IEEE transactions on medical imaging最新文献_第2页

Artifact-Aware Self-Supervised Temporal Learning for Pediatric Lung Ultrasound Scoring 儿童肺超声评分的伪象感知自监督时态学习。

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-21 DOI: 10.1109/TMI.2026.3715740

Ling Zhou;Runtian Yuan;Shang Gao;Yinghua Sun;Xiaojiao Qin;Xiaoling Li;Yuejie Zhang;Xiaobo Zhang;Rui Feng

{"title":"Artifact-Aware Self-Supervised Temporal Learning for Pediatric Lung Ultrasound Scoring","authors":"Ling Zhou;Runtian Yuan;Shang Gao;Yinghua Sun;Xiaojiao Qin;Xiaoling Li;Yuejie Zhang;Xiaobo Zhang;Rui Feng","doi":"10.1109/TMI.2026.3715740","DOIUrl":"10.1109/TMI.2026.3715740","url":null,"abstract":"Accurately assessing the severity of pediatric pneumonia using AI-based lung ultrasound (LUS) video scoring remains challenging due to three main limitations: 1) limited availability of expert annotations, 2) underutilization of temporal dynamics, and 3) lack of mechanisms to capture the directional patterns and temporal evolution of diagnostic artifacts. To address these challenges, we introduce PedLUS, a self-supervised video learning framework for Pediatric Lung Ultrasound Scoring, together with a dedicated dataset comprising 1,646 unlabeled clips for pretraining and 464/156 labeled clips for training/test severity scoring. PedLUS mitigates annotation scarcity by using a self-supervised pretraining strategy that masks the central 8 frames of a 24-frame sequence and reconstructs them from bidirectional temporal context, thereby learning motion-aware representations. To improve reconstruction fidelity, we design a Semantic-Aware Clustering Reconstruction (SACR) module that groups latent features into learnable prototypes and uses them to regenerate masked regions, reducing redundancy while emphasizing clinically meaningful artifact cues. To capture the directional and evolving characteristics of diagnostic artifacts, PedLUS further integrates two specialized modules: Spatial Directional Attention (SDA) for orientation-specific feature refinement, and Temporal-Aware Attention (TAA) for modeling progression across respiratory phases. Experimental results over ten repeated runs demonstrate that the proposed model achieves outstanding performance, with an accuracy, F1-score and AUC of <inline-formula> <tex-math>$82.37~pm ~1.36$ </tex-math></inline-formula>%, <inline-formula> <tex-math>$81.56~pm ~1.47$ </tex-math></inline-formula>%, and <inline-formula> <tex-math>$92.03~pm ~0.94$ </tex-math></inline-formula>%, respectively. Code and dataset will be released at: PedLUS.","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"5097-5111"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148539159","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

WDK-Net: Lightweight Wavelet Diffusion With Kolmogorov--Arnold Network for Limited-Angle Cardiac CT Reconstruction WDK-Net：基于Kolmogorov-Arnold网络的轻型小波扩散心脏CT有限角度重建

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-09 DOI: 10.1109/TMI.2026.3711942

Changsheng Fang;Bahareh Morovati;Shuo Han;Yu Shi;Li Zhou;Shuyi Fan;Dayang Wang;Hengyong Yu

