In terms of computers, the edge is the very end device that sees things and measures parameters. Deep learning on edge devices implies injecting AI into the edge device so that along with seeing, it can also analyze an image and report its content. An example of an edge device for computer vision is a camera. Edge computing makes image recognition on-premises quick and efficient. The AI component inside a camera consists of a powerful yet tiny processor that has deep learning capabilities.
This AI on the edge can perform a mix of three separate functions, depending on the choice of hardware and software platforms you use:
- Hardware acceleration to make the device run faster
- Software optimization to reduce the model size and remove unnecessary components
- Interacting with the cloud to batch process image and tensors
The benefit of this...
