Vehicle re-identification aims to retrieve and match vehicles under non-overlapping cameras. Although this technology has made some progress in recent years, the problem of vehicle appearance ambiguity caused by perspective shift still can not be well resolved. To alleviate the above problem, this paper proposes a vehicle re-identification method based on keypoint-guided semantic feature alignment and graph matching strategy to enhance complementary information under the transformer framework. The method first uses the pre-trained vehicle attitude detection model to extract keypoints, and proposes an information mapping strategy. Through the coordinate information of keypoints, the corresponding local feature tokens under the transformer framework are extracted and given semantic content to achieve the same semantics. Feature alignment for attribute tokens. Then, a graph matching network is constructed to realize the transfer of semantic information between similar samples, and the expressive ability of features is further improved through the multivariate interaction of information. The proposed model achieves state-of-the-art performance and superiority compared to similar methods on two large vehicle re-ID datasets.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.