Digital Health Hub for AMR researcher wins best poster prize at Diagnostics for Low- and Middle-Income Countries conference

We are delighted to share that Elliott Rogers, a PhD student working on the EPSRC Digital Health Hub for AMR was awarded Best Poster at the 2025 Diagnostics for Low- and Middle-Income Countries (LMICs) Conference held at Reuben College, University of Oxford, last month. You can find out more about the annual conference here.

Elliott’s winning poster showcased an upcoming pre-print: "Harnessing Mechanistic Simulators for Rapid Diagnostic Test Capture and Deep Learning Classification"

This study presents SynSight, a novel classification pipeline for Rapid Diagnostic Tests (RDTs) trained entirely on mechanistically generated synthetic data. RDTs are widely used for affordable, accessible disease diagnosis, however their interpretation is subjective and prone to errors. Machine learning models can aid their interpretation but typically require large, costly, and clinically sensitive real-world image libraries. SynSight offers a faster, more flexible alternative, combining segmentation of test regions of interest with real-time object detection and test result classification. In validation tests, SynSight achieved:

• 98.2% sensitivity and 99.2% specificity with HIV RDTs

• 99% and 97% accuracy for two COVID-19 antigen RDTs

• 95% accuracy for a COVID-19 antibody RDT

The full pre-print can be accessed here.

Elliott Rogers said: "I am honoured to have received the Best Poster Prize. It was a fantastic opportunity to share our work on SynSight with such a diverse and inspiring group of researchers, clinicians, and policymakers. What excites me most about this project is its potential to accelerate the development of diagnostic tools in settings where they are needed most, even when real-world training data is limited.”

Elliott (right) pictured with Professor Lord Tarassenko CBE FREng FMedSci, President of Reuben College, University of Oxford (middle), and Iva Grujic, DPhil (PhD) student at the Stevens Group, University of Oxford.