Microsoft is notching up its Azure Media services platform by a couple of notches. The company is now going to implement its machine learning tools into its collection of cloud-based tools for video workflows.
Now, you may wonder at the apparent non-existence of a relation between videos and machine learning. Machine learning after all, is used for data analysis. It can’t be used with videos, right? Wrong.
Apparently, Microsoft’s integration will let its Media Services automatically select the most interesting parts from a source video — Basically, a summary of what the whole thing is about.
And hold your horses, because that’s not even the most interesting part yet. The company is also building face detection into its video tools so that you could also recognize how people are reacting to your speech — or whatever else your video is about. And yes, if you’re thinking that the timing is so very perfect — considering that Facebook launched a similar feature just yesterday, yup you are right.
So basically, you would be able to recognize how your audience feels about your video. If their experience goes from indifferent to happy, then you are rolling. And if it goes the other way around, well you know you need to improve stuff.
Also being integrated into media services is Hyperlapse. The technology can be used for stabilizing and time-lapsing videos, and their integration means that users will be able to edit stuff, right from Azure. The 10,000 frames limit for videos that could be Hyperlapsed, has also been removed.
Finally, Optical Character Recognition, which lets computers directly recognize text without any human intervention, has been integrated with the Azure Media Suite. As for its uses, well, a potential example may be how search results could display videos that had your search terms in its subtitles.
The addition will also lead to a greater discoverability and accessibility of media through media intelligence. So yes, good news all around. Meanwhile, the integration gives us yet another look at the vast potential of using machine learning with virtually everything.