K-AI image-classification

Kor readme - link
Code for solution in mask, gender, age classification of boostcamp Aistages

Getting Started

Dependencies

• torch==1.6.0
• torchvision==0.7.0
• tensorboard==2.4.1
• pandas==1.1.5
• opencv-python==4.5.1.48
• scikit-learn==0.24.1
• matplotlib==3.2.1
• timm==0.4.12
• albumentations==1.0.3
• pandas-streaming==0.2.175
• facenet-pytorch==2.5.2

Installation

git clone https://github.com/pudae/kaggle-understanding-clouds.git
pip install -r requirements.txt

Dataset

This dataset consist of face images and labels. label is class that combinates mask, gender and age. Unfortunately Aistages said that the dataset in this competition can not be made public.

Class Description

Class description is like below.

Class	Mask	Gender	Age
0	Wear	Male	<30
1	Wear	Male	>=30 and <60
2	Wear	Male	>=60
3	Wear	Female	<30
4	Wear	Female	>=30 and <60
5	Wear	Female	>=60
6	Incorrect	Male	<30
7	Incorrect	Male	>=30 and <60
8	Incorrect	Male	>=60
9	Incorrect	Female	<30
10	Incorrect	Female	>=30 and <60
11	Incorrect	Female	>=60
12	Not Wear	Male	<30
13	Not Wear	Male	>=30 and <60
14	Not Wear	Male	>=60
15	Not Wear	Female	<30
16	Not Wear	Female	>=30 and <60
17	Not Wear	Female	>=60

Dataset folder path

train/
├─train.csv
└─images/
  ├─(id)_(gender)_(race)_(age)/
  | ├─mask1.jpg
  | ├─mask2.jpg
  | ├─mask3.jpg
  | ├─mask4.jpg
  | ├─mask5.jpg
  | ├─incorrect_mask.jpg
  | └─normal.jpg
  ├─000001_Female_Asian_20/
  | ├─mask1.jpg
  | ├─mask2.jpg
  | ├─mask3.jpg
  | ├─mask4.jpg
  | ├─mask5.jpg
  | ├─incorrect_mask.jpg
  | └─normal.jpg
  .
  .
  .
  └─...
eval/
├─train.csv
└─images/
  ├─(id).jpg/
  ├─....jpg/
  .
  .
  .
  └─....jpg/

Code Components

├── FaceCrop.ipynb
├── README.md
├── dataset.py
├── evaluation.py
├── inference.py
├── total_result.py
├── loss.py
├── model
├── model.py
├── requirements.txt
├── sh
│   ├── inference_ViT.sh
│   ├── inference_effnet.sh
│   ├── inference_resnet.sh
│   ├── train_ViT.sh
│   ├── train_ViT_optuna.sh
│   ├── train_effnet.sh
│   ├── train_resnet.sh
│   └── train_resnet_multi.sh
└── train.py

Model

Models are included like below

• ResNet
• EfficientNet
• VGG
• Xception
• ViT

Train

We train models EfficientNet_b7. You can train using our train.py file. Simply you can train model our using sh/train_effnet.sh script. If you can need other option, change args in sh script.

sh ./sh/train_effnet.sh

Inference

If you finish training model. You can create output.csv file using inference.py. We give sample inference_effnet.sh script.

sh ./sh/inference_effnet.sh

Evaluation

If you have ground-truth of evaluation dataset. you can evaluate using evaluation.py file.