An Example of Solution for Data Preparation Required for Some Purposes of People Identification or Re-Identification

Abstract

In this paper, we present an example of solution for obtaining some key elements required for people identification or re-identification. During an identification or re-identification process, different elements are in use. Depending on the method to be implemented, said elements can be RGB (red, green, blue) images, grayscale images, different types of features extracted from them, etc. In this work, we focused on obtaining certain elements suitable for use in the gait recognition methods or re-identification methods that use the obtained features from the images with people in gait. The presented solution can be useful in many applications of identification or re-identification, since key elements required by the implemented methods can be obtained in a simple way. Also, different methods for identification or re-identification can be added in the presented solution. Based on this, we have developed a simple system for people re-identification. An experiment was conducted on our own dataset containing 13 people in gait. The results obtained were over 90% for 4 out of 5 types of features used in the presented re-identification system. It is important to emphasize that more reliable results can be obtained by combining different types of features.