SSOAR Logo
    • Deutsch
    • English
  • English 
    • Deutsch
    • English
  • Login
SSOAR ▼
  • Home
  • About SSOAR
  • Guidelines
  • Publishing in SSOAR
  • Cooperating with SSOAR
    • Cooperation models
    • Delivery routes and formats
    • Projects
  • Cooperation partners
    • Information about cooperation partners
  • Information
    • Possibilities of taking the Green Road
    • Grant of Licences
    • Download additional information
  • Operational concept
Browse and search Add new document OAI-PMH interface
JavaScript is disabled for your browser. Some features of this site may not work without it.

Download PDF
Download full text

(external source)

Citation Suggestion

Please use the following Persistent Identifier (PID) to cite this document:
https://doi.org/10.22178/pos.119-20

Exports for your reference manager

Bibtex export
Endnote export

Display Statistics
Share
  • Share via E-Mail E-Mail
  • Share via Facebook Facebook
  • Share via Bluesky Bluesky
  • Share via Reddit reddit
  • Share via Linkedin LinkedIn
  • Share via XING XING

Deep Learning Approaches to Identify Subtle Anomalies in Prenatal Ultrasound Imaging

[journal article]

Agboola, Olasoji O.
Kuye, Oludare Olukayode
Adenowo, Thomas K.

Abstract

This research investigated deep learning approaches for detecting subtle anomalies in prenatal ultrasound imaging. Congenital anomalies affect approximately 6% of births worldwide, with detection rates for subtle defects varying significantly based on operator expertise. A multi-institutional datase... view more

This research investigated deep learning approaches for detecting subtle anomalies in prenatal ultrasound imaging. Congenital anomalies affect approximately 6% of births worldwide, with detection rates for subtle defects varying significantly based on operator expertise. A multi-institutional dataset comprising 12,450 prenatal ultrasound examinations from three tertiary care centres was employed to develop and evaluate multiple deep learning architectures, including modified convolutional neural networks, generative adversarial networks, autoencoders, and feature fusion approaches. The ensemble approach, which combines these architectures, achieved an overall accuracy of 91.4% and 89.8% accuracy for subtle anomalies, specifically substantially exceeding previous benchmarks. Feature visualisation confirmed that models focused on anatomically appropriate regions when making predictions. Performance varied across anomaly categories, with cardiac defects presenting the most significant challenges. The research identified meaningful relationships between model confidence and clinical significance, with higher sensitivity for anomalies requiring immediate intervention. Expert evaluation confirmed that models occasionally detected subtle findings that were missed during routine interpretation, suggesting a potential complementary role between automated systems and human expertise. The findings demonstrate significant progress toward addressing the challenges of subtle anomaly detection in prenatal ultrasound while identifying important directions for future refinement.... view less

Keywords
data protection

Classification
Medicine, Social Medicine

Free Keywords
Healthcare policymakers; Healthcare organisations; Data governance; Privacy protocols; Ethical implementation; Robust data management; Integration of diagnostic technologies

Document language
English

Publication Year
2025

Page/Pages
p. 3019-3026

Journal
Path of Science, 11 (2025) 6

ISSN
2413-9009

Status
Published Version; peer reviewed

Licence
Creative Commons - Attribution 4.0


GESIS LogoDFG LogoOpen Access Logo
Home  |  Legal notices  |  Operational concept  |  Privacy policy
© 2007 - 2025 Social Science Open Access Repository (SSOAR).
Based on DSpace, Copyright (c) 2002-2022, DuraSpace. All rights reserved.
 

 


GESIS LogoDFG LogoOpen Access Logo
Home  |  Legal notices  |  Operational concept  |  Privacy policy
© 2007 - 2025 Social Science Open Access Repository (SSOAR).
Based on DSpace, Copyright (c) 2002-2022, DuraSpace. All rights reserved.