Development of a Named Entity Recognition-based Text Masking System for Preventing Personal Information Exposure in Shipping Labels

The widespread use of social media and video platforms has frequently and unintentionally exposed sensitive personal information, such as shipping labels, in shared images and videos. Masking entire shipping labels is not only time-consuming and labor-intensive but it can also lead to video distortion. This paper proposes a system based on named entity recognition and text recognition technologies to automatically detect and mask personal information within shipping labels. Experimental results demonstrate that the system achieved an accuracy of 81.2% in recognizing and masking personal information. The proposed technology selectively masks the relevant personal information, minimizing video distortion, and can be effectively applied to social media platforms and video production environments where there is a high risk of exposing sensitive information.