contador Saltar al contenido

Facebook in the future: facial recognition and automatic tags

Facebook enters the future with new patents for new services. The social network takes photos for us even when the camera is turned off, it will recognize our faces even in the videos to "tag us" and much more. The desire to share and the possibility of inserting tags in posts, videos and photos has made a fortune (…)

Facebook enters the future with new patents for new services.The social network will take pictures for us even when the camera is off, it will recognize our faces even in the videos to "tag us" and much more.

The desire to share and the possibility of inserting tags in posts, videos and photos has made the fortune of the social network. And so – just to give a further boost to its users – the new technologies should allow, shortly, to take one more step and wink at features that only a few years ago would have been part of a spy script.

In 2011 a patent called "was filed, for example"Automatic Photo Capture", Which together with"Preferred images from captured video sequence" is "Image selection from captured video sequence based on social components", Would allow for memorize the various frames of a video shot continuously and – through recognition techniques – link its protagonists, places, but also products to their respective accounts on social networks. An evolved tag that – we read in the description – should respect the privacy settings chosen by the individual (!). The favorite photos taken from the videos can obviously be stored in your mobile device, and it goes without saying, shared.

All three apps date back to October 2011 and were published between April and May 2013 and are owned by Facebook employees Andrea "Boz" Bosworth, David Garcia, and Soleio Cuervo (who now works for Dropbox).

At the technological level these services are based on the fact that the cameras of our smartphones, even when we're not in video mode, can capture video continuously, even between one photo and another. All those frames can be archived, device performance permitting. However, there is no doubt that the performance will increase in terms of software and memory.

The question can only be born spontaneously: what if we make Facebook of all these frames?

Already in the patent description we talk not only about recognizing users but also i places or a product like the drink resting on the background table of a photo. Everything can be tagged and everything can be connected to one or more users. Who spends time with whom, doing what, consuming what …

So Facebook can always suggest us which products and companies we might be interested in! And there will be no need for the end user to put this kind of individual tags, because through image recognition techniques, there may be automatic tags, in pre-set sets.

These patents also have other "exciting" features, such as that of always understand who you are with. Instead of wasting time manually tagging friends or entering where you are or using the future image recognition described above, Facebook could also make all the sensors in your mobile phone and those around you work. Using the sensors GPS can detect a user less than 100 meters from your distance or who is present in a room or an event in which you participate, sending this data to the mobile devices of other connected users.

Not everything is worth being recognized, but Facebook takes care of classifying the most interesting frames for you. To do so, it uses parameters such as your affinity with an object, analyzing how popular it is among your friends and contacts and then extending it to any preference given by the mass audience. This too using the audio contained in the video file.

"The image selection process can analyze the content of the vocal segments (for example, using a speech recognition algorithm) to give an indication of importance (for example, "Say cheese!", "This beauty!", "Incredible!" )?And create a kind of score which also applies to the quality of the frame in the strict sense (for example if the motion sensor has recorded an acceleration that will probably lead to a frame not in focus). At this point, Facebook can collect your favorite sequences in a sort of time-line to share as a news feed. It sounds like science fiction, but an app could really know where, with whom and for what you exclaimed "Wow!". And show it to you with a photo.

In short, there will no longer be a need to look for people, places, or things to "tag". They will already be there, just waiting for approval. And the prize will be to be led gently to one sharing is always easier, which would also lead to receiving more interesting content for each. It does not escape for which huge business implications have these services.

Facebook may have the largest and most promising marketing database, but if you are worried about this, the laziness will you save! If indeed it is true that every day millions and millions of photos, links, various posts are shared, it is equally true that Facebook does not always have such a clear vision of what relationship there is between what is in your account and offline life. And on what advertisers want to do.

The various possibilities to put order with friend lists based on interests or hierarchies, the news feed customization based on interests, they are strongly promoted by the social network, but not used with the same enthusiasm by the majority of its members.

So, waiting for new applications, you know what to do to increase the accuracy of the commercial proposals that will come to you or on the contrary to confuse marketing experts from all over the world. A small but significant free will.