DS-UNeXt: depthwise separable convolution network with large convolutional kernel for medical image segmentation

Huang, Tongyuan; Chen, Jiangxia; Jiang, Linfeng

doi:10.1007/s11760-022-02388-9

Cited by 18 publications

(3 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Depthwise separable convolution networks are also applicable to image segmentation. Huang et al [35] proposes an end-toend depthwise separable U-shaped convolution network for medical image segmentation.…”

Section: A Semantic Segmentation For Remote Sensing Imagerymentioning

confidence: 99%

Multilateral Semantic With Dual Relation Network for Remote Sensing Images Segmentation

Zhao,

Cao,

Dong

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Semantic segmentation of remote sensing images is an extensively employed and demanding task. Although deep convolutional neural networks have significantly increased the accuracy of semantic segmentation, the problems of losing detailed features in segmentation and ignoring rich contextual information of images still exist. To solve these challenges, we propose a multilateral semantic with dual relation network (MSDRNet) for remote sensing images segmentation. The proposed MSDRNet consists of two parallel modules, the detail semantic module (DSM) and the global semantic module (GSM), for extracting image detail and global features, respectively. Subsequently, improved spatial relation block (SRB) and channel relation block (CRB) are introduced in two separate parallel modules to further enhance the contextual connection of the images. Finally, a feature refinement module (FRM) is added to balance the multilateral features between the features extracted from the two branches. We display the robustness and effectiveness of the proposed MSDRNet on the publicly available ISPRS Potsdam and Vaihingen datasets. We further experimented with the Gaofen Image Dataset (GID), which contains information on larger scale features, to demonstrate the validity of our model. The results of extensive experiments conducted on the above three datasets show that the proposed approach outperforms several state-of-the-art semantic segmentation methods.

show abstract

Section: A Semantic Segmentation For Remote Sensing Imagerymentioning

confidence: 99%

Multilateral Semantic With Dual Relation Network for Remote Sensing Images Segmentation

Zhao,

Cao,

Dong

2024

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

show abstract

“…In the medical field, accurate image segmentation is crucial for disease diagnosis and treatment planning. However, this field faces serious challenges, mainly focused on limited annotated data [1] , complex image structures [2] , and the diversity of disease manifestations [3] . Traditional supervised learning methods have achieved some success in medical image segmentation, but they tend to rely on large amounts of labeled data, which is an expensive and time-consuming task in the medical field [4] .…”

Section: Introductionmentioning

confidence: 99%

FR-nnUNet: A Novel Medical Segmentation Network Based on the Fuzzy Regions Recognition Scheme and Improved nnU-Net

Huang,

Wang,

Han

et al. 2024

Preprint

View full text Add to dashboard Cite

The performance of the traditional supervised learning based medical image segmentation methods heavily depends on the expensive and scarce data labeling. The semi-supervised based segmentation strategies, such as the mean teacher model, have achieved milestone success in this domain. However, few studies have focused on the problem of target selection. Therefore, a novel medical image segmentation model is proposed based on the fuzzy regions recognition scheme and improved nnU-Net. The novel fuzzy regions recognition strategy is introduced and the improved nnU-Net with the new semi-supervised mean teacher network architecture is constructed, which can efficiently optimize the representation of image features via the intensity and spatial transformations. The class-conditional labeling noise identification method is applied to the image segmentation task and combines the concepts of confusion matrix and thresholding to improve the robustness of the model to class imbalance and overconfidence predictions. Experimental results show that the proposed fuzzy selective mean teacher model can achieve excellent segmentation performance on the Left Atrium and ACDC datasets. Compared with the traditional methods, it can not only improve the objective segmentation evaluation metrics in terms of the Dice coefficient, Jaccard coefficient, and Hausdorff95 distance but also obtain better segmentation results especially when less labeled data is used.

show abstract

“…In the early stages, it was challenging and time-consuming for radiologists to manually analyze ultrasound scans and distinguish abnormal from normal breast tissue, which delayed the diagnosis process. Automatic segmentation of tumor regions is an important task in computer-aided diagnosis systems for the early detection of cancer signs [2][3][4], which contributes to reducing cancer mortality [5].…”

Section: Introductionmentioning

confidence: 99%

A Dual-branch Framework Based on Implicit Continuous Representation for Tumor Image Segmentation

Wang,

Zheng,

Wang

et al. 2023

Preprint

View full text Add to dashboard Cite

Breast tumor segmentation has important significance for early detection and determination of treatment plans. However, segmenting early-stage small tumors in breast images is challenging due to small and low-resolution tumor regions, variation of tumor shapes, and blurred tumor boundaries. More importantly, breast scans are usually noisy and include metal artifacts. Most of the existing tumor segmentation methods have difficulty in extracting lesion discriminative information, leading to the problem that small tumors are ignored or predictions contain a lot of noise. In addition, common reconstruction and segmentation algorithms are based on discrete images and ignore the continuity of feature space. Therefore, in this paper, we investigate a novel and flexible dual-branch framework, named High-Resolution and Information Bottleneck-based Segmentation Network (HR-IBS), for breast tumor segmentation. For the first time, this method introduces the high-resolution tumor region reconstruction (HR-TR) branch via implicit neural representations to learning functions that map the discrete input signal to continuous density. The branch enables reconstruction from lesion regions for another segmentation branch. Furthermore, we design an Information bottleneck-based segmentation (IBS) branch, which adopts information bottleneck and U-Net to retain the features most relevant while removing noisy regions, and discovering more informative lesion regions. The reconstruction and segmentation branches interact with each other to facilitate performance. Comprehensive experiments are conducted on segmentation benchmarks of two modalities of breast images. The results show that the proposed method outperforms other segmentation models and contributes to optimizing hand-crafted ground-truths.

show abstract

DS-UNeXt: depthwise separable convolution network with large convolutional kernel for medical image segmentation

Cited by 18 publications

References 35 publications

Multilateral Semantic With Dual Relation Network for Remote Sensing Images Segmentation

Multilateral Semantic With Dual Relation Network for Remote Sensing Images Segmentation

FR-nnUNet: A Novel Medical Segmentation Network Based on the Fuzzy Regions Recognition Scheme and Improved nnU-Net

A Dual-branch Framework Based on Implicit Continuous Representation for Tumor Image Segmentation

Contact Info

Product

Resources

About