{"title":"WDK-Net: Lightweight Wavelet Diffusion With Kolmogorov--Arnold Network for Limited-Angle Cardiac CT Reconstruction","authors":"Changsheng Fang;Bahareh Morovati;Shuo Han;Yu Shi;Li Zhou;Shuyi Fan;Dayang Wang;Hengyong Yu","doi":"10.1109/TMI.2026.3711942","DOIUrl":"10.1109/TMI.2026.3711942","url":null,"abstract":"Limited-angle cardiac CT reconstruction is a severely ill-posed problem, where incomplete angular coverage leads to strong artifacts and structural distortions. Although diffusion-based methods have shown strong potential for improving reconstruction quality, their high computational cost, large memory demand, and slow inference remain major barriers to practical clinical deployment. To address this bottleneck, we propose WDK-Net, a lightweight dual-domain reconstruction framework with a structure–detail decoupled design, aiming to make diffusion-based reconstruction more computationally feasible for cardiac CT. Leveraging wavelet transforms, WDK-Net models global structures and local details in different domains. Specifically, although shallow low-frequency components preserve more complete anatomical structures, deeper decomposition provides a more compact low-frequency representation for efficient generative modeling. Experiments on simulated, in-house clinical, and public cardiac CT datasets demonstrate that the proposed method achieves competitive or superior reconstruction quality across multiple limited-angle settings (60°, 90°, 120°), while substantially reducing the computational burden compared with existing diffusion-based reconstruction methods. These results suggest that frequency-selective low-dimensional wavelet diffusion, combined with explicit structure–detail decoupling, provides a feasible pathway toward more efficient and clinically deployable diffusion-based limited-angle cardiac CT reconstruction.","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"4923-4936"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148414748","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Parallel Hyper-Prior and Dual-Domain Autoregressive Transformer for Lossless JPEG Recompression of Pathology Images 用于病理图像无损JPEG再压缩的并行超先验双域自回归变压器。

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-22 DOI: 10.1109/TMI.2026.3715893

Tao Song;Yicheng Wu;Zhonghua Chen;Linda Wei;Yi Guo;Feng Xu;Shaoting Zhang

{"title":"Parallel Hyper-Prior and Dual-Domain Autoregressive Transformer for Lossless JPEG Recompression of Pathology Images","authors":"Tao Song;Yicheng Wu;Zhonghua Chen;Linda Wei;Yi Guo;Feng Xu;Shaoting Zhang","doi":"10.1109/TMI.2026.3715893","DOIUrl":"10.1109/TMI.2026.3715893","url":null,"abstract":"Due to their high resolution, pathology images incur substantial storage costs. Although most vendor formats have already adopted lossy compression (typically JPEG) to process such images, the storage requirements remain considerable. However, achieving further reductions in storage requirements while ensuring truly lossless recompression of already lossy-compressed JPEG images remains a critical yet largely unexplored research problem. To address this challenge, we propose a Parallel Hyper-Prior and Dual-Domain Autoregressive Transformer model that processes the three discrete cosine transform (DCT) components in parallel, thereby reducing the complexity of the attention mechanism. The Dual-Domain Autoregressive Transformer module employs a spatial fractal autoregressive transformer, which not only substantially lowers the computational complexity of attention but also captures the inherent hierarchical structure of the spatial domain, and a frequency-domain autoregressive transformer that more effectively models dependencies among coefficients at different frequencies. In addition, we introduce a parameter-sharing and scale-regularized Gaussian–Laplace mixture entropy model to achieve more accurate entropy estimation, thereby enhancing the network’s conditional modeling capability and improving compression performance. We validate the proposed method on the public CAMELYON16 dataset (in-distribution) and multiple TCGA datasets (out-of-distribution). Experimental results show that, for lossless JPEG recompression, our approach achieves state-of-the-art performance compared with both traditional and learned JPEG lossless recompression methods, attaining up to 33.4% additional lossless compression on standard JPEG files. Moreover, the method maintains strong generalization to out-of-distribution data. The source code has been made publicly available at: <uri>https://github.com/zhuiguangzhe123/PLR</uri>","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"5126-5137"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148552394","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

MoST-Net: Motion Memory Spatiotemporal Transformer Network for Myocardial Echocardiography Segmentation 心肌超声心动图分割的运动记忆时空变换网络。

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-22 DOI: 10.1109/TMI.2026.3715753

