Deep learning classification of liver lesions from contrast-enhanced ultrasound images (CEUS) to improve diagnostic confidence.
Designed and trained neural networks for automated classification of liver lesion types from noisy, low-contrast CEUS imagery a modality that requires specialized preprocessing and domain-specific augmentation.
Stack
Challenge
Bracco Suisse needed to improve diagnostic confidence in contrast-enhanced ultrasound examinations by automatically classifying liver lesion types from noisy, low-contrast CEUS images where temporal contrast dynamics and image artifacts make standard classification approaches unreliable.
Approach
Designed custom neural network architectures for CEUS image processing, implementing domain-specific augmentation strategies that account for ultrasound artifacts, contrast agent dynamics, and temporal perfusion patterns. Built a preprocessing pipeline tailored to the unique characteristics of contrast-enhanced ultrasound data.
Results
- ●Multi-class lesion classification
- ●CEUS-specific preprocessing pipeline
- ●Domain-adapted augmentation strategies
Working on a similar problem?
Let's talk about how production-grade ML can move your roadmap forward.