatically learns anatomical landmark point skeletons in medical images using Fiedler regularization and graph sparsification, and locates them based on heatmap regression.
Learning 3D Gaussians for Extremely Sparse-View Cone-Beam CT Reconstruction
Lin, Yiqun (Hong Kong University of Science and Technology), Li, Xiaomeng (Southern Medical University)
RestorationOptimizationGaussian SplattingImageBiomedical DataComputed Tomography
🎯 What it does: A DIF-Gaussian framework based on 3D Gaussian distribution is proposed for limited projection (≤10 views) CBCT reconstruction, incorporating Test-Time Optimization (TTO) during inference to enhance generalization capability.
Learning a Clinically-Relevant Concept Bottleneck for Lesion Detection in Breast Ultrasound
Bunnell, Arianna (University of Hawai'i), Shepherd, John A. (University of Hawai'i)
ClassificationObject DetectionSegmentationExplainability and InterpretabilityHyperparameter SearchConvolutional Neural NetworkSupervised Fine-TuningImageBiomedical DataUltrasound
🎯 What it does: A concept bottleneck model based on the BI-RADS vocabulary (BI-RADS-CBM) is proposed, achieving lesion detection, segmentation, interpretable feature prediction, and cancer classification in breast ultrasound images, while supporting concept-level interventions.
Learning from Partial Label Proportions for Whole Slide Image Segmentation
Matsuo, Shinnosuke (Kyushu University), Bise, Ryoma (Kyoto University Hospital)
SegmentationConvolutional Neural NetworkImage
🎯 What it does: This paper proposes a weakly supervised learning method for tumor subtype segmentation in whole slide images (WSI), utilizing a 'partial label proportion' (LPLP) that only provides the tumor subtype ratio to classify each image patch.
Learning Representations by Maximizing Mutual Information Across Views for Medical Image Segmentation
Weng, Weihao (University of Aizu), Zhu, Xin (University of Aizu)
SegmentationRepresentation LearningConvolutional Neural NetworkContrastive LearningImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A semi-supervised medical image segmentation method called Mutual Exemplar is proposed, which trains three identical structure networks on the same image with different intensity augmentations to maximize the mutual information of different view features, thereby enhancing segmentation performance.
Learning Temporally Equivariance for Degenerative Disease Progression in OCT by Predicting Future Representations
Emre, Taha (Medical University of Vienna), Bogunović, Hrvoje (Medical University of Vienna)
Representation LearningConvolutional Neural NetworkContrastive LearningTime SeriesBiomedical DataAlzheimer's Disease
🎯 What it does: This paper proposes a framework TC based on time-invariant contrastive self-supervised learning, which learns representations that evolve over time using unlabeled long-sequence OCT data and predicts the progression of AMD through these representations.
Learning to Segment Multiple Organs from Multimodal Partially Labeled Datasets
Liu, Hong (Eindhoven University of Technology), Wang, Liansheng (Shanghai Changhai Hospital)
SegmentationConvolutional Neural NetworkImageMultimodalityBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: This paper proposes a framework for learning multi-modal partially labeled multi-organ segmentation using cross-modal prior probability graphs;
Let Me DeCode You: Decoder Conditioning with Tabular Data
Szczepański, Tomasz (Sano Centre for Computational Medicine), Sitek, Arkadiusz (Massachusetts General Hospital)
SegmentationConvolutional Neural NetworkImageTabularComputed Tomography
🎯 What it does: A model called DeCode has been developed for decoder conditioning using shape features in 3D segmentation tasks, capable of achieving label-free conditioning during inference through learned embeddings.
Letting Osteocytes Teach SR-microCT Bone Lacunae Segmentation: A Feature Variation Distillation Method via Diffusion Denoising
Poles, Isabella (Politecnico di Milano), D’Arnese, Eleonora
SegmentationKnowledge DistillationConvolutional Neural NetworkDiffusion modelImageBiomedical DataComputed Tomography
🎯 What it does: Using unpaired bone tissue pathological images as a teacher model, knowledge distillation is performed on synchrotron radiation microCT (SR-microCT) images to train a model that only requires SR-microCT for bone cavity segmentation.
Leveraging Coarse-to-Fine Grained Representations in Contrastive Learning for Differential Medical Visual Question Answering
Liang, Xiao, Wang, Quan (Xidian University)
GenerationRepresentation LearningConvolutional Neural NetworkGraph Neural NetworkTransformerContrastive LearningImageTextBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A two-stage framework is proposed, first enhancing chest X-ray visual representation through anatomical knowledge graphs, then using multi-transformations to capture subtle differences between the main/reference images, and finally utilizing coarse-fine hierarchical contrastive learning to align the differing features with keywords and generate difference answers.
