Manufacturers performing visual inspections on their products have historically had two options: perform the inspection manually, or program a rule-based vision system. Both of these options are slow, inflexible, and costly to maintain. However, as Graphical Processing Units (GPUs) have become more powerful, a third approach has become available: deep learning.
Deep learning-based vision systems learn directly from annotated images to perform inference on images. They offer numerous advantages including rapid deployment, greater detection flexibility, and lower development and acquisition costs. They are also easily retrained and reconfigured with minimal to no programming. As such, they represent a massive opportunity to automate quality inspection for small to medium sized manufacturers.
In this presentation, we will discuss the technology that enables AI-based computer vision, the advantages they provide, use cases, and how to best implement them. Attendees will walk away with a clear understanding of where deep learning fits into their operation and how to get started with the right tools and strategy.
