Deep Learning-Based Computer-Aided Classification of Amniotic Fluid Using Ultrasound Images from Saudi Arabia

Author:

Khan Irfan UllahORCID,Aslam NidaORCID,Anis Fatima M.ORCID,Mirza SamihaORCID,AlOwayed AlanoudORCID,Aljuaid Reef M.ORCID,Bakr Razan M.ORCID,Qahtani Nourah Hasan Al

Abstract

Amniotic Fluid (AF) refers to a protective liquid surrounding the fetus inside the amniotic sac, serving multiple purposes, and hence is a key indicator of fetal health. Determining the AF levels at an early stage helps to ascertain the maturation of lungs and gastrointestinal development, etc. Low AF entails the risk of premature birth, perinatal mortality, and thereby admission to intensive care unit (ICU). Moreover, AF level is also a critical factor in determining early deliveries. Hence, AF detection is a vital measurement required during early ultrasound (US), and its automation is essential. The detection of AF is usually a time-consuming process as it is patient specific. Furthermore, its measurement and accuracy are prone to errors as it heavily depends on the sonographer’s experience. However, automating this process by developing robust, precise, and effective methods for detection will be beneficial to the healthcare community. Therefore, in this paper, we utilized transfer learning models in order to classify the AF levels as normal or abnormal using the US images. The dataset used consisted of 166 US images of pregnant women, and initially the dataset was preprocessed before training the model. Five transfer learning models, namely, Xception, Densenet, InceptionResNet, MobileNet, and ResNet, were applied. The results showed that MobileNet achieved an overall accuracy of 0.94. Overall, the proposed study produces an effective result in successfully classifying the AF levels, thereby building automated, effective models reliant on transfer learning in order to aid sonographers in evaluating fetal health.

Publisher

MDPI AG

Subject

Artificial Intelligence,Computer Science Applications,Information Systems,Management Information Systems

Cited by 4 articles. 订阅此论文施引文献 订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献

1. Automated Detection of Abnormalities in Ultrasound Images Using Deep Learning;2024 International Conference on Optimization Computing and Wireless Communication (ICOCWC);2024-01-29

2. ESDPCOS: Effectiveness of Combined CNN and Handcrafted Features for Ovarian Cyst Detection in PCOS Patients using Ultrasound Images;2023 10th IEEE Uttar Pradesh Section International Conference on Electrical, Electronics and Computer Engineering (UPCON);2023-12-01

3. DLBCNet: A Deep Learning Network for Classifying Blood Cells;Big Data and Cognitive Computing;2023-04-14

4. A Deep Learning Fusion Approach to Diagnosis the Polycystic Ovary Syndrome (PCOS);Applied Computational Intelligence and Soft Computing;2023-02-14

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3