Leveraging Image Captions for Selective Whole Slide Image Annotation
Qiu, Jingna (FAU Erlangen Nurnberg), Breininger, Katharina (Friedrich-Alexander-Universität Erlangen-Nürnberg)
Object DetectionSegmentationConvolutional Neural NetworkTransformerContrastive LearningImageTextMultimodalityMagnetic Resonance Imaging
🎯 What it does: A method is proposed to extract task-specific class prototypes based on an image-caption database, and to select WSI annotation areas using prototype similarity, thereby reducing the cost of full-image annotation.
Leveraging the Mahalanobis Distance to enhance Unsupervised Brain MRI Anomaly Detection
Behrendt, Finn (Hamburg University of Technology), Schlaefer, Alexander (Hamburg University of Technology)
SegmentationAnomaly DetectionConvolutional Neural NetworkDiffusion modelImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes using Mahalanobis distance to perform pixel-level anomaly scoring on multiple pseudo-healthy reconstruction results generated by diffusion models, in order to improve the segmentation accuracy of unsupervised anomaly detection in brain MRI.
LGA: A Language Guide Adapter for Advancing the SAM Model’s Capabilities in Medical Image Segmentation
Hu, Jihong (Ritsumeikan University), Chen, Yen-Wei (Ritsumeikan University)
SegmentationTransformerSupervised Fine-TuningImageTextMultimodalityComputed Tomography
🎯 What it does: A parameter-efficient fine-tuning framework LGA is proposed, which integrates the text features of medical reports extracted by BERT with the SAM image encoder through a lightweight adapter (Feature Fusion Module) to achieve multi-modal medical image segmentation.
LGRNet: Local-Global Reciprocal Network for Uterine Fibroid Segmentation in Ultrasound Videos
Xu, Huihui (Shanghai Artificial Intelligence Laboratory), Zhu, Lei (Hong Kong University of Science and Technology)
SegmentationVideoBiomedical DataUltrasound
🎯 What it does: This paper proposes a Local-Global Recursive Network (LGRNet) for the segmentation of uterine fibroids in ultrasound videos, and constructs the UFUV dataset with 100 videos and a total of 5000 annotated frames for the first time.
LGS: A Light-weight 4D Gaussian Splatting for Efficient Surgical Scene Reconstruction
Liu, Hengyu (Chinese University of Hong Kong), Yuan, Yixuan (Southern University of Science and Technology)
CompressionComputational EfficiencyKnowledge DistillationGaussian SplattingImage
🎯 What it does: A lightweight 4D Gaussian splatting reconstruction framework LGS is proposed for efficient real-time surgical scene reconstruction.
LIBR+: Improving Intraoperative Liver Registration by Learning the Residual of Biomechanics-Based Deformable Registration
Wang, Dingrong (Rochester Institute of Technology), Wang, Linwei (Rochester Institute of Technology)
Graph Neural NetworkMesh
🎯 What it does: This paper studies a hybrid registration method LIBR+ that combines linear elastic biomechanics with deep learning for intraoperative liver registration.
LIDIA: Precise Liver Tumor Diagnosis on Multi-Phase Contrast-Enhanced CT via Iterative Fusion and Asymmetric Contrastive Learning
Huang, Wei (Sichuan University), Yan, Ke (China Medical University)
ClassificationSegmentationConvolutional Neural NetworkTransformerContrastive LearningImageComputed Tomography
🎯 What it does: This paper proposes a precise diagnostic network for liver tumors, LIDIA, based on multi-phase contrast-enhanced CT. It achieves multi-phase information fusion and heterogeneity suppression through an iterative fusion module and asymmetric contrast learning, thereby completing tumor segmentation and classification.
Lifelong Histopathology Whole Slide Image Retrieval via Distance Consistency Rehearsal
Zhu, Xinyu (Beihang University), Zheng, Yushan (Beihang University)
RetrievalTransformerContrastive LearningImage
🎯 What it does: A lifelong whole slide image retrieval framework (LWSR) is proposed, which achieves continuous learning and avoids catastrophic forgetting through a local memory pool and a Distance Consistency Replay (DCR) module.
LighTDiff: Surgical Endoscopic Image Low-Light Enhancement with T-Diffusion
Chen, Tong (University of Sydney), Zhou, Luping (University of Sydney)
RestorationDiffusion modelImage
🎯 What it does: The research objective is to improve the quality of low-light images in surgical endoscopy through a lightweight diffusion model.
