论文标题
Disda-2020的第一名解决方案：消除域自适应行人重新识别的偏见
1st Place Solution to VisDA-2020: Bias Elimination for Domain Adaptive Pedestrian Re-identification
论文作者
论文摘要
本文介绍了我们在视觉域自适应挑战中提出的针对域自适应行人重新识别（RE-ID）任务的方法（VISDA-2020）。考虑到源域和目标域之间的较大差距，我们专注于解决两种影响域自适应行人重新ID的偏见，并提出了两阶段的训练程序。在第一阶段，基线模型接受了从源域转移到目标域以及从单个相机到多个相机样式的图像训练。然后，我们引入了一个域适应框架，以同时在源数据和目标数据上训练模型。采用不同的伪标签生成策略来不断提高模型的判别能力。最后，通过采用多个模型和其他后处理方法，我们的方法在测试集中获得了76.56％的MAP和84.25％的排名。代码可在https://github.com/vimar-gu/bias-eliminate-da-reid上找到
This paper presents our proposed methods for domain adaptive pedestrian re-identification (Re-ID) task in Visual Domain Adaptation Challenge (VisDA-2020). Considering the large gap between the source domain and target domain, we focused on solving two biases that influenced the performance on domain adaptive pedestrian Re-ID and proposed a two-stage training procedure. At the first stage, a baseline model is trained with images transferred from source domain to target domain and from single camera to multiple camera styles. Then we introduced a domain adaptation framework to train the model on source data and target data simultaneously. Different pseudo label generation strategies are adopted to continuously improve the discriminative ability of the model. Finally, with multiple models ensembled and additional post processing approaches adopted, our methods achieve 76.56% mAP and 84.25% rank-1 on the test set. Codes are available at https://github.com/vimar-gu/Bias-Eliminate-DA-ReID
