SAMURAI
Sensor-Aware MRI Using Real-time AI Inference
The SAMURAI project aims to boost the market position of the Netherlands as one of the world leading manufacturers of medical technology. Philips, a global, leading manufacturer of medical scanners with rich expertise in advanced medical technology, and TU Delft a leading research university with strong track record in real-time, application specific computing, will join forces to make MRI technology faster, more predictable, maintainable and highly cost effective by using modern AI methods.
Data driven foundation model for MRI scanners
This proposal postulates a new paradigm: using a data-driven foundation model trained on Philips' extensive repository of MRI scans, patient metadata, and sensor readout records. The model will take into account the actual physical location of different data sources within the image reconstruction chain of the scanner and will learn the mapping between partial scan data and optimal sensor configurations, thereby enabling real-time, predictive adaptation of image acquisition control.