No Need of Retina or Fingerprint Scanner, AI Can Identify People from Their Gait
AI-Based Footstep Recognition System Can Spot People by Their Manner of Walk

It is irritating to stop by every office door just to punch the keycard or scan the fingerprint. Moreover, especially when you are in hurry, these scanners detect some error and you have to perform the procedure all over again. To solve this conundrum, scientists developed an artificial intelligence (AI) that can observe footsteps and identify people by observing their walking pattern. This novel technology is expected to be used at airports as a biometric instead of retina or fingerprint scanners. The AI can perfectly identify a person by analyzing footsteps and real-time data. The best part about this technology is that you will not have to stop at any check points, just walk on the platform and your entry will get registered.

The researchers at the University of Manchester, the U.K. and the University of Madrid, Spain recently published their article in one of the top research journals, IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). The team successfully developed AI-based gait recognition technology that identifies people at almost 100 percent of the times, with just 0.7 error rate. The most commonly used biometrics are fingerprints, retinal scans, and facial recognition. However, the new behavioral biometrics including footstep recognition can capture unique signature based on a person’s manner of walk and movement patterns. Omar Costilla Reyes, a researcher at the University of Manchester, said, “Everyone has about 24 different patterns and movements when walking, which gives an individual a unique walking style and signature. Thus, if these movements carefully monitored, they can be used to identify a person similar to retinal and fingerprint scans.”

The team tested the technology by using large number of so-called “imposters” and small number of users in three real-world scenarios including airport security checkpoint, workplace, and home environment. To create any AI system, computers need to consider ample of possibilities and patterns. the team collected huge amount of footstep database, called as SfootBD, including about 20,000 footstep signals from more than 125 individuals. Costilla Reyes explained that focusing on no-intrusive gait recognition by monitoring the force exerted on the floor during a footstep is very challenging. The most difficult part is to differentiate the subtle variations between two individuals. The gait recognition system uses floor-only sensors and high-resolution cameras. Unlike other biometrics, this technology is non-intrusive and immune to noise and environmental changes. Moreover, a person does not even need to remove footwear as it does not consider footprint shape, but only the manner of walking. After the successful trial for security purposes, Costilla Reyes wants to use this AI for smart steps technology in healthcare sector to recognize neuro degeneration.

It has been few years that scientists want to replace vexing key cards and biometric. Moreover, Swedish employees are implanting microchips under their skin to avoid use constant use of key cards and biometrics. However, instead of turning into cyborg just to get rid of biometric, AI-based gait recognition system seems more convenient.