Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/cahya-wirawan/ws-dan.pytorch

A PyTorch implementation of WS-DAN (Weakly Supervised Data Augmentation Network) for FGVC (Fine-Grained Visual Classification)

https://github.com/cahya-wirawan/ws-dan.pytorch

Science Score: 10.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ○
    codemeta.json file
  • ○
    .zenodo.json file
  • ○
    DOI references
  • ✓
    Academic publication links
    Links to: arxiv.org
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (6.1%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

A PyTorch implementation of WS-DAN (Weakly Supervised Data Augmentation Network) for FGVC (Fine-Grained Visual Classification)

Basic Info
  • Host: GitHub
  • Owner: cahya-wirawan
  • License: mit
  • Language: Python
  • Default Branch: master
  • Homepage:
  • Size: 447 KB
Statistics
  • Stars: 0
  • Watchers: 2
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Fork of GuYuc/WS-DAN.PyTorch
Created about 7 years ago · Last pushed about 7 years ago

https://github.com/cahya-wirawan/WS-DAN.PyTorch/blob/master/ 

# WS-DAN.PyTorch
A PyTorch implementation of WS-DAN (Weakly Supervised Data Augmentation Network) for FGVC (Fine-Grained Visual Classification). (Hu et al., ["See Better Before Looking Closer: Weakly Supervised Data Augmentation
Network for Fine-Grained Visual Classification"](https://arxiv.org/abs/1901.09891v2), arXiv:1901.09891)

**NOTICE: This is NOT an official implementation by authors of WS-DAN.**


## Attention Cropping and Attention Dropping
![Fig1](./images/Fig1.png)

The framework introduce an attention based method for extracting more detailed features and more object's parts by Attention Cropping and Attention Dropping, see Fig 1. 

## Training Process and  Testing Process 
![Fig2a](./images/Fig2a.PNG)

![Fig2b](./images/Fig2b.PNG)

## Bilinear Attention Pooling (BAP)

![Fig3](./images/Fig3.PNG)

## Usage
This code repo contains WS-DAN with feature extractors including VGG19, ResNet(34, 50, 101, 152), and Inception_v3 in PyTorch form. The default feature extractor is Inception_v3, and this can be modified conveniently in ```train_wsdan.py```: 

```python
# feature_net = vgg19_bn(pretrained=True)
# feature_net = resnet101(pretrained=True)
feature_net = inception_v3(pretrained=True)

net = WSDAN(num_classes=num_classes, M=num_attentions, net=feature_net)
```

1. ``` git clone``` this repo.
2. Prepare image data and rewrite ```dataset.py``` for your CustomDataset.
3. ```$ nohup python3 train_wsdan.py -j  -b  --sd  (etc.) 1>log.txt 2>&1 &``` (see ```train_wsdan.py``` for more training options)
4. ```$ tail -f log.txt``` for logging information.

Owner

  • Name: Cahya Wirawan
  • Login: cahya-wirawan
  • Kind: user
  • Location: Vienna, Austria

System engineer, currently working on NLP, CV and Speech Recognition for fun and curiosity

GitHub Events

Total
Last Year