Pu Huang;Peisheng Wang;Mei Zhu;Xiangyu Zhai;Jie Xue;Yao Cheng;Pan Li;Dengwang Li

{"title":"MoST-Net: Motion Memory Spatiotemporal Transformer Network for Myocardial Echocardiography Segmentation","authors":"Pu Huang;Peisheng Wang;Mei Zhu;Xiangyu Zhai;Jie Xue;Yao Cheng;Pan Li;Dengwang Li","doi":"10.1109/TMI.2026.3715753","DOIUrl":"10.1109/TMI.2026.3715753","url":null,"abstract":"Myocardial echocardiography segmentation is useful for cardiac function assessment, yet remains challenging due to continuous myocardial motion and deformation during the cardiac cycle, as well as boundary ambiguity caused by speckle noise. Existing memory-based segmentation methods typically construct memory banks by accumulating single-frame representations, where the stored memory items fail to explicitly capture motion dynamics across consecutive frames. Meanwhile, spatiotemporal Transformer-based methods often rely on full-sequence, limiting their applicability to online scenarios where frames arrive sequentially. In this work, we propose a Motion memory Spatiotemporal Transformer Network (MoST-Net) for myocardial echocardiography segmentation, MoST-Net integrates motion modulation into long-term memory learning, enabling the long-term memory to be updated as frames arrive while continuously accumulating temporal dependencies. Specifically, motion memory learning leverages progressively estimated motion states to modulate image features and capture short-term motion memory, which are further extended to long-range dependencies through long-term memory modeling. The memory prompt encoder maps both image and memory features into motion prompts, which are adapted by a spatiotemporal decoder to obtain the final segmentation, and the boundary uncertainty enhancement module is introduced within the decoder to alleviate ambiguous myocardial boundaries. In addition, the adaptive ranking strategy maintains an effective memory length by jointly considering temporal consistency and segmentation quality. Experiments on CAMUS, HMC-QU, and a private myocardial dataset demonstrate that MoST-Net outperforms state-of-the-art methods even under severe noise. Further analysis shows that longer memory banks may not be necessary when motion-aware memory learning is employed, demonstrating its potential for online myocardial segmentation in clinical scenarios.","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"5112-5125"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148552395","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

VSS-SAM++: Visual State Space-Aware SAM for 3-D Medical Image Segmentation 用于三维医学图像分割的视觉状态空间感知SAM。

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-23 DOI: 10.1109/TMI.2026.3716353

Jinxuan Lyu;Bin Zhang;Yipan Wang;Shengping Liu;Wang Li;Zhuoxu Cui;Haifeng Wang;Dong Liang;Yue Hu;Yihang Zhou

{"title":"VSS-SAM++: Visual State Space-Aware SAM for 3-D Medical Image Segmentation","authors":"Jinxuan Lyu;Bin Zhang;Yipan Wang;Shengping Liu;Wang Li;Zhuoxu Cui;Haifeng Wang;Dong Liang;Yue Hu;Yihang Zhou","doi":"10.1109/TMI.2026.3716353","DOIUrl":"10.1109/TMI.2026.3716353","url":null,"abstract":"The Segment Anything Model (SAM) has demonstrated groundbreaking performance in natural image segmentation, yet its direct application to medical imaging remains suboptimal due to domain shifts in data distributions and the inherent 3D nature of medical data. Although recent SAM-based methods have employed parameter-efficient transfer learning (PETL) to adapt SAM for medical tasks, they often overlook the critical 3D contextual information essential for accurate volumetric segmentation. To address this limitation, we introduce VSS-SAM++, a novel dual-branch architecture that combines SAM’s foundational visual priors with Vision Mamba’s capacity for modeling long-range spatial dependencies. In this framework, SAM serves as the primary encoder for high-level feature extraction, while a parallel Mamba branch captures cross-slice dependencies in 3D medical volumes. A gated hybrid attention module then dynamically fuses complementary features from both branches, adaptively weighting multi-view representations to minimize feature ambiguity and enhance segmentation precision. Extensive evaluations across nine public CT and MRI datasets demonstrate that VSS-SAM++ outperforms existing methods by <bold>0.2–11.3%</b> in Dice score on diverse organ segmentation tasks. The framework’s robustness to domain shifts and scalability across diverse modalities highlights its potential for clinical deployment.","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"5166-5177"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148564954","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

REACT: Relation-Aware Class Activation-Guided Mixture-of-Experts for Spatio-Temporal Triplet Recognition REACT：关系感知类激活引导的时空三元识别混合专家。

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-22 DOI: 10.1109/TMI.2026.3716009

Bohan Qu;Yuhao Xiao;Wenxuan Liu;Wenbo Zheng;Zheng Wang;Xian Zhong;Bin Sheng

{"title":"REACT: Relation-Aware Class Activation-Guided Mixture-of-Experts for Spatio-Temporal Triplet Recognition","authors":"Bohan Qu;Yuhao Xiao;Wenxuan Liu;Wenbo Zheng;Zheng Wang;Xian Zhong;Bin Sheng","doi":"10.1109/TMI.2026.3716009","DOIUrl":"10.1109/TMI.2026.3716009","url":null,"abstract":"Video-based surgical action-triplet recognition represents each surgical action as an <inline-formula> <tex-math>$langle $ </tex-math></inline-formula>instrument, verb, target<inline-formula> <tex-math>$rangle $ </tex-math></inline-formula> triplet, which is fundamental for understanding complex surgical scenes. However, this task remains challenging due to heterogeneous spatio-temporal characteristics and complex inter-component dependencies. To address these issues, we propose a Relation-Aware Class Activation-guided Mixture-of-Experts framework (REACT) for surgical triplet recognition. REACT jointly models spatial details and temporal dynamics while capturing inter-component relations via a Class Activation-guided Mixture-of-Experts (CAMoE) architecture. Class Activation Maps (CAMs) serve as weak semantic priors to guide expert routing, while Feature Modulation Experts (FMEs) adaptively align heterogeneous component features into a shared semantic space. Extensive experiments on CholecT45 and CholecTriplet show that REACT achieves state-of-the-art triplet-level performance with strong relational consistency. Additional evaluation on SIRNet dataset further demonstrates cross-procedure generalization. Code is available at <uri>https://github.com/accusemaker/REACT</uri>","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"5138-5152"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148552412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Topology Optimization of Magnetocardiographic Array Based on Cardiac Electromagnetic Simulation Model 基于心脏电磁仿真模型的心磁成像阵列拓扑优化

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-13 DOI: 10.1109/TMI.2026.3712004

Haitao Niu;Ziyuan Huang;Maotong Pang;Yuan Yuan;Yifan Li;Jingyi Zhang

{"title":"Topology Optimization of Magnetocardiographic Array Based on Cardiac Electromagnetic Simulation Model","authors":"Haitao Niu;Ziyuan Huang;Maotong Pang;Yuan Yuan;Yifan Li;Jingyi Zhang","doi":"10.1109/TMI.2026.3712004","DOIUrl":"10.1109/TMI.2026.3712004","url":null,"abstract":"Array topology in magnetocardiography system dictates spatial resolution and clinical applicability, impacting topographic reconstruction, which affects the accuracy of feature and source localization. Existing array designs rely on task-driven or empirical criteria without accounting for intrinsic imaging properties to ensure complete high-precision reconstruction and broad applicability, and analyses regarding the impact of array parameters on imaging remain lacking. To address these limitations, a computationally efficient electromagnetic simulation framework is proposed to guide array topological optimization. A three-dimensional electrophysiological forward model, combining the monodomain equation with phenomenological formulation, is constructed and regional parameters are optimized to rapidly simulate transmembrane potential. Following rigorous multidimensional validation of the mapped magnetic distribution against real data, high-fidelity signals are utilized to demonstrate the necessity of parameter optimization for imaging and to establish universally applicable criteria for the effective coverage to completely capture signals. The Manifold Harmonic Transform and the sampling theorem are employed to analyze dynamic anti-aliasing sampling requirements across cardiac cycle. Ultimately, integrating the spatial coverage and sampling criteria with a highly efficient hexagonal layout yields a broadly applicable and cost-effective topology. The reliability of array is verified and results demonstrate that the optimized configuration can achieve complete and robust reconstruction of spatiotemporally evolving signals, thereby establishing a quantitative foundation and reliable paradigm for the clinical deployment and application of system. Meanwhile, the analysis process of array optimization can serve as a reference for other application scenarios.","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"4937-4953"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148432030","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physics-Driven Autoregressive State Space Models for Medical Image Reconstruction 医学图像重建的物理驱动自回归状态空间模型。

