Eunpyeong-gu Introduces Smart Deep Learning Selection Surveillance System

Real-time Selection and Monitoring of Suspicious Videos Using AI-based CCTV Analysis... Selected for '2020 Smart City Service Demonstration Support Project' (Seoul Digital Foundation), Secured 130 Million KRW

[Asia Economy Reporter Park Jong-il] The ‘Smart Deep Learning Selective Surveillance System’ is an intelligent service that utilizes visual artificial intelligence-based video analysis technology to recognize objects such as people and vehicles appearing in CCTV footage, analyze and select movement events, and prioritize displaying CCTV footage expected to be dangerous situations on the surveillance screen.

Unlike the 1st and 2nd generation intelligent surveillance systems that extract objects from CCTV footage, compare them with pre-input patterns, and obtain results, this project applies ‘deep learning’ technology where the AI system learns autonomously. The goal is to increase the object recognition rate to over 80% to achieve accurate selective surveillance.

Eunpyeong-gu was selected for the ‘2020 Smart City Service Demonstration Support Project’ announced by the Seoul Digital Foundation and secured a project budget of 130 million KRW to build the ‘Smart Deep Learning Selective Surveillance System.’

This project, which requires a total budget of 620 million KRW including 490 million KRW already allocated from Eunpyeong-gu’s budget for building an intelligent CCTV system, is planned to be applied first to 273 CCTV locations within Eunpyeong-gu that are highly vulnerable to crime.

By selectively displaying only problematic CCTV footage identified by the artificial intelligence (AI) system from all CCTV footage, the collaborative system where CCTV monitoring personnel verify the footage can resolve surveillance gaps caused by sequential visual monitoring, reduce fatigue, and improve concentration of the monitoring personnel.

It was also announced that the displayed problematic CCTV footage will be linked with the ‘Eunpyeong-gu Smart City Integrated Platform Service’ and provided in real-time to police stations and fire stations, thereby establishing a two-way situation response system.

The ‘Smart Deep Learning Selective Surveillance System’ is scheduled to be completed by November this year, after which the results will be evaluated and supplemented, and a decision will be made on whether to expand its application to all CCTV operated by Eunpyeong-gu.

Kim Mi-kyung, Mayor of Eunpyeong-gu, said, “Through this project applying advanced information technologies such as artificial intelligence and big data to the CCTV surveillance system, we will come closer to truly real-time surveillance,” and added, “We will continue to integrate smart city services into various fields to ensure that all residents can enjoy a happy and safe life.”