A study on the surveillance system using soft biometric information

Abstract

Recently, with the spread of surveillance cameras, high-performance surveillance systems have been attracting much interest in information security and computer vision. On the other hand, there is also a need for simple surveillance systems that use less computer resources, have lower installation and operating costs, and consider the privacy of the people being identified. Moreover, human tracking and identification within and between cameras' views is required in surveillance systems. For this type of simple surveillance system, we propose a human tracking and identification method for within and between cameras' views by using soft biometric information. The results of the simulation using human gait video data confirmed the effectiveness of the proposed method.