Pedestrian re-identification method suitable for field generalization

• Main Authors: XU HONGBING, DING JIN, ZHOU XUE, ZOU JIANXIAO
• Format: Patent
• Language: Chinese; English
• Published: 27.06.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses a pedestrian re-identification method suitable for field generalization, and belongs to the technical field of computer vision and machine learning. The method comprises the following steps: constructing a domain normalization module DN by combining distribution information of an IN layer and a BN layer, and replacing a normalization module BN of a deep network structure in a deep learning network model by using the domain normalization module DN to construct a strong baseline model; in addition, a hidden distribution space is constructed by using the distribution information of all the source domains, and distribution alignment is realized by projecting the distribution of the source domains and the target domain to the hidden distribution space; meanwhile, an anti-normalization module AN is designed, each source domain is regarded as a simulation target domain of other source domains, and on the basis that style information and identity authentication information of the target domain