South Korea has launched a state-sponsored project to develop an advanced gait analysis system for scientific criminal investigations based on artificial intelligence and deep learning. It comes nearly nine years after related technology was demonstrated by a British expert who played a crucial role in catching a criminal.
In 2013, a criminal ran away after throwing a firebomb at the home of a former intelligence agency chief in Seoul. Investigators failed to identify him due to the poor image quality of closed-circuit television (CCTV) cameras. The police asked for help from Hayden Kelly, a British consultant podiatrist listed in the Guinness Book of World Records as the first to use gait analysis in a court case.
Based on Kelly's analysis, the investigation was completed. Since then, the National Police Agency (NPA) has tried to master Kelly's skills in cooperation with researchers, but technology development has been unsuccessful for various reasons – until recent advances in AI and machine learning.
This week, South Korea’s Ministry of Science and ICT announced that it had endorsed a joint project with police to develop an AI gait analysis system.
To improve the reliability and accuracy of an investigative technique that tracks suspects based on how they walk, the ministry said that CCTV images would be combined with biomechanical data to select and track criminal suspects and missing persons quickly. For commercialization, a university research team will work with SEO, a private company that develops video surveillance systems using technologies such as deep learning.
Gait analysis is a system of scientific analysis used by podiatrists to study how the body moves – or its biomechanics and effects on the feet, hips, lower back, and limbs. During gait analysis, your body's movements are observed, measured, recorded, and assessed.
Gait analysis is commonly used in sports biomechanics to help athletes run more efficiently and identify posture-related or movement-related problems in people with injuries.
Image credit: iStockphoto/vchal