Indoor technology for detecting large crowds and analyzing behavior. Courtesy of LSK iRobot.

Indoor technology for detecting large crowds and analyzing behavior. Courtesy of LSK iRobot.

View original image

LSK iRobot, a company specializing in AI and robotics, is expanding its optical vision artificial intelligence (AI) technology, previously used in manufacturing processes, into the security sector. The company plans to develop solutions that detect devices difficult to identify with existing surveillance systems, such as AI smart glasses and ultra-small cameras, and target the markets for exam venues, public institutions, and high-security facilities.


LSK iRobot said on Nov. 19 that it is working with its subsidiary OneTouch AI to expand its optical-based vision AI technology into the detection of illicit recording and disguised devices at exam venues and high-security facilities.


Demand for related detection technologies is growing as cheating with AI smart glasses and ultra-small cameras has emerged as a problem in the College Scholastic Ability Test and national qualification exams. AI smart glasses that look similar to ordinary glasses and button-sized cameras that can be attached to shirts are difficult for proctors to distinguish with the naked eye alone. Radio-frequency detectors may also have difficulty detecting devices that do not use communication functions or that use information stored on the devices themselves.


To address these problems, LSK iRobot plans to use OneTouch AI's optical-based computer vision technology. The company aims to develop technology that identifies ordinary glasses and wearable devices with built-in cameras by using special lighting and 2.5D three-dimensional shape analysis to detect subtle surface contours and differences in the optical properties of objects.


The company will also apply vision AI technology used in existing manufacturing processes. According to LSK iRobot, its industrial vision AI system achieved an inspection speed of about 0.5 seconds and a classification accuracy of 98% in an ultra-high-speed inspection environment. LSK iRobot plans to apply this inspection technology and its anomaly-detection algorithms to security, establishing a technological foundation for identifying even atypical appearances not defined in advance and subtle signs of anomalies.


However, the inspection performance achieved in manufacturing processes will not necessarily be replicated in the security sector. The company plans to further develop detection performance and systems tailored to security environments, then refine them through technical validation.


The initial targets are devices requiring optical identification, such as AI smart glasses and ultra-small cameras. The company plans to expand the scope of application to exam venues, national qualification exam sites, public institutions, and industrial facilities. Its strategy is to target both the business-to-business (B2B) and government and public-sector (B2G) markets, with applications including contactless device inspections at entry points and the detection of illicit recording devices at high-security facilities.



"As wearable devices incorporating AI become smaller and more intelligent, security blind spots that are difficult to address through conventional visual inspections or simple radio-frequency detection alone are expanding," said Choi Jongwon, CEO of OneTouch AI and a professor at Chung-Ang University. "We will apply optical vision AI and 2.5D shape-analysis technologies developed through our work in manufacturing, along with technology that identifies test takers displaying abnormal behavior on closed-circuit television (CCTV), to the security sector. We will continue developing visual-intelligence-based detection technology to address emerging threats."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing