Request a call

AI to Recognise Gait for Non-Intrusive Security Scan

Last updated on May 29, 2018 by Dotsquares

Researchers at the University of Manchester and University of Madrid, have developed a new AI system that recognises people from their walk. They intend to use it for more accurate, yet non-intrusive, security scans at critical places like Airports and concerts.

The system, as reported by the Financial Express, is incredibly accurate. It correctly identifies individuals almost a hundred percent of the times and has the error rate of mere 0.7. To create such a meticulous system, the researchers have created the largest ever database of footsteps to train the algorithm on different movement patterns. For this purpose, the researchers have collected the data of 20,000 footsteps from 127 people, by using tools like high-definition cameras and floor-only sensors.

The collected dataset is named SfootBD and it was fed to the automatic footprint biometric verification model, which was recently presented in the IEEE Transactions on Pattern Analysis and Machine Intelligence.

A researcher from the University of Manchester, Omar Costilla Reyes has this to say about the new system, “Focusing on non-intrusive gait recognition by monitoring the force exerted on the floor during a footstep is very challenging. That’s because distinguishing between the subtle variations from person to person is extremely difficult to define manually, that is why we had to come up with a novel AI system to solve this challenge from a new perspective.”

The system has already been tested in three real-world scenarios such as in airports, the workplace and home, wherein a few individual users were placed amongst a large number of impostors. The system then created the aforementioned results with unbelievable accuracy leading the Reyes to make this statement, “…every individual person has a unique, singular walking pattern. Therefore monitoring these movements can be used, like a fingerprint or retinal scan, to recognise and clearly identify or verify an individual.

Indeed, if these metrics work as good as the present security scans that primarily use biometrics, it will be a revolutionary step in redefining the security procedures in crucial places. The new systems will be non-intrusive and will even work in noisy environments. Although the research has maximum potential in the revolution of security systems, its impact on other sectors like healthcare should not be overlooked. The positive impacts on healthcare include helping to more efficiently and effectively diagnose conditions such as neurodegeneration.

Resources:

https://www.financialexpress.com/industry/technology/new-ai-system-can-identify-people-from-their-walk/1183066/