TL;DR
GTUNet is proposed, a novel hybrid framework that synergistically integrates Graph Attention Networks, Transformers, and UNet for robust 3D medical image segmentation and significantly outperforms existing methods in ASSD, highlighting its precision in delineating vascular boundaries.
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.
Medical image segmentation requires high accuracy and robustness, yet practical commercial deployment also demands privacy preservation and computational efficiency. In this context, the U-Net architecture, which can be inherently decoupled into independent encoder and decoder components, serves as a natural commercial choice. However, pure Transformer-based variants like Swin-UNet often suffer from insufficient local detail capture and limited interpretability. In this paper, we propose a lightweight hybrid architecture built upon the Swin-UNet framework. Our model integrates a parallel CNN encoder to complement the shallow layer reasoning of Swin Transformers with local texture features. To bridge the semantic gap and enhance fine-grained spatial detail recovery, we design an asymmetric feature fusion strategy and introduce cross-layer skip (XSkip) connections that explicitly propagate shallow CNN features into the decoder. We further incorporate novel loss functions and an auxiliary supervision head (Aux-Head) to strengthen training stability, boundary delineation, and intermediate feature interpretability. Extensive experiments on the Synapse multi-organ segmentation dataset demonstrate that our approach achieves state-of-the-art competitive Dice scores and Hausdorff distances, offering an accurate, efficient, and interpretable solution for clinical deployment.
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.
Automatic cardiac image segmentation is pivotal for diagnosing and treating cardiac diseases. In this work, we introduce MCSeg, a volumetric transformer-based network tailored for multi-modal cardiac segmentation. To overcome the architectural mismatch inherent in existing hybrid networks, we propose a novel Scaling Feature Pyramid (SFP). Unlike conventional skip connections, the SFP effectively bridges the single-scale 3D Vision Transformer (ViT) encoder and the multi-scale CNN decoder by transforming the ViT's output into a hierarchical feature pyramid, ensuring that global contextual information is effectively leveraged. For the training paradigm, the ViT encoder first undergoes self-supervised pre-training via masked image modeling. Subsequently, the network is fine-tuned on downstream tasks, during which a regional mutual information (RMI) loss is integrated to improve boundary segmentation accuracy. In experiments, MCSeg consistently outperforms eleven SOTA methods on CT dataset ImageCHD, multi-modal dataset MM-WHS, MRI dataset HVSMR-2.0 and MSD Heart, highlighting the effectiveness of our MCSeg for multi-modal cardiac segmentation tasks. Furthermore, MCSeg's superior performance in few-shot experiment showcases its significant potential in adapting to limited data scenarios. Codes and pre-trained ViT-B weights are open-sourced at https://openi.pcl.ac.cn/OpenMedIA/MCSeg
For medical image segmentation, accurately balancing local details and global long-range dependencies is critical to tackling thyroid nodule challenges (variable sizes, ambiguous boundaries, complex context). Traditional CNNs excel at local feature extraction but are constrained by local receptive fields, hindering efficient global dependency modeling. To address this, we propose a Parallel Mamba Dual-U Network (PM-DUNet). It adopts a cascaded dual U-Net encoder-decoder for two-stage “coarse-to-fine” segmentation refinement. We design a Multi-Path Parallel Mamba (MPM) module—using State Space Models (SSMs)—to efficiently model global context with linear complexity. Additionally, Squeeze-Excitation Downsampling (SED) and Spatial Attention Upsampling (SAU) modules are integrated to adaptively enhance key features in encoding/decoding. Results show PM-DUNet achieves highly competitive performance and outperforms state-of-the-art methods on most core metrics, verifying its effectiveness and robustness for complex medical image segmentation. Our code is available on https://github.com/Andrevict/MPDUNet.
Abstract. The pancreas and its associated tumors typically exhibit minuscule spatial proportions, high heterogeneity, and poorly contrasted with surrounding tissues, rendering automated segmentation from computed tomography scans highly challenging. To address these problems, we propose DMFU-Net, a spatial–frequency attention-guided segmentation framework built upon a U-Net-style architecture. Integrating multi-scale standard and atrous convolutions, the encoder balances global context and local details to extract rich semantic information. For further feature refinement, the bottleneck layer leverages multi-scale receptive fields to aggregate cross-scale features, significantly fortifying deep semantic representations. In the final stage, the dual-domain attention mechanism within the decoder leverages spatial and frequency domain complementarity, sustaining semantic consistency while precisely delineating minuscule structures. Experimental results on the Medical Segmentation Decathlon Task07_Pancreas dataset show that DMFU-Net achieves Dice similarity coefficient (DSC) scores of 56.79% and 82.53% for tumors and the pancreas, respectively. Notably, for ultra-small-volume tumors (<0.015%), our method achieves a 14.4% DSC gain over nnU-Net. These results demonstrate its superior reliability in challenging small-lesion cases and its potential value for downstream clinical quantitative analysis.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
