AI Emotion Detection

About

AI Emotion Detection is a series of images that were put through Microsoft's experimental Project Oxford, which was touted as a way to be able to use AI to tell what emotions are being expressed within a picture. The project saw a lot of old memes, as well as original pictures, used to try and tell what emotions were being given off, before the project was officially scrapped, causing the remnants to be the last of their kind.

Origin

At Microsoft's Build2015 press conference the company showed off the next step in AI image vision technology, showcased by having Project Oxford be able to tell if two faces looked similar to each other, or compile together a list of similar looking faces. This was revealed early in March 2015, but it wasn't until November 13th, 2015, that people really started to pay attention. On November 13th, Microsoft added emotion detection, meaning that the AI would attempt to read the emotions presented on the face of a picture, and would then give an emotional breakdown of the face, with an early example of this in action^[1] being Success Kid, whom Microsoft labeled as being more angry than happy or successful (shown below).

Spread

On November 19th, 2015, a now deleted redditor^[2] uploaded an image of Sheldon from The Big Bang Theory with a grin on his face. The emotion panel to the right indicating his happiness mixed with a small tinge of surprise were accurate enough for Redditors to give the post over 100 upvotes in seven years (shown below).

The idea of taking existing memes and throwing them through the emotion detection was the main brunt of exploration of this technology, with a Me.Me^[3] post uploaded on February 24th, 2017, by the site admins shows Michael Rosen in the middle of one of his famous sketches, paused to show his face in the middle of an expression that the AI detected to mean 99.9% happiness (shown below).

Various Examples

Search Interest

External References