人工智能
计算机科学
计算机视觉
超声波
人工神经网络
生物医学工程
模式识别(心理学)
特征(语言学)
医学
噪音(视频)
信号处理
自动化
作者
Zhihao Li,Liyan Chen,Ye Ding,Xiuxiu Hao
出处
期刊:
[Institute of Electrical and Electronics Engineers]
日期:2026-04-21
卷期号:: 8052-8056
标识
DOI:10.1109/icassp55912.2026.11463995
摘要
The classification of fetal ultrasound (US) planes is a critical step in prenatal diagnosis and biometric assessment. Manual screening is time-consuming and subject to considerable inter-operator variability, hindering standardized evaluation. Existing automated methods often struggle to simultaneously capture fine-grained local features and long-range dependencies. To overcome this limitation, we propose FetalUS-Net, a four-stage hierarchical network designed for fetal US plane classification. Specifically, we introduce a novel dual-branch block named CBAMamba, which integrates a Convolutional Block Attention Module (CBAM) for local discriminative features with Mamba to capture long-range dependencies. We evaluated FetalUS-Net on two fetal US datasets collected from different regions. Experimental results demonstrate that our method achieves robust classification accuracy and strong cross-domain generalization, highlighting its potential as a reliable solution for automated fetal US plane recognition.
科研通智能强力驱动
Strongly Powered by AbleSci AI