A novel framework for segmentation of small targets in medical images

Abstract Medical image segmentation represents a pivotal and intricate procedure in the domain of medical image processing and analysis. With the progression of artificial intelligence in recent years, the utilization of deep learning techniques for medical image segmentation has witnessed escalatin...

Full description

Saved in:
Bibliographic Details
Main Authors: Longxuan Zhao, Tao Wang, Yuanbin Chen, Xinlin Zhang, Hui Tang, Fuxin Lin, Chunwang Li, Qixuan Li, Tao Tan, Dezhi Kang, Tong Tong
Format: Article
Language:English
Published: Nature Portfolio 2025-03-01
Series:Scientific Reports
Online Access:https://doi.org/10.1038/s41598-025-94437-9
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850094866908839936
author Longxuan Zhao
Tao Wang
Yuanbin Chen
Xinlin Zhang
Hui Tang
Fuxin Lin
Chunwang Li
Qixuan Li
Tao Tan
Dezhi Kang
Tong Tong
author_facet Longxuan Zhao
Tao Wang
Yuanbin Chen
Xinlin Zhang
Hui Tang
Fuxin Lin
Chunwang Li
Qixuan Li
Tao Tan
Dezhi Kang
Tong Tong
author_sort Longxuan Zhao
collection DOAJ
description Abstract Medical image segmentation represents a pivotal and intricate procedure in the domain of medical image processing and analysis. With the progression of artificial intelligence in recent years, the utilization of deep learning techniques for medical image segmentation has witnessed escalating popularity. Nevertheless, the intricate nature of medical image poses challenges on the segmentation of diminutive targets is still in its early stages. Current networks encounter difficulties in addressing the segmentation of exceedingly small targets, especially when the number of training samples is limited. To overcome this constraint, we have implemented a proficient strategy to enhance lesion images containing small targets and constrained samples. We introduce a segmentation framework termed STS-Net, specifically designed for small target segmentation. This framework leverages the established capacity of convolutional neural networks to acquire effective image representations. The proposed STS-Net network adopts a ResNeXt50-32x4d architecture as the encoder, integrating attention mechanisms during the encoding phase to amplify the feature representation capabilities of the network. We evaluated the proposed network on four publicly available datasets. Experimental results underscore the superiority of our approach in the domain of medical image segmentation, particularly for small target segmentation. The codes are available at https://github.com/zlxokok/STSNet .
format Article
id doaj-art-dbb6cc701a754b56922be30ad8864cf7
institution DOAJ
issn 2045-2322
language English
publishDate 2025-03-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-dbb6cc701a754b56922be30ad8864cf72025-08-20T02:41:33ZengNature PortfolioScientific Reports2045-23222025-03-0115111710.1038/s41598-025-94437-9A novel framework for segmentation of small targets in medical imagesLongxuan Zhao0Tao Wang1Yuanbin Chen2Xinlin Zhang3Hui Tang4Fuxin Lin5Chunwang Li6Qixuan Li7Tao Tan8Dezhi Kang9Tong Tong10College of Physics and Information Engineering, Fuzhou UniversityCollege of Physics and Information Engineering, Fuzhou UniversityCollege of Physics and Information Engineering, Fuzhou UniversityCollege of Physics and Information Engineering, Fuzhou UniversityCollege of Physics and Information Engineering, Fuzhou UniversityDepartment of Neurosurgery, Neurosurgery Research Institute, The First Affiliated Hospital, Fujian Medical UniversityDepartment of Neurosurgery, Neurosurgery Research Institute, The First Affiliated Hospital, Fujian Medical UniversityDepartment of Neurosurgery, Neurosurgery Research Institute, The First Affiliated Hospital, Fujian Medical UniversityMacao Polytechnic UniversityDepartment of Neurosurgery, Neurosurgery Research Institute, The First Affiliated Hospital, Fujian Medical UniversityCollege of Physics and Information Engineering, Fuzhou UniversityAbstract Medical image segmentation represents a pivotal and intricate procedure in the domain of medical image processing and analysis. With the progression of artificial intelligence in recent years, the utilization of deep learning techniques for medical image segmentation has witnessed escalating popularity. Nevertheless, the intricate nature of medical image poses challenges on the segmentation of diminutive targets is still in its early stages. Current networks encounter difficulties in addressing the segmentation of exceedingly small targets, especially when the number of training samples is limited. To overcome this constraint, we have implemented a proficient strategy to enhance lesion images containing small targets and constrained samples. We introduce a segmentation framework termed STS-Net, specifically designed for small target segmentation. This framework leverages the established capacity of convolutional neural networks to acquire effective image representations. The proposed STS-Net network adopts a ResNeXt50-32x4d architecture as the encoder, integrating attention mechanisms during the encoding phase to amplify the feature representation capabilities of the network. We evaluated the proposed network on four publicly available datasets. Experimental results underscore the superiority of our approach in the domain of medical image segmentation, particularly for small target segmentation. The codes are available at https://github.com/zlxokok/STSNet .https://doi.org/10.1038/s41598-025-94437-9
spellingShingle Longxuan Zhao
Tao Wang
Yuanbin Chen
Xinlin Zhang
Hui Tang
Fuxin Lin
Chunwang Li
Qixuan Li
Tao Tan
Dezhi Kang
Tong Tong
A novel framework for segmentation of small targets in medical images
Scientific Reports
title A novel framework for segmentation of small targets in medical images
title_full A novel framework for segmentation of small targets in medical images
title_fullStr A novel framework for segmentation of small targets in medical images
title_full_unstemmed A novel framework for segmentation of small targets in medical images
title_short A novel framework for segmentation of small targets in medical images
title_sort novel framework for segmentation of small targets in medical images
url https://doi.org/10.1038/s41598-025-94437-9
work_keys_str_mv AT longxuanzhao anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT taowang anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT yuanbinchen anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT xinlinzhang anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT huitang anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT fuxinlin anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT chunwangli anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT qixuanli anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT taotan anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT dezhikang anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT tongtong anovelframeworkforsegmentationofsmalltargetsinmedicalimages
AT longxuanzhao novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT taowang novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT yuanbinchen novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT xinlinzhang novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT huitang novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT fuxinlin novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT chunwangli novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT qixuanli novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT taotan novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT dezhikang novelframeworkforsegmentationofsmalltargetsinmedicalimages
AT tongtong novelframeworkforsegmentationofsmalltargetsinmedicalimages