Vehicle re-identification method and device based on vehicle global attribute perception and local discriminant feature learning

By using the global attribute perception module, local area perception module and fine-grained feature extraction module in vehicle recognition, the vehicle features are extracted and trained in multi-scale, which solves the problems of intra-class differences and inter-class similarity in vehicle recognition, and achieves high-precision vehicle recognition.

CN119693888AActive Publication Date: 2025-03-25HUBEI CHUTIAN HIGH SPEED DIGITAL TECH CO LTD
5 Cites 0 Cited by

Patent Information

Application Number
CN202411830048.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-12
Publication Date
2025-03-25
Estimated Expiration
2044-12-12

AI Technical Summary

Technical Problem

The existing deep learning-based vehicle recognition methods face great in-class differences caused by changes in vehicle perspective and camera perspective, and the extreme inter-class similarities caused by different vehicles that may have the same color, model and other attributes.

Method used

The global attribute perception module based on the Transformer mechanism and the vehicle local area perception module based on the attention mechanism are adopted, combined with the pyramid-shaped local fine-grained feature extraction module, the vehicle features are global attribute perception, local area feature extraction and fine-grained feature extraction, and the backbone network is trained through multi-scale loss functions.

Benefits of technology

The recognition accuracy of vehicle re-identification is improved, and excellent re-identification results can be displayed on data sets in different domains such as drastic viewing angle, low picture pixels, uneven vehicle attributes, and day and night cross-domain, effectively responding to the challenges of vehicle re-identification inter-class similarity and intra-class differences.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a vehicle re-identification method and device based on vehicle global attribute perception and local discriminant feature learning. The method comprises the following steps: acquiring a vehicle re-identification data set, and inputting the vehicle re-identification data set into a backbone network to obtain vehicle features; inputting the vehicle features into a global attribute sensing module based on a Transform mechanism, a vehicle local area sensing module based on an attention mechanism and a pyramid-shaped local fine-grained feature extraction module to obtain global attribute features, local area features and local fine-grained features of the vehicle; calculating a global attribute loss function, a local area loss function and a local fine-grained loss function respectively; training the backbone network based on the loss function; and inputting the vehicle re-identification data set into the trained backbone network to obtain a plurality of vehicle picture features, calculating the similarity of the plurality of vehicle picture features, and arranging the vehicle picture features according to the similarity to obtain a vehicle re-identification result. According to the invention, the recognition precision of vehicle re-recognition can be improved.
Need to check novelty before this filing date? Find Prior Art