Inference for Intervention Translation (in4it)

This platform will develop a software suite with custom-tested hardware for deploying modern machine learning algorithms in intervention and surgery.

There are growing number of surgical and interventional applications employing machine learning with deep neural networks, some having already been validated with retrospective imaging data. Translating these innovative assistive and enabling algorithms into procedure rooms or surgical theatres requires specific hardware testing, efficient implementation, network re-training, robust and generalisable learning models and in-use data collection. Many of them are common yet specific requirements for integrating deep learning solutions in time-critical procedures guided by real-time medical imaging, while there are not supported by existing deep-learning packages.