论文阅读 [TPAMI-2022] Pose-Guided Representation Learning for Person Re-Identification

论文阅读 [TPAMI-2022] Pose-Guided Representation Learning for Person Re-Identification

论文搜索(studyai.com)

搜索论文: Pose-Guided Representation Learning for Person Re-Identification

搜索论文: http://www.studyai.com/search/whole-site/?q=Pose-Guided+Representation+Learning+for+Person+Re-Identification

关键字(Keywords)

Feature extraction; Training; Pose estimation; Robustness; Image segmentation; Body regions; Measurement; Person re-identification; pose variation; misalignment; pose-guided representation

机器视觉; 自然语言处理

检测分割; 姿态估计; 行人(再)识别; 语言表示学习

摘要(Abstract)

The large pose variations and misalignment errors exhibited by person images significantly increase the difficulty of person Re-Identification (ReID).

人体图像显示的巨大姿势变化和错位误差显著增加了人体重新识别（ReID）的难度。.

Existing works commonly apply extra operations like pose estimation, part segmentation, etc., to alleviate those issues and improve the robustness of pedestrian representations.

现有的工作通常采用额外的操作，如姿势估计、部分分割等，以缓解这些问题并提高行人表示的鲁棒性。.

While boosting the ReID accuracy, those operations introduce considerable computational overheads and make the deep models complex and hard to tune.

在提高里德精度的同时，这些操作会带来相当大的计算开销，并使深层模型变得复杂且难以调整。.

To chase a more efficient solution, we propose a Part-Guided Representation (PGR) composed of Pose Invariant Feature (PIF) and Local Descriptive Feature (LDF), respectively.

为了寻求更有效的解决方案，我们提出了一种分别由姿态不变特征（PIF）和局部描述特征（LDF）组成的零件引导表示（PGR）。.

We call PGR “Part-Guided” because it is trained and supervised by local part cues.

我们称PGR为“部分引导”，因为它由本地部分线索进行训练和监督。.

Specifically, PIF approximates a pose invariant representation inferred by pose estimation and pose normalization.

具体来说，PIF近似于通过姿势估计和姿势规范化推断出的姿势不变表示。.

LDF focuses on discriminative body parts by approximating a representation learned with body region segmentation.

LDF通过近似通过身体区域分割学习的表示来关注有区别的身体部位。.

In this way, extra pose extraction is only introduced during the training stage to supervise the learning of PGR, but is not required during the testing stage for feature extraction.

这样，额外的姿势提取只在训练阶段引入，以监督PGR的学习，而在特征提取的测试阶段不需要额外的姿势提取。.

Extensive comparisons with recent works on five widely used datasets demonstrate the competitive accuracy and efficiency of PGR…

与最近在五个广泛使用的数据集上的工作进行的广泛比较证明了PGR具有竞争力的准确性和效率。。.

作者(Authors)

[‘Jianing Li’, ‘Shiliang Zhang’, ‘Qi Tian’, ‘Meng Wang’, ‘Wen Gao’]