Q-Cast
PODCAST | Benefiting from Machine Learning
In a joint Assembly/Quality podcast, Jennifer Pierce, multimedia editor with Assembly Magazine and the host of Assembly Audible, and Michelle Bangert, managing editor of Quality, explore an area where assembly and quality intersect: machine vision, with David Dechow, founder and owner of Machine Vision Source. He'll be speaking at the Assembly Show this fall. His session is "Benefiting from Machine Learning in Machine Vision Applications."
Quality/Assembly: So, you'll be speaking at the show this October in Rosemont, Illinois. Can you tell us a little bit about how you decided to speak about machine learning?
David: Well, first of all, I think there's a certain amount of confusion about what machine learning is versus AI versus deep learning. They're all part of the same bucket in general, with machine learning being more a part of computer and data science, but a very close relationship to AI. In many cases, people consider deep learning, which is a lot more familiar, I think, to most people, consider deep learning a part of machine learning. The thing is, it's a really valuable set of algorithms within machine vision that can help do simple classification, simple segmentation, and other tasks without having to go to the full scale convolutional neural networks that are part of the deep learning algorithms. Sounds good.
I know in the past you've mentioned trends are good to follow, but as engineers, you don't want to get too caught up in them and, know, follow the buzzwords or want to use something just because people are talking about it, whatever. So. Absolutely. And when we talk, when we think about as a vision engineers, when we think about machine learning, it's something that doesn't come up too often. So, it's not one of those kind of trendy topics, but again, really, really valuable in certain use cases. And I'm excited to present those options to the audience.
Quality/Assembly: You mentioned machine vision. Is that a sector within machine learning? I'm just trying to learn the proper terminologies.
David: Right, right. And I shouldn't assume everyone has those kind of broad terms under their belt completely. We usually consider, and when I go into the definitions of these things, many people consider machine vision to be a part of, and also an offshoot from the early days of what's called computer vision. Today, many people have picked up that term computer vision as a sort of a catchall for anything associated with using the computer to do analysis.
Machine vision, however, in the early years of AI back in literally the 50s, machine vision was not really a part of computer vision and emerged later into the late 60s. So, we can consider machine vision a part of what the broad, in the broader sense, we consider computer vision. Now the other differentiation is that machine vision most often is recognized for using discrete algorithms like machine learning or many other associated kind of algorithms to extract data from the image. Whereas computer vision in today's marketplace or in today's industrial environment is more considered to be defining using deep learning to extract those features and segment those images.
Listen to the Full Podcast Here:
Looking for a reprint of this article?
From high-res PDFs to custom plaques, order your copy today!
