Is this machine vision?

I have a pretty tight definition of what constitutes machine vision: it’s the use of image acquisition, processing and analysis tools for automated decision-making in industrial environments. I realize this means that some of the computer vision applications I cover here – security, autonomous vehicles and so on – are not really “machine vision,” but I’m Okay with that. Machine vision is after all, a subset of computer vision.

But what about analyzing video content? Is that machine vision? California start-up Anvato believes so. Anvato has developed tools for analyzing video in order to detect copyright infringement. As I understand it, and the information on their website is pretty sketchy, they produce a kind of digital signature of a series of frames, and can compare this against the signature extracted from another video sequence.

Presumably, and I’m just guessing here, the software indicates a degree of match for a selected video, which could be used as the basis for litigation. It’s an interesting idea that seems to open the way for automated detection of copyright violations. In an age when companies are starting to recognize the value of posting information on YouTube, that has to be a pretty big deal.

But is it machine vision?