U-shape-based network for left ventricular segmentation in echocardiograms with contrastive pretraining
Zhengkun Qian, Tao Hu, Jianming Wang, Zizhong Yang · 2024
Read U-shape-based network for left ventricular segmentation in echocardiograms with contrastive pretraining by Zhengkun Qian, Tao Hu, Jianming Wang, Zizhong Yang (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-11607360.txt.
Description
Zhengkun Qian and three co-authors build a lean AI model to outline the heart’s main pumping chamber on ultrasound scans, in a 2024 paper. Doctors draw that outline by hand to get chamber volume and ejection fraction, two core measures of heart health. Many top models chase tiny gains in outline accuracy with heavy compute that clinics cannot spare at scan time.
Their model pairs a fast encoder design with a light decoder in a U-shaped layout. Before the main training, a contrastive pre-training step teaches the encoder general image traits. The aim is near-top accuracy at a cost a normal clinic machine can bear. Faster, cheaper outlines could speed heart checks in busy and rural clinics alike.
Catalog record
| CREATOR | Zhengkun Qian, Tao Hu, Jianming Wang, Zizhong Yang |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |