웹 서비스에서 공유되는 이미지 파일에 노출된 중요 개인정보의 자동 검출 및 보호 기법

In today's Internet environment, various multimedia files, especially image files, are shared in real time, and these image files are relatively small compared to video and audio files, and are actively shared with the benefits of ease of expression and meaning. On the other hand, image files containing sensitive personal information, such as identification cards and passports, can be misused in cybercrime if they are inadvertently leaked or shared on the Internet. Laws and policies to protect personal information are in place and are applied, but at the same time, efforts to develop technical protection measures are also needed. Protection techniques proposed in the existing studies have limitations in protecting the personal information contained in images transmitted in real time. For this reason, this study proposes a technique that utilizes Optical Character Recognition and data masking techniques to automatically check and protect personal information exposed in image files intended to be transmitted from the client-side on major Internet service platforms.To verify the proposed technique, we implemented two platforms, Node.js-based SNS chat rooms and online bulletin boards, which provide similar functionality to real-time web services, and confirmed that automatic inspection and protection functionality works well when posted in real time on images containing personal information.