The proposed MaxStripViT is a novel hybrid architecture that effectively integrates distortion modeling with multi-axis strip attention mechanisms and Local Position-Aware blocks and provides a robust solution for health scenarios requiring both geometric flexibility and multi-scale interpretability.
Abstract
Accurate diagnosis in medical imaging is often hampered by two intrinsic factors: the complex, anomalous geometric distortion of anatomical structures and the extreme size variability of pathological lesions. Existing convolutional neural networks (CNNs) struggle with global context, while Vision Transformer (ViT) models are limited by quadratic computational costs or reliance on window-based attention mechanisms that disrupt semantic continuity. To address these limitations, we propose MaxStripViT, a novel hybrid architecture that effectively integrates distortion modeling with multi-axis strip attention mechanisms and Local Position-Aware blocks. Our method introduces three key contributions: 1) geometric-adaptive stem (GAS) leverages learnable offsets via Deformable Convolutions (DCNv2) to dynamically align the sampling grid with irregular organ boundaries at the earliest feature extraction stage. It effectively mitigates background noise; 2) position-aware local block is proposed to enhance the Mobile Inverted Bottleneck (MBConv) with Coordinate Attention. This mechanism explicitly models long-range dependencies along spatial axes, improving the precise localization of subtle lesions; and 3) novel sequential multi-axis strip attention mechanism is proposed to replace shift window-based self-attention. The proposed method is experimented on three dataset benchmarks of CT and ChestXray imaging for multi-class lung disease such as CheXtImageNet, IQ-OTH/NCCD, and ChestXray-Image dataset. The results demonstrated that MaxStripViT achieves robust performance and outperforms standard ViT models, and improves performance compared to the most advanced hybrid models currently available, including Swin Transformer, MaxViT, CSWin, ConvNeXt, ConvNeXtv2, and EfficientNet-B7, in terms of classification accuracy and computational efficiency. The proposed method provides a robust solution for health scenarios requiring both geometric flexibility and multi-scale interpretability.
Vision Transformer has achieved significant performance improvements in natural image segmentation tasks owing to its superior global modeling capabilities. However, applying vision Transformers to 3D medical image segmentation is challenging because of the quadratic computational complexity of the self-attention mechanism and their limited generalization on small-scale datasets. To address these limitations, we propose a hybrid CNN-Transformer architecture guided by channel attention, referred to as CACFormer, for 3D medical image segmentation. Specifically, we design a simple and effective channel attention module to guide the fusion of local and global features in each channel. This module adaptively assigns weights to each channel based on its semantic contribution to accurate segmentation. Meanwhile, we introduce a novel linear Transformer variant that integrates a linear attention mechanism with tanh activation. This design encourages the model to focus on the target regions and produce robust segmentation outcomes. The effectiveness and competitive generalization of the proposed framework are validated across five benchmark datasets. On AMOS2022, CACFormer achieves an average Dice score of 89.71%, outperforming 3D UX-Net (89.30%) while reducing inference time from 3.77 s to 2.49 s (a 33.95% reduction). On BraTS2021, CACFormer attains an average Dice score of 90.20%, comparable to TransBTS (90.33%), with 28.54% fewer parameters and 15.22% faster inference time (from 0.46 s to 0.39 s), demonstrating a favorable trade-off between performance and efficiency. Moreover, CACFormer demonstrates competitive cross-dataset generalization, achieving an average Dice score of 86.50% on BraTS2021 when trained on BraTS2019, significantly out-performing TransBTS (47.90%). Index Terms-3D Medical Image.
Hongzhong Tang, Xiaopei Zhang, Yabo Li et al.· IEEE Transactions on Medical...· 0 citations
TransCat, a hybrid CNN-Transformer architecture for medical image segmentation, is proposed and an extended deformable attention mechanism with attentive value identification is developed, to control the computational burden caused by the enlarged token set.
Jin Wang, Zheng-Hua Yang, Dong-Ming Zhou et al.· Frontiers in Bioinformatics· 0 citations
Colonoscopy and dermoscopy are essential clinical tools for the early detection and diagnosis of colorectal polyps and skin lesions, respectively. Accurate segmentation of polyp and skin lesion images is critical for subsequent clinical diagnosis and treatment planning. However, the significant variations in lesion morphology, size, and boundary characteristics, together with the insufficient multi-scale feature fusion in existing methods, still pose considerable challenges to accurate medical image segmentation. To address these challenges, we propose NAHFormer, a medical image segmentation framework equipped with boundary-aware and redundancy-aware capabilities. Specifically, NAHFormer employs a pyramid-structured Mix Transformer (MiT) encoder to capture multi-scale features, enhancing the model’s generalization capability across diverse lesion appearances. A Cross-Resolution Semantic Perception (CRSP) module is designed to integrate semantic information across multiple resolutions, enabling precise delineation of lesion contours and boundaries through neighborhood attention. Furthermore, a Hierarchical Feature Fusion (HFF) module progressively aggregates multi-scale features while suppressing redundant information, thereby improving segmentation accuracy. Extensive experiments on five publicly available colonoscopy polyp datasets (Kvasir, CVC-ClinicDB, CVC-ColonDB, EndoScene, and ETIS) and two dermoscopy datasets (ISIC 2017 and ISIC 2018) demonstrate that NAHFormer consistently outperforms state-of-the-art methods, achieving mean Dice scores of 0.821 on the challenging ETIS dataset and 0.909 on ISIC 2018.
