Analysis and Test of External Image Anonymization Algorithm

With the development of intelligent transportation and autonomous driving technology, in-car cameras are becoming more and more common in daily life. The exterior images captured by these cameras may contain sensitive personal information (such as pedestrian faces, license plate numbers, etc.), posing a threat to privacy. In order to protect privacy, anonymization technology is widely used in external vehicle image processing. This study aims to design and implement an off-vehicle image anonymization algorithm to identify the private information in the advanced target detection model and complete the anonymization. The experimental results show that this method has excellent anonymization effect, which is higher than relevant national standards in terms of covering rate, positive number, and has certain practicability.