arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.10313cs.CV
收藏

无监督域自适应Re-ID中改进的交互式均值教学

Improved Mutual Mean-Teaching for Unsupervised Domain Adaptive Re-ID

The Chinese University of Hong Kong(香港中文大学)

Yixiao Ge, Shijie Yu, Dapeng Chen

摘要:

在这份技术报告中,我们提交了我们在ECCV 2020中的VisDA挑战赛,我们在排行榜上取得了最佳成绩之一。我们的解决方案是基于结构化领域自适应(SDA)和相互意义教学(MMT)框架。SDA是一个基于域翻译的框架,它专注于将源域图像仔细地翻译到目标域。MMT是一个基于伪标签的框架,专注于使用健壮的软标签进行伪标签精炼。具体来说,我们的培训管道有三个主要步骤。(i)我们采用SDA来生成源到目标的翻译图像,以及(ii)这些图像作为信息丰富的训练样本来预训练网络。(iii)MMT在目标域上进一步微调预训练的网络。请注意,我们设计了一个改进的MMT(称为MMT+),以进一步减轻标签噪声,通过建模两个域之间的样本间关系,并保持实例的歧视。我们提出的方法在mAP方面取得了74.78%的准确率,在153个团队中排名第二。

英文摘要:

In this technical report, we present our submission to the VisDA Challenge in ECCV 2020 and we achieved one of the top-performing results on the leaderboard. Our solution is based on Structured Domain Adaptation (SDA) and Mutual Mean-Teaching (MMT) frameworks. SDA, a domain-translation-based framework, focuses on carefully translating the source-domain images to the target domain. MMT, a pseudo-label-based framework, focuses on conducting pseudo label refinery with robust soft labels. Specifically, there are three main steps in our training pipeline. (i) We adopt SDA to generate source-to-target translated images, and (ii) such images serve as informative training samples to pre-train the network. (iii) The pre-trained network is further fine-tuned by MMT on the target domain. Note that we design an improved MMT (dubbed MMT+) to further mitigate the label noise by modeling inter-sample relations across two domains and maintaining the instance discrimination. Our proposed method achieved 74.78% accuracies in terms of mAP, ranked the 2nd place out of 153 teams.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