GRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound images

Abstract Ultrasound imaging is widely used for diagnosing various medical conditions. However, lesion segmentation in ultrasound images is challenging due to low contrast, noise, blurred boundaries, and variability in lesion characteristics. To address these issues, we propose a Global Response Doub...

Full description

Saved in:
Bibliographic Details
Main Authors: Xiaokai Jiang, Xuewen Ding, Jinying Ma, Chunyu Liu, Xinyi Li
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:Journal of King Saud University: Computer and Information Sciences
Subjects:
Online Access:https://doi.org/10.1007/s44443-025-00206-z
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849225856269942784
author Xiaokai Jiang
Xuewen Ding
Jinying Ma
Chunyu Liu
Xinyi Li
author_facet Xiaokai Jiang
Xuewen Ding
Jinying Ma
Chunyu Liu
Xinyi Li
author_sort Xiaokai Jiang
collection DOAJ
description Abstract Ultrasound imaging is widely used for diagnosing various medical conditions. However, lesion segmentation in ultrasound images is challenging due to low contrast, noise, blurred boundaries, and variability in lesion characteristics. To address these issues, we propose a Global Response Double U-shaped Network, a hybrid CNN-Transformer architecture designed for lesion segmentation in ultrasound images. The encoder-decoder backbone is constructed using a U-shaped Dilated Convolution module, which effectively captures fine-grained local features and enhances boundary delineation, particularly under low-contrast conditions. To improve global context modeling, this paper proposes the Global Response Transformer Block in the bottleneck, enabling the network to capture long-range dependencies and structural variability in lesion appearance. By modeling interactions across distant regions, the block more effectively captures the variability in lesion shape, size, and location, enhancing segmentation accuracy for complex and irregular structures in ultrasound images. Furthermore, we design a Multi-Scale Linear Attention Gate to refine skip connections by emphasizing salient features and suppressing redundancy, thereby mitigating noise interference and improving decoding efficiency. By suppressing speckle noise and enhancing critical features, this mechanism improves segmentation accuracy and ensures robustness in complex ultrasound imaging scenarios. The proposed method has been extensively evaluated on publicly available ultrasound image datasets, including breast and thyroid lesion data, demonstrating its effectiveness and robustness in segmenting complex and low-contrast lesions under real-world imaging conditions.
format Article
id doaj-art-9a87668857c6434e94541ba8b83f46d5
institution Kabale University
issn 1319-1578
2213-1248
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series Journal of King Saud University: Computer and Information Sciences
spelling doaj-art-9a87668857c6434e94541ba8b83f46d52025-08-24T11:53:35ZengElsevierJournal of King Saud University: Computer and Information Sciences1319-15782213-12482025-08-0137711910.1007/s44443-025-00206-zGRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound imagesXiaokai Jiang0Xuewen Ding1Jinying Ma2Chunyu Liu3Xinyi Li4School of Electronic Engineering, Tianjin University of Technology and EducationSchool of Electronic Engineering, Tianjin University of Technology and EducationSchool of Electronic Engineering, Tianjin University of Technology and EducationSchool of Electronic Engineering, Tianjin University of Technology and EducationSchool of Electronic Engineering, Tianjin University of Technology and EducationAbstract Ultrasound imaging is widely used for diagnosing various medical conditions. However, lesion segmentation in ultrasound images is challenging due to low contrast, noise, blurred boundaries, and variability in lesion characteristics. To address these issues, we propose a Global Response Double U-shaped Network, a hybrid CNN-Transformer architecture designed for lesion segmentation in ultrasound images. The encoder-decoder backbone is constructed using a U-shaped Dilated Convolution module, which effectively captures fine-grained local features and enhances boundary delineation, particularly under low-contrast conditions. To improve global context modeling, this paper proposes the Global Response Transformer Block in the bottleneck, enabling the network to capture long-range dependencies and structural variability in lesion appearance. By modeling interactions across distant regions, the block more effectively captures the variability in lesion shape, size, and location, enhancing segmentation accuracy for complex and irregular structures in ultrasound images. Furthermore, we design a Multi-Scale Linear Attention Gate to refine skip connections by emphasizing salient features and suppressing redundancy, thereby mitigating noise interference and improving decoding efficiency. By suppressing speckle noise and enhancing critical features, this mechanism improves segmentation accuracy and ensures robustness in complex ultrasound imaging scenarios. The proposed method has been extensively evaluated on publicly available ultrasound image datasets, including breast and thyroid lesion data, demonstrating its effectiveness and robustness in segmenting complex and low-contrast lesions under real-world imaging conditions.https://doi.org/10.1007/s44443-025-00206-zUltrasound imagesLesion segmentationU-NetTransformerResponse normalizationAttention mechanism
spellingShingle Xiaokai Jiang
Xuewen Ding
Jinying Ma
Chunyu Liu
Xinyi Li
GRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound images
Journal of King Saud University: Computer and Information Sciences
Ultrasound images
Lesion segmentation
U-Net
Transformer
Response normalization
Attention mechanism
title GRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound images
title_full GRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound images
title_fullStr GRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound images
title_full_unstemmed GRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound images
title_short GRU2-Net: Global response double U-shaped network for lesion segmentation in ultrasound images
title_sort gru2 net global response double u shaped network for lesion segmentation in ultrasound images
topic Ultrasound images
Lesion segmentation
U-Net
Transformer
Response normalization
Attention mechanism
url https://doi.org/10.1007/s44443-025-00206-z
work_keys_str_mv AT xiaokaijiang gru2netglobalresponsedoubleushapednetworkforlesionsegmentationinultrasoundimages
AT xuewending gru2netglobalresponsedoubleushapednetworkforlesionsegmentationinultrasoundimages
AT jinyingma gru2netglobalresponsedoubleushapednetworkforlesionsegmentationinultrasoundimages
AT chunyuliu gru2netglobalresponsedoubleushapednetworkforlesionsegmentationinultrasoundimages
AT xinyili gru2netglobalresponsedoubleushapednetworkforlesionsegmentationinultrasoundimages