There are cheap MRI scanners already in use in Indian district hospitals. However, their imaging is of low quality, noisy and difficult to interpret. The solution: there are the scanners but they are useless.
The technology that makes MRI images interpretable is LARMOR: an approach that uses software to convert the scan data into clean and diagnostic imaging. No new and expensive equipment is required.
Why does it work: while current approaches attempt to make a pretty image hoping that it will be accurate, LARMOR presents the physician the image and shows which areas of the picture correspond to the data received in the scan, and which are inferred by the algorithm. Thus, in case of error, it will be easy to detect it since a doctor will know where the AI created a fiction.
This technology is a retrofit: it can be deployed on scanners the hospital already has.
The current stage of development is ideation: the technology's physics and architecture have been defined, the datasets are public and identified, and the build plan is ready. However, there is no trained model and results available yet.
Log in or sign up for Devpost to join the conversation.