A DEEP LEARNING-BASED ABNORMAL BEHAVIOR DETECTION SYSTEM THAT DETECTS THE BEHAVIOR OF A CAMERA INSTALLATION IN A SAFE TOILET USING DE-IDENTIFIED DATA

摘要

The present invention relates to a deep learning-based abnormal behavior detection system that detects a hidden camera installation behavior in a safe toilet using de-identification data. The present invention is image information de-identified using a de-identification image information generating unit that captures the subject's behavior in the bathroom and generates de-identified image information data of the subject's personal information and a learning model for abnormal behavior It includes an abnormal behavior discrimination unit that classifies data into normal behavior data or abnormal behavior data.