Automated Detection of Behavourial Patterns for Healthcare Applications
We created an embedded platform to monitor patients in their premises and signal an alarm when someone falls off their bed, has an accident, etc., which can be a big problem for residents of hospitals and old-age homes.
The project included the development of an Edge AI framework that provides the necessary tools and processes for rapid deployment of AI/ML research results into AI enabled applications on edge devices. It provides: (a) extensible inference engines, (b) platform-independent libraries supporting image transformations, and (c) validation tools capable of guaranteeing the algorithm functionality across different platforms.
The development included the selection of state-of-the-art object detection algorithms for face-landmarks and body-pose estimation, and porting them to the Nvidia Jetson NX platform.
In production, real-world AI solutions in healthcare, automotive, retail.
A dedicated team of 30 engineers, scientists and developers passionate about embedded AI.
Software & Hardware
Optimising AI to run on a range of embedded platforms suited for specific requirements.
We are hardware agnostic – the engineers at Darwin Edge will work with the board best adapted to the project at hand.
Our operational prototypes use Nvidia Jetson NX board, TensorRT ML library, OpenCV, C/C++ integration of the models and implementation of the application layer, Nvidia DLA acceleration for Deep Neural Networks.
Embedded AI patient monitoring offers real-time care 24/7 without intruding on the privacy of the individual. Our Human Behaviour Detection platform is already capable of:
- On-device real-time detection of basic human activities (e.g., standing, sitting, lying), including anomalous ones, e.g. falling down.
- Improving the application layer to match the field requirements, using novel techniques and data collection.
The future will see embedded AI allowing the elderly population to remain more independent while reassuring their love ones that they remain safe.