CC photo by A200Wells on Flickr
Recently on Spark, we’ve done a number of stories all about how trackable you are. For instance, a little while ago, we talked about facial recognition technology in the wild, and how surprisingly accurate it can be. It’s all a little bit creepy. But here at Spark, we like to get a little creepy once in a while.
Recently, we talked to someone who’s working on another way for computers to identify you in the real world: by the way you walk. Todd Pataky is an assistant professor in the Department of Bioengineering at Shinshu University, and he recently published a study in which he correctly identified individuals by their gait with 99.6% accuracy.
You can hear part of Spark’s interview with Todd below, or download the MP3. [runs 4:28]
Play audio:
If you like hearing these extended interviews, why not subscribe to Spark Plus? You’ll get regular weekly episodes, plus additional blog-only content like this. [Subscribe via RSS] or [Subscribe with iTunes]
And why do we need this….or are researching this?
Do we research Anything that is not Police State related ???
Sheeesh
Hi, I'm the first author of the paper. My intentions were neither police- nor security-related, I simply sought to find out how differently different people actually walk.
We're researching this topic because: (1) human gait research is widely relevant: prosthetic limb development, rehabilitation following injury, sports performance optimization, etc, and (2) we need to know exactly how different walking can be across people; if we want to create a prosthetic leg, for example, we need to know whether its design will be applicable to one person, to ten similar people, to all adults in one country, to all children around the world, etc.
I'm afraid that the 99.6% accuracy result has been misinterpreted as a police-state statistic; in fact it can also mean that different people walk very differently, and we need to consider this carefully when extrapolating our results beyond isolated experiments.
Cheers, Todd.
HELLZ YEAH GEE
What does it matter. In a generation we'll all have some form of high tech barcode that will make gps tracking look like childsplay. It will start with the safety nazis who will sell it as the untilmate protection for kids. Then someone will figure out how to see it as a general convience. Then…. well, George Orwell is already rolling in his grave
whatever
Everyday I'm shuffshufflin.
I like the idea of this being used to identify potentially dangerous people at places like airports. But I wonder if someones gait remains the same after some alcohol or medications. If so, it would be pretty easy to fool it then.
Hi, I'm the first author of the paper. While we weren't really focussed on security (please also see my response to Gee above), if this type of analysis were to be used by others for security purposes I think it would only work in cases where people WANT to be identified. For example: ATMs, automated airport customs, controlled building access. In these cases it wouldn't be to your advantage to be drunk or otherwise intoxicated.