Pgool is a next generation healthcare solution that helps providers to increase its efficiency with integrated decision support tools and utilizes field tested and validated technology. They develop computer-aided detection systems for recognition of all known medicines. Their solutions are based on the state-of-the-art computer vision, deep learning and artificial intelligence technology.
Patients who are on many different medications are always at risk of accidentally taking the wrong pills. Since the consequences of taking the wrong medication could be fatal, the client is working on developing a system that would eliminate such a risk once and for all. The client was looking to create an AI/ML application that could discern between all of the medications and notify the patient in the event they pick up the wrong medication.
Since the client did not have an in-house IT department, they relied on Team Harbour for all aspects of development. We needed to choose the right methodology for creating the AI system. We created a proof of concept of the application by applying new methods to find accurate results from large data sets.
We utilized a wide array of Amazon Web Services to create the solution. More specifically, we used EC2 instances for web-scale cloud computing which allowed us to run large-scale Big Data Analytics workloads and accelerated data processing. We also used CDN CloudFront to deliver dynamic web content to provide a better user experience, GPU instances for model training, and S3 storage services to store Big Data for testing
Team Harbour created an MVP that fulfilled purposes for both pharmaceutical and personal use: • Pharmaceutical use - Packaging process where the person that collects the pills for clients may use the application as a validation tool to verify the right pill and markings • Personal use - is for ordinary people to identify and validate the pills that they have.