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-23 DOI: 10.1109/TMI.2026.3716153

Bilal Kabas;Fuat Arslan;Valiyeh A. Nezhad;Kader K. Oguz;Saban Ozturk;Emine U. Saritas;Tolga Çukur

{"title":"Physics-Driven Autoregressive State Space Models for Medical Image Reconstruction","authors":"Bilal Kabas;Fuat Arslan;Valiyeh A. Nezhad;Kader K. Oguz;Saban Ozturk;Emine U. Saritas;Tolga Çukur","doi":"10.1109/TMI.2026.3716153","DOIUrl":"10.1109/TMI.2026.3716153","url":null,"abstract":"Medical image reconstruction from undersampled acquisitions is an ill-posed inverse problem requiring accurate recovery of anatomical structures from incomplete measurements. Physics-driven (PD) network models have gained prominence for this task by integrating data-consistency mechanisms with learned priors, enabling improved performance over purely data-driven approaches. However, reconstruction quality still hinges on the network’s ability to disentangle artifacts from true anatomical signals—both of which exhibit complex, multi-scale contextual structure. Convolutional neural networks (CNNs) capture local correlations but often struggle with non-local dependencies. While transformers aim to alleviate this limitation, practical implementations involve design compromises to reduce computational cost by balancing local and non-local sensitivity, occasionally resulting in performance comparable to CNNs. To address these challenges, we propose MambaRoll, a novel physics-driven autoregressive state space model (SSM) for high-fidelity and efficient image reconstruction. MambaRoll employs an unrolled architecture where each cascade autoregressively predicts finer-scale feature maps conditioned on coarser-scale representations, enabling consistent multi-scale context propagation. Each stage is built on a hierarchy of scale-specific PD-SSM modules that capture spatial dependencies while enforcing data consistency through residual correction. To further improve scale-aware learning, we introduce a Deep Multi-Scale Decoding (DMSD) loss, which provides supervision at intermediate spatial scales in alignment with the autoregressive design. Demonstrations on accelerated MRI and sparse-view CT reconstructions show that MambaRoll provides improved performance relative to state-of-the-art CNN-, transformer-, and SSM-based methods.","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"5153-5165"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11621983","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148581897","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

TriCD-Net: Triple-Attention Coordinated Cross-Layer Dynamic Network for Few-Shot Medical Image Segmentation TriCD-Net：用于医学图像分割的三注意协调跨层动态网络

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-13 DOI: 10.1109/TMI.2026.3712966

