用于显微镜图像中寄生虫卵检测的轻量级深度学习模型。

IF 3 2区 医学 Q1 PARASITOLOGY
Wenbin Xu, Qiang Zhai, Jizhong Liu, Xingyu Xu, Jing Hua
{"title":"用于显微镜图像中寄生虫卵检测的轻量级深度学习模型。","authors":"Wenbin Xu, Qiang Zhai, Jizhong Liu, Xingyu Xu, Jing Hua","doi":"10.1186/s13071-024-06503-2","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Intestinal parasitic infections are still a serious public health problem in developing countries, and the diagnosis of parasitic infections requires the first step of parasite/egg detection of samples. Automated detection can eliminate the dependence on professionals, but the current detection algorithms require large computational resources, which increases the lower limit of automated detection. Therefore, we have designed a lightweight deep-learning model, YAC-Net, to achieve rapid and accurate detection of parasitic eggs and reduce the cost of automation.</p><p><strong>Methods: </strong>This paper uses the ICIP 2022 Challenge dataset for experiments, and the experiments are conducted using fivefold cross-validation. The YOLOv5n model is used as the baseline model, and then two improvements are made to the baseline model based on the specificity of the egg data. First, the neck of the YOLOv5n is modified to from a feature pyramid network (FPN) to an asymptotic feature pyramid network (AFPN) structure. Different from the FPN structure, which mainly integrates semantic feature information at adjacent levels, the hierarchical and asymptotic aggregation structure of AFPN can fully fuse the spatial contextual information of egg images, and its adaptive spatial feature fusion mode can help the model select beneficial feature and ignore redundant information, thereby reducing computational complexity and improving detection performance. Second, the C3 module of the backbone of the YOLOv5n is modified to a C2f module, which can enrich gradient information, improving the feature extraction capability of the backbone. Moreover, ablation studies are designed by us to verify the effectiveness of the AFPN and C2f modules in the process of model lightweighting.</p><p><strong>Results: </strong>The experimental results show that compared with YOLOv5n, YAC-Net improves precision by 1.1%, recall by 2.8%, the F1 score by 0.0195, and mAP_0.5 by 0.0271 and reduces the parameters by one-fifth. Compared with some state-of-the-art detection methods, YAC-Net achieves the best performance in precision, F1 score, mAP_0.5, and parameters. The precision, recall, F1 score, mAP_0.5, and parameters of our method on the test set are 97.8%, 97.7%, 0.9773, 0.9913, and 1,924,302, respectively.</p><p><strong>Conclusions: </strong>Compared with the baseline model, YAC-Net optimizes the model structure and simplifies the parameters while ensuring the detection performance. It helps to reduce the equipment requirements for performing automated detection and can be used to realize the automatic detection of parasite eggs under microscope images.</p>","PeriodicalId":19793,"journal":{"name":"Parasites & Vectors","volume":null,"pages":null},"PeriodicalIF":3.0000,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11539555/pdf/","citationCount":"0","resultStr":"{\"title\":\"A lightweight deep-learning model for parasite egg detection in microscopy images.\",\"authors\":\"Wenbin Xu, Qiang Zhai, Jizhong Liu, Xingyu Xu, Jing Hua\",\"doi\":\"10.1186/s13071-024-06503-2\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><strong>Background: </strong>Intestinal parasitic infections are still a serious public health problem in developing countries, and the diagnosis of parasitic infections requires the first step of parasite/egg detection of samples. Automated detection can eliminate the dependence on professionals, but the current detection algorithms require large computational resources, which increases the lower limit of automated detection. Therefore, we have designed a lightweight deep-learning model, YAC-Net, to achieve rapid and accurate detection of parasitic eggs and reduce the cost of automation.</p><p><strong>Methods: </strong>This paper uses the ICIP 2022 Challenge dataset for experiments, and the experiments are conducted using fivefold cross-validation. The YOLOv5n model is used as the baseline model, and then two improvements are made to the baseline model based on the specificity of the egg data. First, the neck of the YOLOv5n is modified to from a feature pyramid network (FPN) to an asymptotic feature pyramid network (AFPN) structure. Different from the FPN structure, which mainly integrates semantic feature information at adjacent levels, the hierarchical and asymptotic aggregation structure of AFPN can fully fuse the spatial contextual information of egg images, and its adaptive spatial feature fusion mode can help the model select beneficial feature and ignore redundant information, thereby reducing computational complexity and improving detection performance. Second, the C3 module of the backbone of the YOLOv5n is modified to a C2f module, which can enrich gradient information, improving the feature extraction capability of the backbone. Moreover, ablation studies are designed by us to verify the effectiveness of the AFPN and C2f modules in the process of model lightweighting.</p><p><strong>Results: </strong>The experimental results show that compared with YOLOv5n, YAC-Net improves precision by 1.1%, recall by 2.8%, the F1 score by 0.0195, and mAP_0.5 by 0.0271 and reduces the parameters by one-fifth. Compared with some state-of-the-art detection methods, YAC-Net achieves the best performance in precision, F1 score, mAP_0.5, and parameters. The precision, recall, F1 score, mAP_0.5, and parameters of our method on the test set are 97.8%, 97.7%, 0.9773, 0.9913, and 1,924,302, respectively.</p><p><strong>Conclusions: </strong>Compared with the baseline model, YAC-Net optimizes the model structure and simplifies the parameters while ensuring the detection performance. It helps to reduce the equipment requirements for performing automated detection and can be used to realize the automatic detection of parasite eggs under microscope images.</p>\",\"PeriodicalId\":19793,\"journal\":{\"name\":\"Parasites & Vectors\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":3.0000,\"publicationDate\":\"2024-11-06\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11539555/pdf/\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Parasites & Vectors\",\"FirstCategoryId\":\"3\",\"ListUrlMain\":\"https://doi.org/10.1186/s13071-024-06503-2\",\"RegionNum\":2,\"RegionCategory\":\"医学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"PARASITOLOGY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Parasites & Vectors","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1186/s13071-024-06503-2","RegionNum":2,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"PARASITOLOGY","Score":null,"Total":0}
引用次数: 0

