Doctors already use MRI to see a brain tumor. The hard part is drawing the tumor on the images, slice by slice.
To break down what's happening here is that a 3D model studied MRI scans from 300 real people. Then it was shown 60 new scans it had never seen and asked to mark the tumor areas.
The software guesses where the dead core is, where the swelling is, and where the active tumor is. A doctor can then check and fix the map instead of starting from a blank scan.
These are but a few ways AI has been implemented in the medical imaging field.
In real clinics, tools like this only help if a human still reviews them.
Mapping a brain tumour by hand takes scarce specialist time. AI gives doctors a faster starting point.
On AxonOS, a 3D model learned from MRI data from 300 real subjects, then mapped tumour regions in 60 scans it had never seen before.
Powered by AxonOS.