Facebook’s Lumos AI Lets Users Search For Images By Content
Just a year ago, Facebook launched a feature that allows visually challenged mobile users to understand what is in the image through the use of AI (artificial intelligence). Thanks to this development, the company recently discovered that it could make use of the same AI technology to enable users to find photos based on what is in them.
While Google Photos has been offering a similar function in its app for several years now, this is certainly a significant addition for Facebook. Through Facebook’s Lumos AI technology, users can look through photos according to their content (what is seen in them), rather than searching for them according to what they are titled. This however means that all photos must have a sufficient number of tags or captions to be identified.
With the vision to fully comprehend images down to the last pixel, Facebook has pushed computer vision to the next level. According to Facebook’s Director of Applied Machine Learning, Joaquin Quinonero Candela, “This helps our systems do things like recognise what’s in an image, what type of scene it is, if it’s a well-known landmark, and so on. This, in turn, helps us better describe photos for the visually impaired and provide better search results for posts with images and videos”.
Built on top of the system called FBLearner Flow which was designed by Facebook, Lumos was created to understand videos and images. The Lumos platform is consistently improving, as Facebook regularly feeds in new data with detailed labels. This is fed through the interpreted data from the applications built by Facebook’s teams.
While the image recognition software is available to users from the US, there has not been any word so far about when it will make its way to other regions. As time progresses, we shall keep an eye on the engaging development of machine learning. What will be interesting is if Facebook can find greater techniques to implement its machine learning tools to make its application easier to use.
You might also like –
Picture courtesy -tnwcdn.com
by techtalks @TechTalks March 23, 2017 12:28 PM UTC