Kev-Bio improves precision in ultrasound landmark detection
Every Pixel Is an Option: A System One Model for Medical Ultrasound Biometry
Computer Vision and Pattern RecognitionMachine Learning
Summary
Ultrasound exams need exact measurements taken by placing markers on body images, but these often vary between operators. The authors propose Kev-Bio, a new method that uses language-based instructions to find landmarks more accurately by predicting where each marker should go. Their approach models uncertainty in placement by creating probability maps rather than single points. They tested Kev-Bio and found it to be more precise than other methods on a large dataset of ultrasound images. This could help make ultrasound measurements more consistent and reliable.
What this means in practice
- •For medical imaging developers: Build ultrasound tools that use text-guided landmark detection to reduce errors in anatomical measurements across multiple views.
- •For cardiac ultrasound technicians: Improve precision in placing measurement markers for cardiac scans by applying language-driven automated landmark detection.
Authors
Moein Heidari, Junbo Rao, Yudan Chen, Yasamin Medghalchi, River Jiang, Wenjin Chen, David J. Foran, Jasmine Grewal, Ilker Hacihaliloglu
Abstract
**Purpose:** Ultrasound biometry is widely used across various clinical applications, including prenatal and cardiac examinations, to obtain quantitative measurements by placing calipers at anatomical landmarks. However, these measurements remain subject to interoperator variability as they rely largely on manual caliper placement. Conventional methods often use specialized prediction heads for different ultrasound tasks, whereas vision-language models offer flexibility through text instructions. However, direct coordinate prediction does not explicitly model spatial likelihood distributions, limiting the representation of localization ambiguity. **Methods:** We propose Kev-Bio, a vision-language model that uses clinical measurement questions to guide ultrasound landmark detection. Inspired by System One decision models, Kev-Bio uses decision representations of the language model to generate queries for each caliper endpoint. These queries are matched against shared dense ultrasound features to produce spatial probability maps, from which landmark coordinates and clinical measurements are derived. The model is trained end-to-end using both landmark localization and clinical measurement losses. **Results:** On the nine-view FU/GU Biometry 2026 dataset, Kev-Bio achieves a mean radial error (MRE) of 22.50 pixels and a clinical measurement mean absolute error (MAE) of 15.38, outperforming the compared baselines in overall performance. **Conclusion:** These results demonstrate the effectiveness of Kev-Bio for ultrasound landmark detection and clinical measurement across different anatomical views. Our code and models are available at https://github.com/moeinheidari7829/Kev-Bio.