SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion Segmentation

Segmentation of skin lesions in dermoscopic images is critical for the accurate diagnosis of skin cancers, particularly malignant melanoma, yet it is hindered by irregular lesion shapes, blurred boundaries, low contrast, and artifacts, such as hair interference. Conventional deep learning methods, t...

Full description

Saved in:
Bibliographic Details
Main Authors: Haijiao Yun, Qingyu Du, Ziqing Han, Mingjing Li, Le Yang, Xinyang Liu, Chao Wang, Weitian Ma
Format: Article
Language:English
Published: MDPI AG 2025-07-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/15/4652
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849239622024953856
author Haijiao Yun
Qingyu Du
Ziqing Han
Mingjing Li
Le Yang
Xinyang Liu
Chao Wang
Weitian Ma
author_facet Haijiao Yun
Qingyu Du
Ziqing Han
Mingjing Li
Le Yang
Xinyang Liu
Chao Wang
Weitian Ma
author_sort Haijiao Yun
collection DOAJ
description Segmentation of skin lesions in dermoscopic images is critical for the accurate diagnosis of skin cancers, particularly malignant melanoma, yet it is hindered by irregular lesion shapes, blurred boundaries, low contrast, and artifacts, such as hair interference. Conventional deep learning methods, typically based on UNet or Transformer architectures, often face limitations in regard to fully exploiting lesion features and incur high computational costs, compromising precise lesion delineation. To overcome these challenges, we propose SGNet, a structure-guided network, integrating a hybrid CNN–Mamba framework for robust skin lesion segmentation. The SGNet employs the Visual Mamba (VMamba) encoder to efficiently extract multi-scale features, followed by the Dual-Domain Boundary Enhancer (DDBE), which refines boundary representations and suppresses noise through spatial and frequency-domain processing. The Semantic-Texture Fusion Unit (STFU) adaptively integrates low-level texture with high-level semantic features, while the Structure-Aware Guidance Module (SAGM) generates coarse segmentation maps to provide global structural guidance. The Guided Multi-Scale Refiner (GMSR) further optimizes boundary details through a multi-scale semantic attention mechanism. Comprehensive experiments based on the ISIC2017, ISIC2018, and PH2 datasets demonstrate SGNet’s superior performance, with average improvements of 3.30% in terms of the mean Intersection over Union (mIoU) value and 1.77% in regard to the Dice Similarity Coefficient (DSC) compared to state-of-the-art methods. Ablation studies confirm the effectiveness of each component, highlighting SGNet’s exceptional accuracy and robust generalization for computer-aided dermatological diagnosis.
format Article
id doaj-art-e697dc01879d4d0e9d435e0eaea71859
institution Kabale University
issn 1424-8220
language English
publishDate 2025-07-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-e697dc01879d4d0e9d435e0eaea718592025-08-20T04:00:54ZengMDPI AGSensors1424-82202025-07-012515465210.3390/s25154652SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion SegmentationHaijiao Yun0Qingyu Du1Ziqing Han2Mingjing Li3Le Yang4Xinyang Liu5Chao Wang6Weitian Ma7School of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSchool of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSchool of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSchool of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSchool of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSchool of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSchool of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSchool of Electronic Information Engineering, Changchun University, Changchun 130022, ChinaSegmentation of skin lesions in dermoscopic images is critical for the accurate diagnosis of skin cancers, particularly malignant melanoma, yet it is hindered by irregular lesion shapes, blurred boundaries, low contrast, and artifacts, such as hair interference. Conventional deep learning methods, typically based on UNet or Transformer architectures, often face limitations in regard to fully exploiting lesion features and incur high computational costs, compromising precise lesion delineation. To overcome these challenges, we propose SGNet, a structure-guided network, integrating a hybrid CNN–Mamba framework for robust skin lesion segmentation. The SGNet employs the Visual Mamba (VMamba) encoder to efficiently extract multi-scale features, followed by the Dual-Domain Boundary Enhancer (DDBE), which refines boundary representations and suppresses noise through spatial and frequency-domain processing. The Semantic-Texture Fusion Unit (STFU) adaptively integrates low-level texture with high-level semantic features, while the Structure-Aware Guidance Module (SAGM) generates coarse segmentation maps to provide global structural guidance. The Guided Multi-Scale Refiner (GMSR) further optimizes boundary details through a multi-scale semantic attention mechanism. Comprehensive experiments based on the ISIC2017, ISIC2018, and PH2 datasets demonstrate SGNet’s superior performance, with average improvements of 3.30% in terms of the mean Intersection over Union (mIoU) value and 1.77% in regard to the Dice Similarity Coefficient (DSC) compared to state-of-the-art methods. Ablation studies confirm the effectiveness of each component, highlighting SGNet’s exceptional accuracy and robust generalization for computer-aided dermatological diagnosis.https://www.mdpi.com/1424-8220/25/15/4652computer-aided diagnosisstructure guidancefeature fusionVisual Mamba
spellingShingle Haijiao Yun
Qingyu Du
Ziqing Han
Mingjing Li
Le Yang
Xinyang Liu
Chao Wang
Weitian Ma
SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion Segmentation
Sensors
computer-aided diagnosis
structure guidance
feature fusion
Visual Mamba
title SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion Segmentation
title_full SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion Segmentation
title_fullStr SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion Segmentation
title_full_unstemmed SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion Segmentation
title_short SGNet: A Structure-Guided Network with Dual-Domain Boundary Enhancement and Semantic Fusion for Skin Lesion Segmentation
title_sort sgnet a structure guided network with dual domain boundary enhancement and semantic fusion for skin lesion segmentation
topic computer-aided diagnosis
structure guidance
feature fusion
Visual Mamba
url https://www.mdpi.com/1424-8220/25/15/4652
work_keys_str_mv AT haijiaoyun sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation
AT qingyudu sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation
AT ziqinghan sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation
AT mingjingli sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation
AT leyang sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation
AT xinyangliu sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation
AT chaowang sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation
AT weitianma sgnetastructureguidednetworkwithdualdomainboundaryenhancementandsemanticfusionforskinlesionsegmentation