Several efforts by scientists to use gait analysis as a biometric have gone in vain, until now.
The Chinese scientists have come up with a new biometric identification technology that negates the need to look at people’s faces in order to recognize them. Rather, it uses the exclusive manner with which different people walk to identify them.
The Gait Recognition System
According to researchers from the Automation Institute of the Chinese Academy of Sciences (CAS), the gait recognition system works by utilizing a camera that is capable of identifying people within a distance of 50 meters.
The camera works by analyzing the way people carry themselves.
One leading technique known as the gait energy image uses video images of a person to create a blurred silhouette that is characteristic of their gait. The silhouette then allows a human operator to identify a person’s identity and allows the system to automatically detect that person when they next appear on film.
Martin Hofmann and colleagues at the Technical University of Munich in Germany have developed a technology that does not just use a blank silhouette but also uses information such as the shadow of a person’s clothing to provide a more detailed signature.
Hofmann also used Microsoft’s gaming sensor Kinect, to measure depth that allows him to separate the target from the background.
The resulting system is capable of identifying a person with variations such as him/her carrying a briefcase and provides results with 80 percent accuracy.
Daigo Muramatsu and colleagues at Osaka University in Japan are working on the problem of identifying a person captured from different angles. The preliminary tests revealed low error rates and Muramatsu describes the results as “promising”.
The Technology Comes With Advantages
- According to an associate researcher from CAS- Huang Yongzhen, contrary to the iris recognition system and the face recognition system which operate at much shorter ranges, the newly developed system operates at a much longer range.
- While the iris system requires its target to be no more than 30cm away, and the face recognition system is capable of working within five meters, the new system allows its target to be as far as 50 meters
- Iris scans and face recognition require reasonably high-quality images, a person’s gait can be recognized from low-quality CCTV footage.
- It takes between 300 and 400 milliseconds for the human eye to blink, however, the detection time for the system is even less than that – it takes 200 milliseconds to detect a person making it faster than the blink of an eye.
- The gait recognition system is capable of identifying a person with minimal effort from the user’s side. It can identify a person even if they walk past the surveillance camera hiding their face.
- The recognition system integrates the real-time people counting technology to estimate the density of a large population. It can detect up to 1,000 people within an area of 1,000 square meters.
- Among other fields, the technique can be utilized to improvise security as well as in the field of public transport.
The CAS said that China is currently at the forefront of gait recognition technology worldwide. Still, let’s hope the technology is able to identify an injured person’s limp and doesn’t raise alarms just because someone can’t walk straight.