LiverUSRecon: Automatic 3D Reconstruction and Volumetry of the Liver with a Few Partial Ultrasound Scans
Sivayogaraj, Kaushalya (University of Moratuwa), Liyanaarachchi, Rukshani (University of Moratuwa)
SegmentationGenerationConvolutional Neural NetworkTransformerImageBiomedical DataComputed TomographyUltrasound
🎯 What it does: This paper utilizes three partially visible transverse ultrasound scans and a statistical shape model (SSM) constructed based on CT to generate shape parameters using TransUNet segmentation and multilayer perceptron regression, ultimately achieving automatic reconstruction of a 3D liver model and volume calculation.
LKM-UNet: Large Kernel Vision Mamba UNet for Medical Image Segmentation
Wang, Jinhong (Zhejiang University), Wu, Jian (Zhejiang University)
SegmentationConvolutional Neural NetworkImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A UNet based on large kernel Mamba (LKM-UNet) is proposed for medical image segmentation, utilizing a multi-scale large kernel structure to enhance local and global feature modeling.
LLM-guided Multi-modal Multiple Instance Learning for 5-year Overall Survival Prediction of Lung Cancer
Kim, Kyungwon (Yonsei University), Hwang, Dosik (DoctorWorks Co., Ltd.)
TransformerLarge Language ModelImageTextMultimodalityComputed Tomography
🎯 What it does: This paper proposes a multimodal multi-instance learning framework guided by LLM, which jointly predicts the 5-year overall survival rate of lung cancer patients using CT, pathological slices, and clinical text information.
LM-UNet: Whole-body PET-CT Lesion Segmentation with Dual-Modality-based Annotations Driven by Latent Mamba U-Net
Liu, Anglin (ShanghaiTech University), Shen, Dinggang (Shanghai United Imaging Intelligence Co., Ltd.)
SegmentationConvolutional Neural NetworkMultimodalityBiomedical DataComputed TomographyPositron Emission Tomography
🎯 What it does: This study proposes the LM-UNet framework, which employs a dual-modal (PET+CT) labeling strategy, a CNN-Mamba hybrid encoder, and an anatomical enhancement module to achieve automatic segmentation of whole-body PET-CT lesions, significantly improving DSC and HD95 on both a self-built dual-modal dataset and the public autoPET dataset.
Lobar Lung Density Embeddings with a Transformer encoder (LobTe) to predict emphysema progression in COPD
Curiale, Ariel H. (Brigham and Women's Hospital), San José Estépar, Raúl (Brigham and Women's Hospital)
TransformerImageBiomedical DataComputed Tomography
🎯 What it does: A Transformer-based LobTe model is proposed and validated to predict the progression of emphysema (%LAA-950 change) in patients with chronic obstructive pulmonary disease (COPD) within five years from baseline CT images.
Location embedding based pairwise distance learning for fine-grained diagnosis of urinary stones
Jin, Qiangguo (Northwestern Polytechnical University), Lu, Yueh-Hsun (Taipei Medical University)
ClassificationSegmentationConvolutional Neural NetworkImage
🎯 What it does: This paper proposes the LEPD-Net model, which combines low-dose abdominal X-rays, position embedding, and pairwise distance learning for fine-grained diagnosis of urinary stones.
LoCI-DiffCom: Longitudinal Consistency-Informed Diffusion Model for 3D Infant Brain Image Completion
Zhu, Zihao (ShanghaiTech University), Zhang, Han (Sichuan University)
RestorationGenerationTransformerDiffusion modelImageBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper proposes LoCI-DiffCom, a long-term consistency-guided diffusion model that utilizes the fusion of images from different time points to complete missing infant brain MRI images.
LOMIA-T: A Transformer-based LOngitudinal Medical Image Analysis framework for predicting treatment response of esophageal cancer
Sun, Yuchen (Beijing Institute of Technology), Zhang, Shuaitong (Beijing Institute of Technology)
TransformerContrastive LearningImageBiomedical DataComputed Tomography
🎯 What it does: A Transformer-based longitudinal medical image analysis framework LOMIA-T was developed to predict the preoperative treatment response (pCR) in patients with esophageal squamous cell carcinoma.