摘要

背景:在发展中国家,肠道寄生虫感染仍是一个严重的公共卫生问题,而寄生虫感染的诊断首先需要对样本进行寄生虫/卵检测。自动检测可以消除对专业人员的依赖,但目前的检测算法需要大量的计算资源,这增加了自动检测的下限。因此,我们设计了一种轻量级深度学习模型--YAC-Net,以实现寄生虫卵的快速准确检测,降低自动化检测成本:本文使用 ICIP 2022 挑战赛数据集进行实验,实验采用五重交叉验证。以 YOLOv5n 模型作为基线模型,然后根据虫卵数据的特异性对基线模型做了两点改进。首先,将 YOLOv5n 的颈部从特征金字塔网络(FPN)修改为渐近特征金字塔网络(AFPN)结构。与 FPN 结构主要整合相邻层次的语义特征信息不同,AFPN 的分层渐近聚合结构可以充分融合鸡蛋图像的空间上下文信息,其自适应空间特征融合模式可以帮助模型选择有利特征,忽略冗余信息,从而降低计算复杂度,提高检测性能。其次,将 YOLOv5n 主干网的 C3 模块修改为 C2f 模块,可以丰富梯度信息,提高主网的特征提取能力。此外,我们还设计了消融研究,以验证 AFPN 和 C2f 模块在模型轻量化过程中的有效性:实验结果表明,与 YOLOv5n 相比,YAC-Net 的精度提高了 1.1%,召回率提高了 2.8%,F1 分数提高了 0.0195,mAP_0.5 提高了 0.0271,参数降低了五分之一。与一些最先进的检测方法相比,YAC-Net 在精度、F1 分数、mAP_0.5 和参数方面都达到了最佳性能。我们的方法在测试集上的精度、召回率、F1 分数、mAP_0.5 和参数分别为 97.8%、97.7%、0.9773、0.9913 和 1 924 302:与基线模型相比,YAC-Net 优化了模型结构,简化了参数,同时保证了检测性能。它有助于降低进行自动检测的设备要求,可用于实现显微镜图像下寄生虫卵的自动检测。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
A lightweight deep-learning model for parasite egg detection in microscopy images.

Background: Intestinal parasitic infections are still a serious public health problem in developing countries, and the diagnosis of parasitic infections requires the first step of parasite/egg detection of samples. Automated detection can eliminate the dependence on professionals, but the current detection algorithms require large computational resources, which increases the lower limit of automated detection. Therefore, we have designed a lightweight deep-learning model, YAC-Net, to achieve rapid and accurate detection of parasitic eggs and reduce the cost of automation.

Methods: This paper uses the ICIP 2022 Challenge dataset for experiments, and the experiments are conducted using fivefold cross-validation. The YOLOv5n model is used as the baseline model, and then two improvements are made to the baseline model based on the specificity of the egg data. First, the neck of the YOLOv5n is modified to from a feature pyramid network (FPN) to an asymptotic feature pyramid network (AFPN) structure. Different from the FPN structure, which mainly integrates semantic feature information at adjacent levels, the hierarchical and asymptotic aggregation structure of AFPN can fully fuse the spatial contextual information of egg images, and its adaptive spatial feature fusion mode can help the model select beneficial feature and ignore redundant information, thereby reducing computational complexity and improving detection performance. Second, the C3 module of the backbone of the YOLOv5n is modified to a C2f module, which can enrich gradient information, improving the feature extraction capability of the backbone. Moreover, ablation studies are designed by us to verify the effectiveness of the AFPN and C2f modules in the process of model lightweighting.

Results: The experimental results show that compared with YOLOv5n, YAC-Net improves precision by 1.1%, recall by 2.8%, the F1 score by 0.0195, and mAP_0.5 by 0.0271 and reduces the parameters by one-fifth. Compared with some state-of-the-art detection methods, YAC-Net achieves the best performance in precision, F1 score, mAP_0.5, and parameters. The precision, recall, F1 score, mAP_0.5, and parameters of our method on the test set are 97.8%, 97.7%, 0.9773, 0.9913, and 1,924,302, respectively.

Conclusions: Compared with the baseline model, YAC-Net optimizes the model structure and simplifies the parameters while ensuring the detection performance. It helps to reduce the equipment requirements for performing automated detection and can be used to realize the automatic detection of parasite eggs under microscope images.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
Parasites & Vectors
Parasites & Vectors 医学-寄生虫学
CiteScore
6.30
自引率
9.40%
发文量
433
审稿时长
1.4 months
期刊介绍: Parasites & Vectors is an open access, peer-reviewed online journal dealing with the biology of parasites, parasitic diseases, intermediate hosts, vectors and vector-borne pathogens. Manuscripts published in this journal will be available to all worldwide, with no barriers to access, immediately following acceptance. However, authors retain the copyright of their material and may use it, or distribute it, as they wish. Manuscripts on all aspects of the basic and applied biology of parasites, intermediate hosts, vectors and vector-borne pathogens will be considered. In addition to the traditional and well-established areas of science in these fields, we also aim to provide a vehicle for publication of the rapidly developing resources and technology in parasite, intermediate host and vector genomics and their impacts on biological research. We are able to publish large datasets and extensive results, frequently associated with genomic and post-genomic technologies, which are not readily accommodated in traditional journals. Manuscripts addressing broader issues, for example economics, social sciences and global climate change in relation to parasites, vectors and disease control, are also welcomed.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信