The Korea Internet & Security Agency (KISA) conducted a joint demonstration project with the Ministry of Science and ICT to verify CCTV capabilities in monitoring various disasters and school safety.
In the field of disaster prevention, MarkAny Inc., an information security company and Goseong County implemented fire and break-in detection features using smart CCTV systems.
These new features go beyond simple fire detection to enable day/night monitoring and fire risk analysis to prevent disasters.
Through machine learning and advanced algorithms, the margin of error for fire detection has been reduced from 9 percent to 0.07 percent. The margin of error for fire risk analysis has been reduced from 25 percent to 0.09 percent.
In the field of school safety, local camera manufacturer Hanwul Multimedia Communication Co. implemented a deep learning-based video analysis solution using CCTV cameras and sensor data at two public schools to demonstrate the detection of school violence and fires.
The new feature detects school violence based on behavioral analysis technology that incorporates various cases of violence such as assaults and break-ins.
The CCTV system automatically detect any case of violence in 10 seconds, raising the accuracy rate from 70 percent to 91.5 percent, the agency explained.