Twitter Improves Image Quality via Machine Learning with Magic Pony Purchase
The truth of it isn't anywhere near as Grimm-esque, or amusing, but it's certainly interesting, and kind of weird. Magic Pony is a London based start up dedicated to machine learning technology. In particular, their focus is improving low quality images and videos in real time though pattern recognition. The results are, at least in the case of the example images, a little creepy, but given that before this the only way to achieve the same result was hours of painstaking edge sharpening and colour fiddling on Photoshop, it's very impressive.
It can do more than just sharpening up a previously pixelated image, though. The pattern recognition enables the system to replicate or even expand images in higher quality, and it can do the same for GIFs and video. The more images it is shown, the more complexity it can apply, as is the way with machine learning, so theoretically this system will only become more sophisticated with time.
The real trick of the process is using information that isn't necessarily within the picture itself. Instead, it creates composites from collections of other, smaller images, which is comparable to the way the human eye works.
As far as Twitter are concerned, the aim here is likely to better optimise image and video posting for their mobile platforms, as quality is often lost in the transition. More than that though, working with Magic Pony will put Twitter across the finish line first with regards to this kind of technology, which will force other companies to deal with them if they want access to it.
Callum is a film school graduate who is now making a name for himself as a journalist and content writer. His vices include flat whites and 90s hip-hop. Follow him @Songbird_Callum
Twitter Improves Image Quality via Machine Learning with Magic Pony Purchase Reviewed by Callum Davies on Wednesday, June 22, 2016 Rating: