Facebook patents for emotion-detecting selfie filters

Seems like Facebook is seeking to improve itself in unique ways as it has filed a patent for ‘emotion-detecting’ se
29 Jun, 2018

Seems like Facebook is seeking to improve itself in unique ways as it has filed a patent for ‘emotion-detecting’ selfie filters.

A recent public patent demonstrates that the popular social media app Facebook is looking to launch a tech that can detect emotions and automatically choose an animated selfie based on the particular emotional state.

The patent was originally filed back in 2016, but was made public on Thursday June 28 this year. The patent describes a system for ‘identifying an emotion and selecting, based on the emotion, a mask from a set of masks’. The term ‘mask’ refers to Facebook’s preferred word for the selfie filters that adds animations to users’ face, reported Mashable.

Currently, users have to select a filter manually that fits their mood, but with this new emotion-detecting technology, the masks will be selected automatically based on the image it detects. The firm explains that its machine learning systems can determine the emotions based on the users’ facial features.

Facebook’s new feature will show how much time users spend on the app

Facebook’s patent explains, “If the app detects happiness, for example, it could bring up a mask named ‘happy panda’, [While] the emotion ‘surprise’ [maps] to a mask named ‘surprised eyes’, the emotion ‘anger’ to a mask named ‘angry bird’, and the emotion ‘sadness’ to a mask named ‘gushing tears’.”

However, the patent also explains that there are other kinds of image recognition too and the mask can also be changed based on other factors like users’ location, profile data or even the contents of an image itself.

“For example, if a user is at a zoo looking at a panda and a digital photograph is taken of his or her face having a happy expression, then a happy panda face mask may be selected for the user based on the user’s happy expression and the input image of the panda detected by a camera (e.g., a camera on the user's smartphone) in the background behind the user’s face.”

“If the input image depicts a heart-shape, such as that made by two hands touching at the fingertips and palms, with the fingertips below the knuckles, then the emotion ‘likes’ or ‘feeling loved’ may be identified,” the company wrote.

However, it’s early to get all excited at this point since it is just a patent as of now and it doesn’t indicate that the feature will come to reality or if it would be a consumer product.

Copyright Business Recorder, 2018

Read Comments