Optical Flow-Enhanced Mamba U-Net for Cardiac Phase Detection in Ultrasound Videos
Lu, Yuhuan and Tan, Guanghua and Pu, Bin and Yeung, Pak-Hei and Wang, Hang and Li, Shengli and Rajapakse, Jagath C. and Li, Kenli
IEEE Transactions on Medical Imaging (TMI), 2025.
paper / code
An optical flow-enhanced Mamba U-Net framework designed to leverage both short-term motion cues and long-term temporal dependencies for accurate cardiac phase detection in ultrasound videos.

AP-Net: Semi-Supervised Ultrasound Cardiac Segmentation Using Enhanced Anatomical Prior
Lu, Yuhuan and Li, Jintang and Lin, Jianxin and Yuan, Ying and Rajapakse, Jagath C. and Zhu, Ningbo and Wang, Chunlian and Li, Kenli
ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2025.
paper / code
A semi-supervised segmentation model that effectively leverages anatomical priors through an anatomical prior generation module, a prior-feature fusion module, and a category-aware cropping strategy.

SKGC: A General Semantic-Level Knowledge Guided Classification Framework for Fetal Congenital Heart Disease
Lu, Yuhuan and Tan, Guanghua and Pu, Bin and Wang, Hang and Liang, Bocheng and Li, Kenli and Rajapakse, Jagath C.
IEEE Journal of Biomedical and Health Informatics (JBHI), 2024.
paper / code
A universal framework for identifying normal and abnormal four-chamber heart images, guided by a limited number of annotation masks, with significantly improved accuracy.

A YOLOX-Based Deep Instance Segmentation Neural Network for Cardiac Anatomical Structures in Fetal Ultrasound Images
Lu, Yuhuan and Li, Kenli and Pu, Bin and Tan, Ying and Zhu, Ningbo
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB), 2022.
paper
A YOLOX-based deep instance segmentation network designed for localizing and segmenting cardiac anatomical structures in fetal ultrasound images.

MobileUNet-FPN: A Semantic Segmentation Model for Fetal Ultrasound Four-Chamber Segmentation in Edge Computing Environments
Pu, Bin and Lu, Yuhuan and Chen, Jianguo and Li, Shengli and Zhu, Ningbo and Wei, Wei and Li, Kenli
IEEE Journal of Biomedical and Health Informatics (JBHI), 2022.
paper
MobileUNet-FPN is a lightweight AI model that integrates MobileNet, UNet, and an explicit FPN to segment 13 key anatomical structures in fetal apical four-chamber ultrasound images with high accuracy.