Longitudinal Mammogram Risk Prediction
Karaman, Batuhan K. (Cornell University and Cornell Tech), Sabuncu, Mert R. (Cornell University and Cornell Tech)
ClassificationTransformerImageBiomedical Data
🎯 What it does: A Transformer model called LoMaR was constructed and evaluated, capable of receiving an arbitrary number of longitudinal mammography images and predicting future breast cancer risk.
Longitudinally Consistent Individualized Prediction of Infant Cortical Morphological Development
Yuan, Xinrui (University of North Carolina at Chapel Hill), Li, Gang (University of North Carolina at Chapel Hill)
GenerationData SynthesisConvolutional Neural NetworkAuto EncoderTime SeriesBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A longitudinally consistent individualized prediction framework is proposed, utilizing surface-based deep networks to generate complete cortical property (thickness, sulcal depth, area, myelination) trajectories of infant scalp at any time point within 0-24 months.
Loose Lesion Location Self-supervision Enhanced Colorectal Cancer Diagnosis
Gao, Tianhong (Zhejiang University), Feng, Zunlei (Zhejiang University)
ClassificationObject DetectionSegmentationExplainability and InterpretabilityConvolutional Neural NetworkImageBiomedical DataComputed Tomography
🎯 What it does: A self-supervised framework for diagnosing colorectal cancer with loosely defined lesion locations is proposed, utilizing temporal and cross-modal consistency constraints to achieve precise lesion localization with minimal annotations, and enhancing diagnostic reliability through a masking loop mechanism.
Lost in Tracking: Uncertainty-guided Cardiac Cine MRI Segmentation at Right Ventricle Base
Zhao, Yidong (Delft University of Technology), Tao, Qian (Delft University of Technology)
SegmentationConvolutional Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: This paper addresses the challenge of segmenting the right ventricular (RV) base in cardiac magnetic resonance imaging by re-labeling the RV base (including RVOT) in the ACDC dataset and proposing a dual-encoding U-Net model guided by motion uncertainty.
Low-Rank Continual Pyramid Vision Transformer: Incrementally Segment Whole-Body Organs in CT with Light-Weighted Adaptation
Zhu, Vince (DAMO Academy Alibaba Group), Jin, Dakai (Alibaba Group)
SegmentationTransformerSupervised Fine-TuningImageBiomedical DataComputed Tomography
🎯 What it does: The paper proposes a continuous organ segmentation framework LoCo-PVT based on PVT, achieving a no-forgetting extension by freezing the base model and adding low-rank adapters for each new task.
Low-Rank Mixture-of-Experts for Continual Medical Image Segmentation
Chen, Qian (Peking University), Lu, Yanye (Peking University)
SegmentationTransformerMixture of ExpertsBiomedical DataComputed Tomography
🎯 What it does: A low-rank Mixture-of-Experts network has been developed for continuous learning in medical image segmentation, addressing the problem of catastrophic forgetting.
Low-Shot Prompt Tuning for Multiple Instance Learning based Histology Classification
Chikontwe, Philip (Harvard Medical School), Park, Sang Hyun (DGIST)
ClassificationTransformerPrompt EngineeringVision Language ModelImageBiomedical Data
🎯 What it does: This paper proposes a method for multi-instance learning (MIL) for whole slide image (WSI) pathology classification through prompt tuning with a small number of samples.
LS+: Informed Label Smoothing for Improving Calibration in Medical Image Classification
Sambyal, Abhishek Singh (Indian Institute of Technology Ropar), Bathula, Deepti R. (Indian Institute of Technology Palakkad)
ClassificationConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: The LS+ (Label Smoothing Plus) method is proposed, which constructs class-specific soft labels using the class accuracy from the validation set to improve the calibration performance of medical image classification.
LSSNet: A Method for Colon Polyp Segmentation Based on Local Feature Supplementation and Shallow Feature Supplementation
Wang, Wei (Changsha University of Science and Technology), Wang, Xin (Chengdu University of Information Technology)
SegmentationTransformerImage
🎯 What it does: LSSNet is designed for colon polyp segmentation, utilizing two structures: local feature compensation and shallow feature compensation.
LUCIDA: Low-dose Universal-tissue CT Image Domain Adaptation For Medical Segmentation
Chen, Yixin (Peking University), Xie, Zhaoheng (United Imaging Healthcare, Co., Ltd.)
SegmentationDomain AdaptationConvolutional Neural NetworkGenerative Adversarial NetworkImageBiomedical DataComputed Tomography
🎯 What it does: LUCIDA is proposed, an unsupervised domain adaptation method for achieving annotation-free multi-organ segmentation in low-dose CT images.
