Deep Learning is a data-driven process that builds an algorithm based on good and bad example images of an inspection point that is so complex a human could not do it themselves. The algorithm gets better with more data and more time. Inspection cameras can then catch subtle and varied mistakes in numerous situations. It can also handle far more complex visuals than a classical approach. Deep learning systems can look for everything at once, rather than looking for one specific type of failure.