一种单阶段无监督可见光-红外跨模态行人重识别方法
Single Stage Unsupervised Visible-infrared Person Re-identification
娄刃 1和任强 2赵三元 3郝昕 2周跃琪 1汪心渊 1李方芳4
作者信息
- 1. 浙江省交通运输科学研究院 杭州 310000
- 2. 北京理工大学计算机学院 北京 100081
- 3. 北京理工大学计算机学院 北京 100081;北京理工大学长三角研究院(嘉兴) 浙江嘉兴 314011
- 4. 浙江交投高速公路运营管理有限公司企业研究院 杭州 310000
- 折叠
摘要
无监督"可见光-红外"跨模态行人重识别任务能够缓解智能监控场景中需要大量人工标注的问题.常见多阶段模型用于处理不同模态数据.文中提出了一种有效的单阶段无监督跨模态行人重识别的方法,设计了基于置信因子的聚类算法和图嵌入的跨模态特征处理方法,分别用于解决无标签问题和跨模态问题.实验结果表明,相较于现有算法,所提方法在r=1时精度至少取得了7％的提高.
Abstract
The unsupervised visible-infrared multi-modal person re-identification can alleviate the problem that a lot of manual la-beling is required in the intelligent monitoring scene.Common multi-stage models are used to process different modal data sepa-rately.This paper proposes an effective single-stage unsupervised cross-modal pedestrian recognition method,and designs a clus-tering algorithm based on confidence factor and a cross-modal feature processing method based on graph embedding to solve the unlabeled problem and cross-modal problem respectively.Experimental results show that compared with the existing algorithms,the proposed algorithm has achieved an improvement of at least 7％in the case of r=1.
关键词
跨模态学习/无监督行人重识别/可见光-红外行人重识别/无监督学习/跨模态特征处理Key words
Cross-modal learning/Unsupervised person re-identification/Visible-infrared person re-identification/Unsupervised learning/Cross-modal feature processing引用本文复制引用
基金项目
浙江省交通厅科技项目(202209)
浙江省科技厅公益性项目(LGC22E080003)
出版年
2024
NETL
NSTL
万方数据