M2Fusion: Multi-time Multimodal Fusion for Prediction of Pathological Complete Response in Breast Cancer
Zhang, Song (Chinese Academy of Sciences), Tian, Jie (Beihang University)
ClassificationSegmentationData-Centric LearningConvolutional Neural NetworkTransformerContrastive LearningImageMultimodalityBiomedical DataMagnetic Resonance Imaging
🎯 What it does: A multi-time multi-modal fusion model M2Fusion is proposed, which combines pre/post MRI and whole slide images (WSI) to predict the pathological complete response (pCR) of breast cancer patients after neoadjuvant chemotherapy.
M4oE: A Foundation Model for Medical Multimodal Image Segmentation with Mixture of Experts
Jiang, Yufeng (Hong Kong Baptist University), Shen, Yiqing (Johns Hopkins University)
SegmentationTransformerMixture of ExpertsAuto EncoderImageMultimodalityMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A multi-modal medical image segmentation foundation model M oE based on SwinUNet is proposed, utilizing modality-specific experts and a gating network to achieve cross-modal learning and parameter efficiency.
MAdapter: A Better Interaction between Image and Language for Medical Image Segmentation
Zhang, Xu, Zhang, Lefei (Wuhan University)
SegmentationConvolutional Neural NetworkVision Language ModelImageTextBiomedical DataComputed Tomography
🎯 What it does: A medical image segmentation framework based on a bidirectional MAdapter is proposed, which utilizes pre-trained visual and language encoders to extract multi-scale features, and achieves the fusion and alignment of visual and textual information through a bidirectional interaction module and a dedicated decoder, ultimately generating pixel-level segmentation results.
MambaMIL: Enhancing Long Sequence Modeling with Sequence Reordering in Computational Pathology
Yang, Shu (Hong Kong University of Science and Technology), Chen, Hao (Harvard Medical School)
ClassificationComputational EfficiencyConvolutional Neural NetworkImageBiomedical Data
🎯 What it does: MambaMIL is proposed within the framework of Multiple Instance Learning (MIL), utilizing Mamba (a state space model with linear time complexity) and the newly designed Sequence Reordering Mamba (SR-Mamba) to model long sequence instances in Whole Slide Images (WSI), thereby enhancing feature extraction and classification/prognosis performance in pathological imaging.
Mammo-CLIP: A Vision Language Foundation Model to Enhance Data Efficiency and Robustness in Mammography
Ghosh, Shantanu (Boston University), Batmanghelich, Kayhan (University of Pittsburgh)
ClassificationObject DetectionTransformerVision Language ModelContrastive LearningImageTextBiomedical DataMagnetic Resonance Imaging
🎯 What it does: We propose Mammo-CLIP, a vision-language foundation model for mammography, and develop Mammo-FActOR for weakly supervised attribute localization and interpretation.
MARVEL: MR Fingerprinting with Additional micRoVascular Estimates using bidirectional LSTMs
Barrier, Antoine (University of Grenoble Alpes), Christen, Thomas (University of Grenoble Alpes)
Recurrent Neural NetworkBiomedical DataMagnetic Resonance Imaging
🎯 What it does: Using the bSSFP MRF sequence, combined with the frequency distribution of simulated microvascular structures, a bidirectional long short-term memory network (BiLSTM) is employed to rapidly reconstruct full-scale quantitative images containing T1, T2, B1, frequency offset, CBV, and vascular radius.
Mask-Enhanced Segment Anything Model for Tumor Lesion Semantic Segmentation
Shi, Hairong (Beihang University), Liu, Si (Peking University)
SegmentationTransformerImageBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: For 3D tumor lesion semantic segmentation, we propose the Mask-Enhanced SAM (M-SAM) framework, which includes a Mask-Enhanced Adapter (MEA) and an iterative refinement scheme; it effectively injects tumor location priors by inserting the adapter into the SAM-Med3D framework and gradually refines the segmentation masks.
Mask-Free Neuron Concept Annotation for Interpreting Neural Networks in Medical Domain
Kim, Hyeon Bae (Kyung Hee University), Kim, Seong Tae (Technical University of Munich)
Object DetectionExplainability and InterpretabilityContrastive LearningImageTextBiomedical DataComputed Tomography
🎯 What it does: This paper studies a mask-free neuron concept annotation method for medical visual models called MAMMI, which is used to explain the internal decisions of the model.
