University of Oxford OMNI Oxford Machine
Learning in
NeuroImaging Lab
Research

Problems we work on

Ultrasound is the most widely used imaging in pregnancy care, and often the only one available. We develop methods that turn routine scans into quantitative measures of the developing brain.

Computational ultrasound: geometry and physics

Geometric and physics-informed methods for ultrasound, from recovering 3D brain anatomy from freehand 2D scans to reducing artefacts such as acoustic shadows.

Methods: Multi-view geometry · physics-based rendering · inverse problems

Explore this theme →

Structure: atlases and segmentation

Building spatiotemporal atlases of the developing brain, and using them for registration, segmentation and shape modelling.

Methods: Segmentation · registration · statistical modelling

Explore this theme →

Robust everywhere: harmonisation and deployment

Building models that generalise across scanners, sites and clinical settings.

Methods: Domain shift · robustness · efficient and federated ML

Explore this theme →

From scan to outcome: brain growth and neurodevelopment

Modelling brain growth across gestation and linking prenatal imaging to neurodevelopment after birth.

Methods: Longitudinal modelling · biostatistics · clinical collaboration

Explore this theme →

Software and data

Open-source software

OMNI Ultrasound Toolkit

Open-source tools that automate fetal brain segmentation and volumetric analysis from 3D ultrasound.

Data and partners

Built on international cohorts

Our work draws on the INTERGROWTH-21st and INTERBIO-21st consortia and our clinical partners.