Strong_Baseline_of_Pedestrian_Attribute_Recognition的readme.md的译文

1. 行人属性识别

论文：Rethinking of Pedestrian Attribute Recognition: Realistic Datasets with Efficient Method.

考虑到各种SOTA基准存在巨大的性能差距，因此我们提供了坚实而强大的基准进行公平比较。

2. 更新

• 20200901 添加了 infer.py文件

3. 依赖的python库

• scipy==1.4.1
• torch==1.4.0（pytorch 1.4.0）
• torchvision==0.5.0
• tqdm==4.43.0
• easydict==1.9
• numpy==1.18.1
• Pillow==7.1.2

4. 技巧

• 样本丢失而不是标签丢失
• 高学习率结合梯度裁剪(clip_grad_norm)
• 增强填充结合随机随机裁剪(RandomCrop)
• 分类层后添加BN(Batch Normalization，批量归一化)

5. 性能比较

5.1 基准性能

• 与MsVAA，VAC，ALM的基准性能相比，我们的基准性能有了很大提升。
• 与复现的MsVAA，VAC，ALM相比，我们的基准更好。
• 我们尽力复现了 MsVAA, VAC ，很感谢他们的贡献
• 我们也尽力复现了ALM ，还有尝试联系该作者, 但是都没有得到答复

SOTA 性能

• 与最新技术方法的性能相比，我们的基准性能相当，甚至更好

• DeepMAR (ACPR15) Multi-attribute Learning for Pedestrian Attribute Recognition in Surveillance Scenarios.
• HPNet (ICCV17) Hydraplus-net: Attentive deep features for pedestrian analysis.
• JRL (ICCV17) Attribute recognition by joint recurrent learning of context and correlation.
• LGNet (BMVC18) Localization guided learning for pedestrian attribute recognition.
• PGDM (ICME18) Pose guided deep model for pedestrian attribute recognition in surveillance scenarios.
• GRL (IJCAI18) Grouping Attribute Recognition for Pedestrian with Joint Recurrent Learning.
• RA (AAAI19) Recurrent attention model for pedestrian attribute recognition.
• VSGR (AAAI19) Visual-semantic graph reasoning for pedestrian attribute recognition.
• VRKD (IJCAI19) Pedestrian Attribute Recognition by Joint Visual-semantic Reasoning and Knowledge Distillation.
• AAP (IJCAI19) Attribute aware pooling for pedestrian attribute recognition.
• MsVAA (ECCV18) Deep imbalanced attribute classification using visual attention aggregation.
• VAC (CVPR19) Visual attention consistency under image transforms for multi-label image classification.
• ALM (ICCV19) Improving Pedestrian Attribute Recognition With Weakly-Supervised Multi-Scale Attribute-Specific Localization.

数据集

PETA： [Paper][Project]

PA100K：[Paper][Github]

RAP ： A Richly Annotated Dataset for Pedestrian Attribute Recognition

• v1.0 [Paper][Project]
• v2.0 [Paper][Project]

Zero-shot Protocal

PETA 和 RAPv2 数据集： Google Drive.

只需要将“ dataset.pkl”替换为“ peta_new.pkl”或“ rapv2_new.pkl”即可在新协议（Protocal）下做这个实验。

预训练模型

预训练模型Google Drive.

使用步骤

1. 克隆项目，项目：github

git clone https://github.com/valencebond/Strong_Baseline_of_Pedestrian_Attribute_Recognition.git

2. 创建目录存放数据集

   cd Strong_Baseline_of_Pedestrian_Attribute_Recognition mkdir data

3. 数据存放的结构：

   ${project_dir}/data
       PETA
           images/
           PETA.mat
           README
       PA100k
           data/
           annotation.mat
           README.txt
       RAP
           RAP_dataset/
           RAP_annotation/
       RAP2
           RAP_dataset/
           RAP_annotation/
   

4. 运行format_xxxx.py 分别生成 dataset.pkl

   python ./dataset/preprocess/format_peta.py
   python ./dataset/preprocess/format_pa100k.py
   python ./dataset/preprocess/format_rap.py
   python ./dataset/preprocess/format_rap2.py
   

5. 使用resnet50进行训练

   CUDA_VISIBLE_DEVICES=0 python train.py PETA
   

致谢

代码来源：
①Dangwei Li
② Houjing Huang.

文献

在研究中使用此方法或此代码，请引用为：

@misc{jia2020rethinking,
    title={Rethinking of Pedestrian Attribute Recognition: Realistic Datasets with Efficient Method},
    author={Jian Jia and Houjing Huang and Wenjie Yang and Xiaotang Chen and Kaiqi Huang},
    year={2020},
    eprint={2005.11909},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}