ID3's computer vision research

How does Facebook suggest tags?

When someone uploads a photo of you, we might suggest that they tag you in it. We're able to compare your friend's photos to information we've put together from your profile pictures and the other photos you're tagged in. If this feature is turned on for you, you can choose whether or not we suggest your name when people upload photos of you. Adjust this in your Timeline and Tagging settings.

We currently use facial recognition software that uses an algorithm to calculate a unique number (“template”) based on someone’s facial features, like the distance between the eyes, nose and ears. This template is based on your profile pictures and photos you’ve been tagged in on Facebook. We use these templates to help you tag photos by suggesting tags of your friends. If you remove a tag from a photo, that photo is not used to create the template for person whose tag was removed. We also couldn’t use a template to recreate an image of you.

If we can’t suggest a name automatically, we'll group similar photos together so you can tag them quickly.

What information does Facebook use to tell that a photo looks like me and to suggest that friends tag me?

Two types of information are required to see that a newly uploaded photo looks like someone who’s been tagged on Facebook before. The types of information required are:

- Information about photos you’re tagged in. When you’re tagged in a photo, or make a photo your profile picture, we associate the tags with your account, compare what these photos have in common and store a summary of this comparison. If you’ve never been tagged in a photo on Facebook or have untagged yourself in all photos of you on Facebook, then we do not have this summary information for you.

Facebook has a facial recognition research project called as DeepFace. DeepFace, is now very nearly as accurate as the human brain. DeepFace can look at two photos, and irrespective of lighting or angle, can say with 97.35% accuracy whether the photos contain the same face.

The DeepFace software, developed by the Facebook AI research group in Menlo Park, California, is underpinned by an advanced deep learning neural network. A neural network, as you may already know, is a piece of software that simulates a (very basic) approximation of how real neurons work. Deep learning is one of many methods of performing machine learning; basically, it looks at a huge body of data (for example, human faces) and tries to develop a high-level abstraction (of a human face) by looking for recurring patterns (cheeks, eyebrow, etc). In this case, DeepFace consists of a bunch of neurons nine layers deep, and then a learning process that sees the creation of 120 million connections (synapses) between those neurons, based on a corpus of four million photos of faces.

Ben Innes on taking a 'selfie' with the EgyptAir hijacker

The 26-year-old British health and safety auditor describes why he posed for picture while held hostage on flight MS181

A British man who posed smiling for a photo with a hijacker on board an EgyptAir flight has said he did it to take “a closer look” at the apparent explosives belt, adding: “I figured if his bomb was real I’d nothing to lose anyway.” Ben Innes, 26, was one of three passengers and four crew held to the end of a hostage siege, after Seif Eldin Mustafa hijacked the EgyptAir flight bound for Cairo from Alexandria and forced it to be redirected to Cyprus. The plane landed at Larnaca airport, where Innes was snapped standing next to Mustafa – who was still wearing what he claimed to be an explosives belt – in the cabin of the Airbus 320. The bomb was later found to be fake. Innes is a health and safety auditor from Leeds, living in Aberdeen. He was returning home on a business trip when flight MS181 was seized. He spoke to the Sun about his motivations. “I’m not sure why I did it, I just threw caution to the wind while trying to stay cheerful in the face of adversity. I figured if his bomb was real I’d nothing to lose anyway, so took a chance to get a closer look at it. “I got one of the cabin crew to translate for me and asked him if I could do a selfie with him. He just shrugged OK, so I stood by him and smiled for the camera while a stewardess did the snap. It has to be the best selfie ever.” Innes was among the last passengers to be released by Mustafa, after the hijacker freed most of those on board. “After about half an hour at Larnaca I asked for a photo with him as we were sitting around waiting. I thought, why not? If he blows us all up it won’t matter anyway.” On closer inspection, Innes suspected that Mustafa’s explosives device was likely to be fake. “So I decided to go back to my seat and plot my next move.” He and the remaining hostages were later released. Innes’ stunt was decried by security experts, praised by relatives, and said to be “totally in character” by friends. “Only Ben could get a selfie! #proud,” reportedly tweeted Sarah Innes, a relative. Her account was later deleted.