Facing a shortage of qualified workers, healthcare facilities across the US are looking for work-arounds that will help them function effectively with a smaller staff. Our client was no exception. They needed to find a way to more efficiently monitor their medical facility video surveillance recordings, and we created an elegant solution for them.
The cloud made it easy to get a working proof of concept up quickly.
We created an AI proof-of-concept using cloud-native serverless technologies
Leverages an AWS serverless implementation that processes surveillance video, detects abnormal events, and sends email notifications to human workers
The system also classifies objects and certain human actions in the video stream
Our prototype proves that the idea to build an application to analyze video streams and detect human action is viable.
The AWS Rekognition API can solve the simple problem of automating video surveillance
It not only has pretrained computer vision models with thousands of labels but also provides tools that allow you to train custom label models on your own data set