A Hybrid Deep Learning Approach for Automated Fetal Brain Diagnosis

DOI

10.2991/978-94-6463-858-5_118How to use a DOI?

Keywords

Fetal Brain Abnormalities; Prenatal Diagnostics; Convolutional Neural Networks (CNNs); Efficient -Net; Texture Analysis; Gray-Level Co-occurrence Matrix (GLCM)

Abstract

Fetal brain abnormality detection is a vital task in prenatal care, requiring the precise and efficient analysis of medical imaging data to ensure early diagnosis and intervention. The study introduces an advanced classification framework that synergizes Convolutional Neural Network (CNN), extraction of texture-based features, and attention mechanisms to achieve superior diagnostic accuracy. This novel approach leverages the pre-trained EfficientNetB0, a state-of-the-art CNN architecture known for its efficient scaling and superior feature extraction capabilities. EfficientNetB0 extracts hierarchical spatial features from fetal brain images, capturing complex patterns and structural information essential for anomaly detection. To complement the spatial features, the Gray-Level Co-occurrence Matrix (GLCM) is employed to extract statistical texture features, including contrast, correlation, energy, and homogeneity. These characteristics offer important new information about the textural patterns in the pictures, enabling the detection of subtle abnormalities that may not be captured by spatial features alone. The integration of spatial and texture-based features enhances the model’s ability to recognize diverse abnormalities in fetal brain structures. An attention mechanism are incorporated into the framework to prioritize the most relevant features from both spatial and texture domains. This mechanism ensures optimal feature fusion by focusing on critical regions and attributes, improving the interpretability and performance of the model. The combined features are processed through fully connected layers, making it possible to distinguish between regular and abnormal fetal brain pictures categories with high precision. The proposed model is trained and evaluated on a diverse dataset of fetal brain ultrasound images, encompassing various gestational stages and conditions. Rigorous testing demonstrates the model’s ability to achieve high classification accuracy, robustness, and generalizability across different imaging scenarios. This approach highlights the potential of integrating advanced CNN architectures, texture analysis techniques, and attention-based feature fusion to create reliable and automated diagnostic systems for prenatal care. The integration of these advanced techniques addresses common challenges in fetal brain abnormality detection, such as noisy data, class imbalance, and variability in imaging protocols. By combining spatial and textural information, the proposed framework achieves a more comprehensive representation of the fetal brain, improving the sensitivity and specificity of anomaly detection. This study underscores the transformative potential of deep learning in prenatal diagnostics, paving the way for early and accurate detection of fetal brain abnormalities and improved neonatal outcomes.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - D.  Anupama
AU  - Vimala Nagabotu
AU  - Giridhar Bagadi
AU  - K. S. L. Prasanna
AU  - A.  Ravi Kumar
AU  - Kappera Ramesh
AU  - Donapati Srikanth
PY  - 2025
DA  - 2025/11/04
TI  - A Hybrid Deep Learning Approach for Automated Fetal Brain Diagnosis
BT  - Proceedings of International Conference on Computer Science and Communication Engineering (ICCSCE 2025)
PB  - Atlantis Press
SP  - 1419
EP  - 1437
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-858-5_118
DO  - 10.2991/978-94-6463-858-5_118
ID  - Anupama2025
ER  -