AI lightweighting and optimisation technology company Nota said on Tuesday it will participate in the World Security Expo (SECON 2026) to be held at KINTEX in Ilsan from March 18 to 20 and present a demo of its vision-language model (VLM)-based security monitoring solution, Nota Vision Agent (NVA).

NVA is a solution that combines VLM with existing computer vision (CV) to go beyond simple object identification, grasp contextual situations and analyse risk factors early. If a user sets detection rules in natural language, the AI immediately understands the situation and responds. It uses an on-device AI method that processes video data on the device itself, enabling both personal information protection and real-time monitoring. Nota said it recorded average accuracy of more than 90 percent across all items in certification by the Korea Internet & Security Agency (KISA).

The demo at the exhibition covers four types: early detection of fires in server rooms; classification of military and civilian vehicles and analysis of abnormal pedestrian behaviour; analysis of crowd density; and real-time detection of illegal dumping of garbage.

Chae Myung-soo (채명수), Nota's chief executive, said, "Based on technological capabilities that understand on-site situations in natural language, we will expand the market to various industrial sites such as smart buildings, logistics and childcare environments."