David: Well, first of all, great show. I'm really enjoying it here today. My session was about machine vision systems integration and really the idea of how do we achieve success with the many, many very exciting machine vision technologies that are in the marketplace. And the premise of the talk, and I've discussed this many times over many years, is that in order to achieve integration success for machine vision systems in an automated environment or quality environment, we need to follow a process. We need to follow something that is a specified group of steps that lead us to a result that will give us the best possible solution for our desired inspection, our desired machine vision application. And we had a good time going into a lot of those details and a very good crowd at the Learning Theater.

Quality: I thought it was a good session, covered a lot. You said there's some kind of misconceptions with machine vision and maybe that term is even kind of incorrect?

David: Yeah. There are, I believe there are, there have been, over the years, a lot of misconceptions about vision. We hear marketing claims by some companies about how machine vision system really doesn't work, so you have to use this. Or machine vision is an outdated technology. It is not.02:04course, it's a growing technology, a thriving technology, and has been used for decades in automation, extremely successfully. One of the big confusion, I find in the marketplace in the last about 10 to15 years is the difference between machine vision and computer vision/AI.

Frankly, those two terms, computer vision and AI, are not necessarily the same thing either. But computer vision is not machine vision. And computer vision is really historically and technically more of a science that utilizes different tools, primarily in today's marketplace of deep learning, to analyze images, classification, defect analysis, and even segmentation. Machine vision, on the other hand, is more of an engineering discipline. It's the ability to put together imaging tools, imaging components, and specific software tools to extract data from an image and from the pixel level data in the image. It might learn, to be fair, it might use and learn using deep learning, but machine vision is more of a broaderapplic ation base within the automation environment. Definitely. That's a good explanation and clarification for people who might be confused. I also thought that it was a great point you made about picking what the application needs instead of just wanting to use AI, for example.