Implement of Deep Multi-attribute Recognition model under ResNet50 backbone network

Preparation

Prerequisite: Python 3.10.13 and Pytorch 2.2.0

1. Install Pytorch

2. Download and prepare the dataset as follow:

   a. PETA

   • Datset內容

   ../Dataset/peta/images/*.png
   ../Dataset/peta/PETA.mat
   ../Dataset/peta/README
   

   • 執行底下程式，整理 training 需要用到的資料包成 dictionary，生成 peta_partition.pkl 後再做後續 training

   python script/dataset/transform_peta.py 
   

   b. RAP

   ../Dataset/rap/RAP_dataset/*.png
   ../Dataset/rap/RAP_annotation/RAP_annotation.mat
   

   python script/dataset/transform_rap.py
   

   c. PA100K

   ../Dataset/pa100k/data/*.png
   ../Dataset/pa100k/annotation.mat
   

   python script/dataset/transform_pa100k.py 
   

   d. RAP(v2)

   ../Dataset/rap2/RAP_dataset/*.png
   ../Dataset/rap2/RAP_annotation/RAP_annotation.mat
   

   python script/dataset/transform_rap2.py
   

3. File location

    PAR_PATH
    ├── Dataset
    │   ├── pa100k
    │   │   │── data
    │   │   │── annotation.mat
    │   │── PETA
    │   │   │── images
    │   │   │── PETA.mat
    │   │── rapv2
    │   │   │── RAP_annotation
    │   │   │── RAP_dataset
    │   │── RAPv1
    ├── pedestrian_attribute_recognition
    |   │── script
    |   │   ├── dataset
    |   │   │   ├── transform_pa100k.py
    |   │   │   ├── transform_peta.py
    |   │   │   ├── transform_rap.py
    |   │   │   ├── transform_rap2.py
    |   │   ├── experiment
    |   │   │   ├── train.sh
    |   │   │   ├── test.sh
    |   │   │   ├── train_deepmar_resnet50.py
    |   │   │   ├── baseline
    |   │   │   |   ├── ...
    |   │── exp
    |   │   ├── deepmar_resnet50
    ├── ...
   

Train the model

sh script/experiment/train.sh

Test the model

sh script/experiment/test.sh

Citation

Please cite this paper in your publications if it helps your research:

@inproceedings{li2015deepmar,
    author = {Dangwei Li and Xiaotang Chen and Kaiqi Huang},
    title = {Multi-attribute Learning for Pedestrian Attribute Recognition in Surveillance Scenarios},
    booktitle = {ACPR},
    pages={111--115},
    year = {2015}
}

Thanks

Partial codes are based on the repository from Dangwei Li.

The code should only be used for academic research.