Tuesday, 5 July 2011

Centralized Medical Diagnosis using AI and Azure

Scientio have been working for some time with a US based medical device company to create a centralized location on Azure for the analysis and reporting of soft tissue injury data collected from semi smart data collection pods at clinics and surgeries.

We’re finalists in a UK Government health challenge with this project, and you can read more here.

Much of what we’re doing with our customer is proprietary and specialized to their needs, but I wanted to share with you the Azure architecture we’re using, because that is generic for a wide range of possible applications in medicine and other fields where data is collected locally and analyzed and reported centrally.


This architecture uses almost every Azure feature. The Client communicates via Blob uploads, Web Services and Queues.

The workers, possibly many, process tasks that are split into the smallest possible chunks and sequenced and parallelized using queues.

SQL Azure is used to record the non-blob data, while the Blob data, time series in this application is sored in Azure Blob storage.

Anyone who’s used Azure to run websites knows that SQL Azure doesn’t do Forms based authentication. You have to use Azure tables instead, so even these get a look-in for the STS, which is used with AppFabric ACS to secure everything.

With our XmlMiner Fuzzy logic inference engine working away within the worker processes, we can create a solution for a wide range of similar projects. This solution is very and rapidly scalable by creating more identical worker instances to handle greater throughput.

Because it’s hosted on Azure there is no need to buy servers or pay maintenance staff.

Contact us if you are interested in a similar solution at sales@scientio.com.

No comments: