Vehicle verification based on deep siamese network with similarity metric

Qian Zhang, Mingtao Pei*, Mei Chen, Yunde Jia

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

4 引用 (Scopus)

摘要

Vehicle verification is a challenging research problem with important practical applications. Most prior work focused on either feature learning or distance metric learning, which could not guarantee the compatibility of the learned feature and the distance metric. In this paper, we propose an end-to-end model based on the Siamese Convolutional Neural Network (CNN), which integrates distance metric learning and feature learning into a unified framework. The network is trained by contrastive loss and a similarity metric loss defined by joint Bayesian to learn more discriminative features for vehicle verification. The experimental results demonstrate the effectiveness of the proposed method.

源语言英语
主期刊名Advances in Multimedia Information Processing – PCM 2017 - 18th Pacific-Rim Conference on Multimedia, Revised Selected Papers
编辑Bing Zeng, Hongliang Li, Abdulmotaleb El Saddik, Xiaopeng Fan, Shuqiang Jiang, Qingming Huang
出版商Springer Verlag
773-782
页数10
ISBN(印刷版)9783319773797
DOI
出版状态已出版 - 2018
活动18th Pacific-Rim Conference on Multimedia, PCM 2017 - Harbin, 中国
期限: 28 9月 201729 9月 2017

出版系列

姓名Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
10735 LNCS
ISSN(印刷版)0302-9743
ISSN(电子版)1611-3349

会议

会议18th Pacific-Rim Conference on Multimedia, PCM 2017
国家/地区中国
Harbin
时期28/09/1729/09/17

指纹

探究 'Vehicle verification based on deep siamese network with similarity metric' 的科研主题。它们共同构成独一无二的指纹。

引用此