Person re-identification (ReID), a pivotal task in computer vision, involves matching a specific individual across various surveillance cameras that do not share overlapping fields of view (Ye et al. 2021; Chen et al. 2021; Fan et al. 2019). Propelled by the advancements in deep learning and the growing demand for smart video surveillance, ReID has garnered significant attention within the research community. Despite the considerable progress made by deep neural network (DNN)-based models, the task of learning robust, discriminative features for accurate person identification in large-scale datasets remains a formidable challenge. This difficulty stems from substantial intra-class variations caused by factors such as changes in pose, occlusions, and cluttered backgrounds.