1st Place Solution to VisDA-2020: Bias Elimination for Domain Adaptive Pedestrian Re-identification

December 25, 2020 · Entered Twilight · 🏛 arXiv.org

"Last commit was 5.0 years ago (≥5 year threshold)"

Evidence collected by the PWNC Scanner

Repo contents: .gitignore, LICENSE, README.md, VisDA.md, config, configs, data, engine, files, layers, modeling, solver, test.py, train_adaptation.py, train_baseline.py, train_camera.py, utils, validate.py

Authors Jianyang Gu, Hao Luo, Weihua Chen, Yiqi Jiang, Yuqi Zhang, Shuting He, Fan Wang, Hao Li, Wei Jiang arXiv ID 2012.13498 Category cs.CV: Computer Vision Citations 5 Venue arXiv.org Repository https://github.com/vimar-gu/Bias-Eliminate-DA-ReID ⭐ 57 Last Checked 2 months ago

Abstract

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