YouTube uses machine learning to improve the appearance of shorts

Machine Learning


As part of a new experiment to clean up your shorts feed, YouTube shorts may get polish after posting.

YouTube has confirmed that it is running a new test aimed at improving the appearance of YouTube shorts clips by using machine learning technology to improve playback.

As explained in YouTube liaison Rene Ritchie:

Just like what modern smartphones do when recording videos, they use traditional machine learning technology for Unblur, Denoise, and run experiments on YouTube shorts of their choice that improve clarity of videos in process. YouTube is always working on how to deliver the best possible video quality and experience, and continues to take author and audience feedback into consideration when iterating and improving these features. ”

Therefore, YouTube is trying to make your shorts look better, even if you're not an expert camera guy.

I think you can make your content look cleaner and make it look more detailed with the shorts feed, while ensuring a better user experience.

But again, some users raised concerns that YouTube's upscaling process makes shorts appear more generated.

As YouTube points out, it will alienate the viewer, but you cannot use AI to modify or change shorts. Therefore, we only apply “traditional” machine learning processes that aim to reduce and improve the presentation of clips.

It's impossible to say that the effects are actually, but if you notice that your shorts look a little different when they go live, this could be the reason for that.

YouTube says this is a small experiment for now and intends to improve the shorts experience for all viewers.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *