Computer vision has come a long way in recent years. Unlike many other forms of machine learning that require complex analysis, the vast majority of computer vision problems come from simple RGB cameras. Machine learning frameworks such as Keras and OpenCV have standard and high-accuracy neural networks built-in. A few years ago, implementing a facial recognition neural net, for example, was complex and challenging to set up in Python, let alone on a high-speed device using C++ or CUDA. Today, this process is easier and more accessible than ever before. In this chapter, we are going to talk about implementing computer vision in the cloud, as well as on Edge devices such as NVIDIA Jetson Nano.
We will cover the following recipes ...