Masked Residual Diffusion Probabilistic Model with Regional Asymmetry Prior for Generating Perfusion Maps from Multi-phase CTA
Cai, Yuxin (Huazhong University of Science and Technology), Qiu, Wu (Huazhong University of Science and Technology)
GenerationData SynthesisDiffusion modelImageBiomedical DataComputed Tomography
🎯 What it does: Based on multi-phase CTA images, CT perfusion (CBF/CBV/Tmax) images are generated using the Masked Residual Diffusion Probabilistic Model (MRDPM).
Masks and Manuscripts: Advancing Medical Pre-training with End-to-End Masking and Narrative Structuring
Gowda, Shreyank N. (University of Oxford), Clifton, David A. (University of Oxford)
ClassificationSegmentationData-Centric LearningTransformerVision Language ModelContrastive LearningImageTextMultimodalityBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: A two-step pre-training framework is proposed: first, medical reports are transformed into standardized triplets and further generate binary observations/conclusions; subsequently, self-supervised reconstruction of images is performed using Meijering-based local masks, and cross-modal alignment and conditional reconstruction are integrated within the MaskVLM structure to achieve a unified representation of medical images and text.
Material Decomposition in Photon-Counting CT: A Deep Learning Approach Driven by Detector Physics and ASIC Modeling
Yu, Xiaopeng (ShanghaiTech University), Lai, Xiaochun (ShanghaiTech University)
Convolutional Neural NetworkImageComputed TomographyPhysics Related
🎯 What it does: A deep learning material decomposition method based on the physical and ASIC modeling of photon counting CT systems is proposed, utilizing a two-stage training process with simulation and experimental calibration to achieve more accurate count predictions and material thickness estimates.
MBA-Net: SAM-driven Bidirectional Aggregation Network for Ovarian Tumor Segmentation
Gao, Yifan (University of Science and Technology of China), Gao, Xin (Chinese Academy of Sciences)
SegmentationTransformerImageMultimodalityBiomedical DataMagnetic Resonance ImagingUltrasound
🎯 What it does: A bidirectional aggregation network MBA-Net based on the Segment Anything Model (SAM) is proposed for fully automatic segmentation of ovarian tumors.
MCAD: Multi-modal Conditioned Adversarial Diffusion Model for High-Quality PET Image Reconstruction
Cui, Jiaqi (Sichuan University), Wang, Yan (East China Normal University)
RestorationGenerationDiffusion modelGenerative Adversarial NetworkImageMultimodalityBiomedical DataPositron Emission Tomography
🎯 What it does: A high-quality PET image reconstruction method based on a multi-modal conditional adversarial diffusion model is proposed, utilizing low-dose PET and clinical tabular data to generate standard-dose PET.
Med-Former: A Transformer based Architecture for Medical Image Classification
Chowdary, G. Jignesh (Stony Brook University), Yin, Zhaozheng (Stony Brook University)
ClassificationTransformerImageBiomedical Data
🎯 What it does: A Transformer-based medical image classification architecture named Med-Former is proposed, which integrates local and global feature learning with spatial attention fusion;
MEDBind: Unifying Language and Multimodal Medical Data Embeddings
Gao, Yuan (University Health Network), McIntosh, Chris (University Health Network)
ClassificationRetrievalTransformerLarge Language ModelContrastive LearningMultimodalityBiomedical DataElectronic Health RecordsElectrocardiogram
🎯 What it does: A tri-modal pre-trained model MEDBind was constructed, embedding chest X-ray (CXR), electrocardiogram (ECG), and medical text into the same vector space to achieve cross-modal retrieval, zero-shot and few-shot classification, and directly integrating this embedding into large language models for clinical prediction.
MedCLIP-SAM: Bridging Text and Image Towards Universal Medical Image Segmentation
Koleilat, Taha (Concordia University), Xiao, Yiming (Concordia University)
SegmentationRetrievalContrastive LearningImageTextBiomedical DataMagnetic Resonance ImagingUltrasound
🎯 What it does: Proposes the MedCLIP-SAM framework, which integrates BiomedCLIP and SAM for text prompt-driven universal medical image segmentation.
MedContext: Learning Contextual Cues for Efficient Volumetric Medical Segmentation
Gani, Hanan (Mohamed Bin Zayed University of Artificial Intelligence), Khan, Salman (Mohamed Bin Zayed University of Artificial Intelligence)
SegmentationKnowledge DistillationBiomedical DataMagnetic Resonance ImagingComputed Tomography
🎯 What it does: This paper proposes a general training framework MedContext, which can jointly optimize supervised segmentation tasks and self-supervised context reconstruction tasks without the need for large-scale labeled data or pre-training, thereby improving the performance of 3D medical image segmentation.