Twitter : Cropping using saliency
The new system will focus on ‘salient’ regions of an image.
Ever post a photo to Twitter only to have it look odd and incomplete when it comes up in tweet timelines, maybe focused on the wrong area of the image?
They’ve been using face detection to focus the view on the key element when displaying a tweet image. But that process has some limitations – as explained by Twitter:
“While this is not an unreasonable heuristic, the approach has obvious limitations since not all images contain faces. Additionally, our face detector often missed faces and sometimes mistakenly detected faces when there were none. If no faces were found, we would focus the view on the center of the image. This could lead to awkwardly cropped preview images.”
But Twitter’s improving it – here’s a comparison of Twitter’s current (left) and new image cropping system, which uses neural networks to predict the most interesting part of the image.
It’s a small improvement, but an important one – particularly, as noted, as an organization that’s looking to put forward a professional representation of their business.
The new Twitter image updates are currently in the process of being rolled out to everyone on twitter.com, iOS and Android.