Vehicle re-identification lightweight method and system based on neural architecture search

• Main Authors: ZHENG HANGJIE, ZHANG XUANQING, ZENG HUANQIANG, SHI YIFAN, ZHU JIANQING, SHEN JIANNAN, CHEN JING, JIA BAOZHI
• Format: Patent
• Language: Chinese; English
• Published: 10.11.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses a vehicle re-identification lightweight method and system based on neural architecture search, and relates to the technical field of computer vision and machine learning, and the method comprises the steps: S1, constructing a network model, giving a hardware constraint to generate a corresponding architecture generator, and training the architecture generator through a test set; s2, inputting the hardware constraint into an architecture generator to obtain a plurality of lightweight neural network architecture models; s3, collecting vehicle re-identification data and performing data enhancement; s4, training and verifying the lightweight neural network architecture model by using the vehicle re-identification data set, and selecting the lightweight neural network architecture model with the best effect as a lightweight vehicle re-identification network; and S5, vehicle re-identification is carried out based on the lightweight vehicle re-identification network. According to the method,