Qiang Chi;Fuzhi Wu;Yuhang Liu;Yue Zhang;Zidong Liu;Huazhong Shu

{"title":"TriCD-Net: Triple-Attention Coordinated Cross-Layer Dynamic Network for Few-Shot Medical Image Segmentation","authors":"Qiang Chi;Fuzhi Wu;Yuhang Liu;Yue Zhang;Zidong Liu;Huazhong Shu","doi":"10.1109/TMI.2026.3712966","DOIUrl":"10.1109/TMI.2026.3712966","url":null,"abstract":"Few-shot Medical Image Segmentation (FSMIS) has garnered increasing attention for its ability to reduce reliance on large-scale pixel-wise annotations. However, most existing methods rely solely on single-level encoder outputs, neglecting the complementary roles of hierarchical features in contour refinement and regional discrimination. This limitation often results in boundary ambiguity and local segmentation errors. To this end, we propose a Triple-Attention Coordinated Cross-layer Dynamic Network (TriCD-Net) for FSMIS. Specifically, we design a Feature-Enhanced Boundary Refinement (FEBR) module in each encoder layer, where three parallel attention branches explicitly model foreground, background, and boundary regions. The resulting features are passed to a Dense Dual-Contrast Segmentation (DDCS) head to perform segmentation, enabling progressive boundary refinement. Considering that query images inherently contain valuable structural information, we introduce a mask reconstruction task and design a Cross-layer Dynamic Convolution Fusion (CDCF) module that dynamically generates convolution kernels to adaptively fuse multi-level features for query reconstruction. Furthermore, we establish an interaction between the reconstruction and segmentation branches through an Uncertainty-Gated Cross-Attention Injection (UCI) module, which selectively injects reconstruction features into the segmentation branch to enhance structural coherence and boundary accuracy. Extensive experiments on three public benchmark datasets demonstrate that TriCD-Net consistently achieves state-of-the-art performance. The code is available at <uri>https://github.com/qchi-code/TriCD-Net</uri>","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"4970-4984"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148432023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dual-Domain Cross-Prompt Learning for Efficient Sparse-View CT 高效稀疏视图CT的双域交叉提示学习。

IEEE transactions on medical imaging Pub Date : 2026-10-01 Epub Date: 2026-07-16 DOI: 10.1109/TMI.2026.3713777

Wenchao Du;Qiao Mu;Huanhuan Cui;Hu Chen;Yi Zhang;Hongyu Yang

{"title":"Dual-Domain Cross-Prompt Learning for Efficient Sparse-View CT","authors":"Wenchao Du;Qiao Mu;Huanhuan Cui;Hu Chen;Yi Zhang;Hongyu Yang","doi":"10.1109/TMI.2026.3713777","DOIUrl":"10.1109/TMI.2026.3713777","url":null,"abstract":"Sparse-view computed tomography (CT) effectively reduces radiation exposure, yet it degrades image signal-to-noise ratio (SNR) and compromises the reliability of clinical diagnosis. Deep unrolling networks, which integrate the merits of optimization-based and data-driven paradigms, have achieved promising performance for sparse-view CT reconstruction. However, existing learned data consistency (DC) and prompt-based reconstruction methods only capture simplistic image priors and rely on elaborately designed regularizers as well as excessive unrolled iterations, leading to heavy computational overhead and over-smoothed results. In this work, we integrate prompt learning into unrolled gradient descent networks and propose a Dual-domain Cross-Prompt Learning (DCPL) framework to address these limitations. Specifically, we first design an implicit pixel-wise learnable step size to adapt to the spatial gradient heterogeneity of CT images. We then integrate learnable prompts separately into the data-fidelity and regularization terms during unrolled iterations, enabling the model to adaptively capture intrinsic CT anatomical and noise priors. Furthermore, a cross-prompt guiding mechanism is developed to enable inter-domain prompt interaction, which facilitates efficient prompt generation and enhances the convergence stability of the model. Extensive experiments on multiple clinical benchmarks under both in-domain and cross-domain settings demonstrate that our DCPL achieves consistent improvements in artifact suppression and fine structural preservation, even under extremely sparse-view sampling. Notably, our method delivers robust reconstruction quality with significantly fewer parameters, higher inference efficiency, and with only a few unrolled iterations.","PeriodicalId":94033,"journal":{"name":"IEEE transactions on medical imaging","volume":"45 10","pages":"5028-5040"},"PeriodicalIF":0.0,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148462312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0