This is a page that curates AI-related papers published worldwide. All content here is summarized using Google Gemini and operated on a non-profit basis. Copyright for each paper belongs to the authors and their institutions; please make sure to credit the source when sharing.
Sequence-aware Pre-training for Echocardiography Probe Movement Guidance
Created by
Haebom
Author
Haojun Jiang, Teng Wang, Zhenguo Sun, Yulin Wang, Yang Yue, Yu Sun, Ning Jia, Meng Li, Shaqi Luo, Shiji Song, Gao Huang
Outline
In this paper, we propose a novel probe motion guidance algorithm to address the shortage of experts in cardiac echocardiography. To address two major challenges, namely the complex structure of the heart and individual differences, we propose a novel sequence-aware self-supervised learning method that learns individual cardiac structures instead of the traditional population-averaged structure learning. The method learns individual 3D cardiac structural features by predicting masked image features and probe motion. Experimental results using a large-scale expert scan dataset containing 1.31 million samples demonstrate that the proposed method effectively reduces probe guidance errors compared to other state-of-the-art baseline methods.
Takeaways, Limitations
•
Takeaways:
◦
Presentation of a novel probe motion guidance algorithm that may contribute to solving the problem of shortage of experts in cardiac ultrasound examination.
◦
Improving the accuracy of cardiac ultrasound image acquisition by considering individual heart structures.
◦
Effective probe guidance error reduction via sequence-aware self-supervised learning.
◦
Presenting the possibility of supporting robotic systems or novice cardiac ultrasound examinations.
•
Limitations:
◦
Validation of the proposed algorithm in practical clinical applications is needed.
◦
Accessibility is limited as code release is scheduled after paper acceptance.
◦
Learning methods that rely on large datasets require consideration of the generalizability and bias issues of the dataset.
