[论文解读] A Novel Teacher-Student Learning Framework For Occluded Person Re-Identification

Jiaxuan Zhuo, Jianhuang Lai+1 more|arXiv (Cornell University)|Jul 7, 2019

Video Surveillance and Tracking Methods参考文献 31被引用 45

一句话总结

本文提出一个两阶段的教师-学生学习框架，以鲁棒地重新识别被遮挡的人，并利用共显著性网络和跨域仿真器将知识从全身域迁移到遮挡域。

ABSTRACT

Person re-identification (re-id) has made great progress in recent years, but occlusion is still a challenging problem which significantly degenerates the identification performance. In this paper, we design a teacher-student learning framework to learn an occlusion-robust model from the full-body person domain to the occluded person domain. Notably, the teacher network only uses large-scale full-body person data to simulate the learning process of occluded person re-id. Based on the teacher network, the student network then trains a better model by using inadequate real-world occluded person data. In order to transfer more knowledge from the teacher network to the student network, we equip the proposed framework with a co-saliency network and a cross-domain simulator. The co-saliency network extracts the backbone features, and two separated collaborative branches are followed by the backbone. One branch is a classification branch for identity recognition and the other is a co-saliency branch for guiding the network to highlight meaningful parts without any manual annotation. The cross-domain simulator generates artificial occlusions on full-body person data under a growing probability so that the teacher network could train a cross-domain model by observing more and more occluded cases. Experiments on four occluded person re-id benchmarks show that our method outperforms other state-of-the-art methods.

研究动机与目标

• 解决在人再识别中因为遮挡导致的识别性能下降的问题。
• 利用大规模全身数据来为鲁棒学习模拟遮挡场景。
• 在受益于教师驱动的遮挡感知特征的同时，使用真实遮挡数据引导学生网络。

提出的方法

• 提出一个由教师阶段使用全身数据来模拟遮挡、学生阶段从真实遮挡数据学习组成的两阶段教师-学生框架。
• 提出一个共显著性网络，包含骨干网络和两个分支（身份分类、共显著性用于显著部位检测），以突出人体部位。
• 在训练期间使用跨域仿真器将全身图像逐步转换为模拟遮挡图像。
• 使用 occluded/non-occluded (OBC) 二分类器来对齐领域迁移，并在多任务损失中整合身份损失和 OBC 损失。
• 将教师的共显著性输出作为学生训练的真实标签，以提高遮挡鲁棒性和显著性估计。

实验结果

研究问题

• RQ1 教师-学生框架是否能够弥合全身人和遮挡人域在再识别中的差距？
• RQ2 共显著性指导和跨域仿真是否提高遮挡人再识别的性能？
• RQ3 growing-probability 遮挡机制如何影响遮挡鲁棒特征的学习？
• RQ4 将从模拟遮挡到真实遮挡的知识迁移对学生模型是否有帮助？
• RQ5 共显著性分支对检测显著遮挡区域是否比其他显著性检测器更有效？

主要发现

• 所提出的框架在四个基准数据集上相较于现有方法，呈现出更好的遮挡人再识别性能。
• 同时具有分类与共显著性分支、再加上跨域仿真和 OBC 损失的教师网络，相较基线配置带来显著提升。
• growing-probability 跨域仿真器在固定概率或无遮挡设置上表现更佳，有助于逐步进行域自适应。
• 共显著性分支提升显著区域定位，并提升了再识别与显著性检测指标。
• 将教师的共显著性输出作为学生的监督，有助于提升遮挡鲁棒性与检测准确性。