FBI to use Face Recognition on 13 Million Faces

October 5, 2012 1 comment

In the latest stage of its one Billion USD worth Next Generation Identification Program, the FBI will deploy free face recognition software nation-wide in the US to law enforcement agencies. Using this software, faces can automatically be compared to a database of close to 13 Million faces. The FBI is stressing that these faces are just mug shots and are not taken from other sources such as social media. Full operational capability of the system is expected in 2014. Of course, there are privacy concerns about this system. Being identified at any time in public places is certainly quite a big intrusion into our personal privacy and freedom and should not be taken lightly.

However, I doubt that the system will work well for public surveillance, at least in the beginning. I just returned from Security Essen, the biggest security trade fair in Europe, and tried various face recognition systems myself. It was once again confirmed that face recognition nowadays works very well for access control (that is for “verification” not “identification”). In this scenario the user wants to be recognized, is voluntarily enrolled in the same system he is recognized with and the system knows which face it should expect. Face recognition in other scenarios and in crowds is much harder and requires very good camera positions and high quality shots of a person, who is enrolled (with various pictures of the face). I do not see this technology being successfully deployed anytime soon in public places and areas where unsuspecting civilians should be recognized. The further development of this project should be observed, especially how the FBI will use the system once it is fully operational.

One Million CCTV Cameras for Bangkok

October 2, 2012 Leave a comment

In an ambitious project (to my knowledge the most ambitious world-wide), Thailand will install one million CCTV cameras in Bangkok in the next three years. Amazingly, this project will only cost 2 Million EUR, which would amount to 2 EUR per camera. This is very unrealistic, to say the least. So how does Thailand plan to finance all these cameras? By requiring residents to pay for this extra protection a monthly fee between 2 and 4 Euros per month.

It will be interesting to see how they plan to make use of these cameras. Currently, there are just vague hints about face recognition and video analysis but no details are given. Obviously, such an amount of cameras cannot be managed manually but automatic analysis can only detect clearly defined events and face recognition is far from anything we see in Enemy of the State or CSI. Without automatic analysis of the videos these cameras are as good as dummy cams. I guess at least privacy won’t be a big issue for them.

Video Surveillance Trend 2012: Smartphone Surveillance

July 23, 2012 Leave a comment

(c) heraldsun.com.au

Mobile apps for video surveillance applications (“Video Surveillance Management” software) exist for some time and are usually add-ons to video surveillance products. They basically allow viewing video streams on a mobile device. However, the next trend in video surveillance seems to be using smartphones as video surveillance cameras. Obvious applications are as spy-cams or as baby monitors. However, additionally, this further paves the way for broader use of sousveillance (“inverse surveillance”) technology, where everyone watches everyone, thus actually creating more privacy than less (as argued by David Brin). This is certainly an area of technology to watch.

Check our this really well-done app which can be used for many applications (such as letting your girlfriend choose which tea she would like while lying sick in bed): Airbeam.

Are You a Sex Predator? Think Before Your Write!

July 19, 2012 Leave a comment

(c) Corner Stock Baby Gifts.
You can get the t-shirt from here.

A recent Reuters story about an interview with Facebook’s Chief Security Officer Joe Sullivan revealed that Facebook is scanning user profiles for criminal behavior, focusing on sexual predators. By comparing several parameters like friend status, age, mutual friends and relationship between users, a monitoring software determines how likely it is that a Facebook user is a sexual predator. If there is a positive match, a Facebook employee gets a warning and checks the information manually. If the monitoring result seems likely, the police is informed. This actually led to the arrest of a thirty-something man who talked with a 13 year old girl on Facebook about sex and planned to meet her the next day. Because of the fast reaction of Facebook this man was arrested before anything else could happen.

This raises the obvious question if it is ok for Facebook to scan our data. No one knows if the age of a Facebook user is correct. Maybe the girl was actually a 60 year old, fat man … On the other hand finding criminals before they do any (more) harm cannot be a bad thing either.

I, for one, do not want to think if what I write might look criminally relevant to somebody before I post something on Facebook! I hope as much is done to educate 13 year olds on not doing something stupid as to monitor if they do.

Twitter Gov Requests Doubled in 2012

July 16, 2012 1 comment

Twitter recently released their first transparency report, outlining how often in the first half of 2012 government or copyright holders requested Twitter account information and how often this information was produced. The majority of information requests (679) came from the United States but also a significant number came from Japan (98). US requests were followed in 75% of the cases while in Japans case only 20% of the requests were fulfilled. Interestingly, only 3 requests to remove a Twitter account by court orders were received (Greece and Turkey) but none of them was followed!

In total, Twitter received in the first half of 2012 as many requests as in all of 2011, which is a much bigger increase than overall Twitter growth (which was at about 20% in the US).

All in all, these numbers do not surprise me that much, taking all of the 140 million active users into account. And it is reassuring that Twitter does not seem eager to give out user data (Twitter already took a stand for an Occupy Wall Street protester at the beginning of this year).

Twitters transparency report is a perfect example on how to build users trust: by making the companies actions transparent. They should be a glowing example for other web companies who basically store all the information of our lives online.

If you are interested in what Facebook sends if they get a subpoena for a user, you can see an example online (it’s 62 pages of Facebook data …).

Flying drones prone to hacking

July 2, 2012 Leave a comment

MQ-9_Reaper_-_090609-F-0000M-777The US department of homeland security was offering a prize for whoever can hack into a flying drone. Now, a team from the University of Texas actually achieved it by spoofing GPS signals. This way, a drone can literally be steered anywhere the hacker likes and could be crashed into a building like a missile. This is an obvious security threat to drones today, more so if they are used in urban environments. While this simple “hack” can probably be fixed, it is a warning sign that these threats have to be considered when developing and deploying this technology.

“Do Not Track” Not So Good After All?

June 12, 2012 Leave a comment

Source: Slashgear

The “Do Not Track” header in websites is a feature that states if a user wishes to be tracked by websites (mainly for advertising purposes through cookies) or not. However, it is optional for the websites if they respect the user’s decision or not. Today, most browsers support this feature (the Chrome browser will support it by the end of 2012), Microsoft recently even announced that it will be turned on by default in IE10. From a privacy perspective this is a very welcome development, which gives power back to the users. However, two recent articles focused on the economical implications of restricting technology that funds big parts of our (free) Internet as we know it. Without ads, websites such as Google or Facebook would have a hard time financing themselves. In Technology Review, Antonio Regalado asks if this feature will kill off innovation in online advertisement, with serious implications for the $40B online ad industry and as such for us as users as well.

Another reason I find the article quite interesting is that it points out the positive sides of tracking the user to deliver highly targeted ads. You might even get relevant information out of ads instead of useless spam.

In order to better understand online tracking I highly recommend the guide from the Guardian. Also check out their nice graph about the biggest advertising companies and websites that use them.

%d bloggers like this: