Tianhan Hu, Jiao Ding, Yuting Liu, Yantao Zhang, Li Yang
{"title":"DAA-UNet: A Dense Connectivity and Atrous Spatial Pyramid Pooling Attention UNet Model for Retinal Optical Coherence Tomography Fluid Segmentation","authors":"Tianhan Hu, Jiao Ding, Yuting Liu, Yantao Zhang, Li Yang","doi":"10.1049/sfw2/6006074","DOIUrl":null,"url":null,"abstract":"<div>\n <p>Retinal optical coherence tomography (OCT) fluid segmentation is a vital tool for diagnosing and treating various ophthalmic diseases. Based on clinical manifestations, retinal fluid accumulation is classified into three categories: intraretinal fluid (IRF), subretinal fluid (SRF), and pigment epithelial detachment (PED). PED is primarily associated with diabetic macular edema (DME). In contrast, IRF and SRF play critical roles in diagnosing age-related macular degeneration (AMD) and retinal vein occlusion (RVO). To address challenges posed by variations in OCT imaging devices, as well as the varying sizes, irregular shapes, and blurred boundaries of fluid accumulation areas, this study proposes DAA-UNet, an enhanced UNet architecture. The proposed model incorporates dense connectivity, Atrous Spatial Pyramid Pooling (ASPP), and attention gate (AG) in the paths of UNet. Dense connectivity expands the model’s depth, whereas ASPP facilitates the extraction of multiscale image features. The AG emphasize critical spatial location information, improving the model’s ability to distinguish different fluid accumulation types. Experimental results on the MICCAI 2017 RETOUCH challenge dataset showed that DAA-UNet demonstrates superior performance, with a mean Dice Similarity Coefficient (<i>mDSC</i>) of 90.2%, 91.6%, and 90.5% on cirrus, spectralis, and topcon devices, respectively. These results outperform existing models, including UNet, SFU, Attention-UNet, Deeplabv3+, nnUNet RASPP, and MsTGANet.</p>\n </div>","PeriodicalId":50378,"journal":{"name":"IET Software","volume":"2025 1","pages":""},"PeriodicalIF":1.3000,"publicationDate":"2025-05-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1049/sfw2/6006074","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IET Software","FirstCategoryId":"94","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1049/sfw2/6006074","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"COMPUTER SCIENCE, SOFTWARE ENGINEERING","Score":null,"Total":0}
引用次数: 0
Abstract
Retinal optical coherence tomography (OCT) fluid segmentation is a vital tool for diagnosing and treating various ophthalmic diseases. Based on clinical manifestations, retinal fluid accumulation is classified into three categories: intraretinal fluid (IRF), subretinal fluid (SRF), and pigment epithelial detachment (PED). PED is primarily associated with diabetic macular edema (DME). In contrast, IRF and SRF play critical roles in diagnosing age-related macular degeneration (AMD) and retinal vein occlusion (RVO). To address challenges posed by variations in OCT imaging devices, as well as the varying sizes, irregular shapes, and blurred boundaries of fluid accumulation areas, this study proposes DAA-UNet, an enhanced UNet architecture. The proposed model incorporates dense connectivity, Atrous Spatial Pyramid Pooling (ASPP), and attention gate (AG) in the paths of UNet. Dense connectivity expands the model’s depth, whereas ASPP facilitates the extraction of multiscale image features. The AG emphasize critical spatial location information, improving the model’s ability to distinguish different fluid accumulation types. Experimental results on the MICCAI 2017 RETOUCH challenge dataset showed that DAA-UNet demonstrates superior performance, with a mean Dice Similarity Coefficient (mDSC) of 90.2%, 91.6%, and 90.5% on cirrus, spectralis, and topcon devices, respectively. These results outperform existing models, including UNet, SFU, Attention-UNet, Deeplabv3+, nnUNet RASPP, and MsTGANet.
期刊介绍:
IET Software publishes papers on all aspects of the software lifecycle, including design, development, implementation and maintenance. The focus of the journal is on the methods used to develop and maintain software, and their practical application.
Authors are especially encouraged to submit papers on the following topics, although papers on all aspects of software engineering are welcome:
Software and systems requirements engineering
Formal methods, design methods, practice and experience
Software architecture, aspect and object orientation, reuse and re-engineering
Testing, verification and validation techniques
Software dependability and measurement
Human systems engineering and human-computer interaction
Knowledge engineering; expert and knowledge-based systems, intelligent agents
Information systems engineering
Application of software engineering in industry and commerce
Software engineering technology transfer
Management of software development
Theoretical aspects of software development
Machine learning
Big data and big code
Cloud computing
Current Special Issue. Call for papers:
Knowledge Discovery for Software Development - https://digital-library.theiet.org/files/IET_SEN_CFP_KDSD.pdf
Big Data Analytics for Sustainable Software Development - https://digital-library.theiet.org/files/IET_SEN_CFP_BDASSD.pdf