Fine-Tuning and Evaluating AI Models for Organ-at-Risk Delineation in Radiotherapy Planning
Manually outlining organs for radiotherapy is slow and subjective. We trained an AI to take over this, and while its accuracy was high, its "failures" were the telling. It turns that it was more consistent than the human, strictly following its own rules even when the original training data didn't.
Preview
Recommended Content
Social Sciences, Scholars' Stories, Research, Durham University, University College London, University of Leeds, University of St Andrews, University of York, Leadership & Research Laidlaw Scholars, Alumni, Opportunities, University of Cambridge, London School of Economics and Political Science, Imperial College London, University of Oxford - SDG Impact Lab
Help a fellow scholar by participating in a 10 minute survey: UK Men aged 18-25.
Please sign in
If you are a registered user on Laidlaw Scholars Network, please sign in