Computer-aided diagnostic technologies have demonstrated substantial advantages in 3D medical image segmentation, particularly in multimodal 3D medical image segmentation tasks, where they play a pivotal role in driving continuous innovation in related architectures. As an integration of U-Net and Transformer, the UNETR architecture has demonstrated remarkable efficacy in 3D medical image segmentation. Nevertheless, despite its successes, UNETR remains challenged by clinical complexities such as intricate tumor localization and anatomical structural diversity in complex clinical settings. To address these issues, we propose an enhanced 3D segmentation framework, UAtten-Unetr, designed to improve segmentation accuracy and robustness in complex medical scenarios. The framework captures global contextual information via hierarchical Transformer layers and incorporates a spatial–channel attention module to enable adaptive fusion of multimodal features, thereby effectively enhancing cross-modal feature alignment capabilities. Concurrently, we innovatively developed a unified loss function based on bimodal modality-specific Dice constraints and uncertainty regularization, optimized for synchronous learning across the ACDC (cardiac MRI) and AMOS22 (abdominal CT/MRI) datasets. Experimental results showed that UAtten-Unetr achieved an average Dice score of 92.20% on the ACDC dataset, exceeding the reported nnU-Net result of 91.61% by 0.59 percentage points. On the AMOS22 dataset, the proposed method achieved an average Dice score of 84.51%, exceeding the reported UNETR result of 78.33% by 6.18 percentage points. However, its myocardium Dice score (84.11%) was lower than those of nnU-Net (89.24%) and MT-UNet (89.04%), indicating a remaining limitation in myocardium boundary segmentation. These results indicate competitive segmentation performance under the reported experimental settings. This method delivers dual improvements in accuracy and generalization across complex anatomical scenarios, providing an effective solution for precise diagnosis in intricate clinical environments.
Accurate three-dimensional medical image segmentation underpins computer-aided diagnosis and surgical planning, yet it is challenged by the spatial anisotropy of clinical computed tomography and magnetic resonance scans and by the topological fragility of fine anatomical structures, such as hepatic vessels, that conventional voxel-overlap supervision may not adequately represent. Conventional dense global self-attention introduces a quadratic voxel-pair interaction term. To address these issues, we propose FSAUnet, a parameter-efficient three-dimensional segmentation framework built upon the self-configuring nnU-Net pipeline and featuring three innovations. First, an adaptive anisotropy-aware focal modulation module aggregates heterogeneous multi-scale context through depth-wise convolutions, including an anisotropic in-plane branch, without constructing a voxel-wise affinity matrix. For fixed batch size, channel width, kernel size, and branch configuration, its theoretical arithmetic cost and principal inference-forward working-memory requirement under sequential branch evaluation scale linearly with the number of spatial voxels. Second, a cascaded shuffle attention module embedded in the skip connections suppresses background noise through a sequential channel-shuffle and spatial-refinement pipeline. Third, a multi-component class-aware loss augments cross-entropy and region-overlap supervision with a boundary-band term and a class-aware clDice term that is selectively activated for tubular structures to encourage topology-aware segmentation. The class-specific coefficients are fixed rather than learned or sample-adaptive, and a lower-bounded denominator is used only as a numerical safeguard when the soft-skeleton volume becomes small. Under full five-fold internal cross-validation against nnU-Net on three computed tomography datasets from the Medical Segmentation Decathlon, FSAUnet achieves mean Dice scores of 66.83, 70.16, and 53.84 percent on Task07, Task08, and Task10, respectively, corresponding to descriptive mean differences of 1.52, 1.32, and 4.50 percentage points. No paired case-level significance test was performed. Additional comparisons with representative segmentation architectures and component ablations are conducted using a controlled fixed fold-0 protocol and are interpreted descriptively. FSAUnet requires approximately 22.0 percent fewer parameters than nnU-Net. Controlled module-level profiling is consistent with the analytical spatial-scaling trends of A2FM and explicit dense global self-attention, while whole-network profiling shows that the parameter reduction does not necessarily translate into lower one-patch latency or peak allocated inference memory. Ablation results indicate dataset- and metric-dependent behavior of the evaluated configurations. Overall, FSAUnet obtains higher cross-validation mean Dice with fewer parameters under the evaluated settings, without establishing statistical significance, strict architecture-only superiority, or end-to-end runtime superiority.
