【论文阅读】行人重识别——Bag of tricks and a strong baseline for deep person re-identification

最新推荐文章于 2024-08-09 08:32:10 发布

**七秒记忆**

最新推荐文章于 2024-08-09 08:32:10 发布

阅读量3k

点赞数 2

分类专栏：论文阅读文章标签：行人重识别训练技巧

本文链接：https://blog.csdn.net/qq_40605139/article/details/97489161

版权

该论文研究了行人重识别（ReID）的简单有效基准，通过结合多种训练技巧，模型在Market1501上达到了94.5%的Rank-1准确率和85.9%的mAP。创新点包括BNNeck结构，解决了ID损失和三联体损失的不一致问题，以及中心损失以提高类内紧凑性和类间可分性。

摘要由CSDN通过智能技术生成

《Bag of tricks and a strong baseline for deep person re-identification》

原文

论文链接：https://arxiv.org/pdf/1903.07071.pdf

代码链接：https://github.com/michuanhaohao/reid-strong-baseline

摘要：

This paper explores a simple and efficient baseline for person re-identification (ReID). Person re-identification(ReID) with deep neural networks has made progress and achieved high performance in recent years.本文探讨了一种简单有效的人员重新识别基准（ReID）。具有深度神经网络的人员重新识别（ReID）近年来取得了进步并取得了很高的性能。However, many state-of-the-arts methods design complex network structure and concatenate multi-branch features. In the literature, some effective training tricks are briefly appeared in several papers or source codes