Snapshots of daily life help researcher build up digital memory database

Dr Cathal Gurrin wears a camera around his neck. All day. And that camera automatically takes pictures several times a minute. All day.

This has been going on for nine years, and he has collected a database of more than 16 million images of the things and places and people he sees as he goes about his daily life.

It could be the most mind-numbing Instagram account ever, but Gurrin is not really interested in looking at the photos themselves. Instead he is building software that can extract and analyse useful information from the images, which he sees as the basis of a “digital memory”.

“Many years ago, nobody was that interested in what I was doing, or maybe they thought I was a bit crazy to be doing it,” says Gurrin, who is a researcher with the Insight Centre for Data Analytics at Dublin City University. “But now more people are asking about what I am doing, and about the research we can do with the database of images.”

The practice of capturing all that information on everyday activities is known as “life-logging”, which Gurrin describes as “using sensors, devices and software to understand activities and create new ways to improve your life”.

“Many people buy things such as devices to count how many steps you take each day or how active you are, and that is a form of life-logging because based on that you can make an effort to keep being active and that improves your fitness,” says Gurrin. “But we want to take it a step further, we want to record as much of ordinary everyday life activity as we can, and use this to build software versions of your memories.”

Banking information

The idea is that by banking information and building software to search and analyse it, specific details from your past can be called up to trigger your recollections of it, he explains. But one of the big challenges is getting the software to recognise objects; humans can recognise a plate of food or a bottle of cola pretty easily, but computers need to “learn” and process those objects in images.

Artificial intelligence systems can now recognise thousands of objects such as a computer, a cup of coffee or a tree, in pictures.

“We can pull out these things with varying degrees of accuracy from the data, the pictures, and that becomes our useful information,” says Gurrin.

‘Walk to work’

“So the software might be able to tell you that you spent six hours in front of the computer today, and you had five cups of coffee, and you saw trees outside on your walk to work.”

Hang on, though; what about privacy? And surely there are some events and activities that are better left unrecorded?

“If people or events are in a public place, then there is no privacy issue, but if I think I should switch the camera off out of respect for myself or for other people, then I do,” he says. “And if I am in someone’s home, which is not a public place, I ask if it is okay to leave the camera on.

“In nine years, I have been asked only three times to switch the camera off.”

I quietly think about how, because of this interview, I am now logged in his image database.

“We don’t sit down and go through all the pictures one by one, the software analyses them,” Gurrin reassures.

“I don’t even view them as pictures, I view them as data, much in the same way as I view step-count data.”

So what can they do with this huge mountain of data? Gurrin’s team at DCU is currently working on spotting brand names that crop up in a person’s environment, and how this might affect their food and drink choices.

They are also continuing to refine the software to build that digital memory, an approach they believe could ultimately be used to help support people with brain conditions that affect memory.

Start small

And while Gurrin is one of the more extreme life-loggers on the planet, he offers ways to start gathering information about aspects of your life.

“Start small,” he says. “Ask for an activity tracker device for Christmas, and you can start looking at the data about how much you are moving and exercising.”