Related Experiment Video
Updated: Dec 19, 2025
State of the Art Cranial Ultrasound Imaging in Neonates
Published on: February 2, 2015
Computer-aided diagnosis for fetal brain ultrasound images using deep convolutional neural networks
Baihong Xie1, Ting Lei2, Nan Wang3
1South China University of Technology, Guangzhou, China.
Insights
Computer-aided diagnosis algorithms using deep learning can accurately detect fetal brain abnormalities from ultrasound images. These tools aid doctors in prenatal assessments, improving early detection and reducing diagnostic errors.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Fetal Medicine
Background:
- Fetal brain abnormalities are common congenital malformations linked to developmental delays.
- Early prenatal detection is crucial for clinical management and parental counseling.
- Accurate antenatal diagnosis aids in identifying syndromic and chromosomal abnormalities.
Purpose of the Study:
- Develop computer-aided diagnosis (CADx) algorithms for five common fetal brain abnormalities.
- Assist clinicians in detecting fetal brain abnormalities during antenatal neurosonographic assessments.
- Enhance the accuracy and efficiency of prenatal diagnosis of fetal brain conditions.
Main Methods:
- Applied a deep convolutional neural network classifier to fetal brain ultrasound images (transventricular and transcerebellar planes).
- Segmented craniocerebral regions, classified segmentations into normal/abnormal categories, and localized lesions using class activation mapping.
- Trained and evaluated algorithms on image-level labeled clinical datasets.
Main Results:
- Achieved a Dice score of 0.942 for craniocerebral region segmentation.
- Obtained an average F1-score of 0.96 for classification of normal versus abnormal images.
- Reached an average mean Intersection over Union (IOU) of 0.497 for lesion localization.
Conclusions:
- Developed effective CADx algorithms for fetal brain ultrasound analysis using deep learning.
- Algorithms show potential for assisting in diagnosis and improving clinical decision-making for junior doctors.
- The system is expected to reduce false negatives in detecting fetal brain abnormalities.
Purpose:
Fetal brain abnormalities are some of the most common congenital malformations that may associated with syndromic and chromosomal malformations, and could lead to neurodevelopmental delay and mental retardation. Early prenatal detection of brain abnormalities is essential for informing clinical management pathways and consulting for parents. The purpose of this research is to develop computer-aided diagnosis algorithms for five common fetal brain abnormalities, which may provide assistance to doctors for brain abnormalities detection in antenatal neurosonographic assessment.
Methods:
We applied a classifier to classify images of fetal brain standard planes (transventricular and transcerebellar) as normal or abnormal. The classifier was trained by image-level labeled images. In the first step, craniocerebral regions were segmented from the ultrasound images. Then, these segmentations were classified into four categories. Last, the lesions in the abnormal images were localized by class activation mapping.
Results:
We evaluated our algorithms on real-world clinical datasets of fetal brain ultrasound images. We observed that the proposed method achieved a Dice score of 0.942 on craniocerebral region segmentation, an average F1-score of 0.96 on classification and an average mean IOU of 0.497 on lesion localization.
Conclusion:
We present computer-aided diagnosis algorithms for fetal brain ultrasound images based on deep convolutional neural networks. Our algorithms could be potentially applied in diagnosis assistance and are expected to help junior doctors in making clinical decision and reducing false negatives of fetal brain abnormalities.
