The Effect of Differents Loss Function in Person Re-identification

Abstract

Person Re-identification (ReID), under the same or different cameras, judges whether two or more pictures are the same person through feature extraction and match of images of pedestrian detection box. In the past period of time, a lot of progress has been made in the field of pedestrian reidentification, and the recognition rate has reached a very high level. One important reason is that people re-examined the loss function, proposed various variants, and the effect has been greatly improved. In this paper, two different loss functions are used for comparison, and the influence of different loss functions on pedestrian re-recognition results is analyzed through the comparison results.