Enabling AI Vision at the Edge
By David Lee, Christopher Moezzi (OpenFive)
Introduction
Computer vision has made tremendous advances in the last several years due to the proliferation of AI technology. The intersection of big data and massive parallel computing changed the way in which machines are programmed to understand unstructured 2D and 3D data, such as video feeds from cameras. Instead of writing a set of rules, programmers use deep convolutional neural network algorithms to enable a machine to generalize a solution from a large, labelled dataset. This new technique, AI vision, has enabled machines to reach super-human capabilities in accurately identifying objects in an image. AI vision makes the camera one of many intelligent sensors.