WS-DAN.PyTorch

A neat 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", arXiv:1901.09891)

NOTICE: This is NOT an official implementation by authors of WS-DAN. The official implementation is available at tau-yihouxiang/WS_DAN (and there's another unofficial PyTorch version wvinzh/WS_DAN_PyTorch).

Innovations

1. Data Augmentation: Attention Cropping and Attention Dropping

   

2. Bilinear Attention Pooling (BAP) for Features Generation

   

3. Training Process and Testing Process

   

Performance

• PyTorch experiments were done on a Titan Xp GPU (batch_size = 12).

Dataset	Object	Category	Train	Test	Accuracy (Paper)	Accuracy (PyTorch)	Feature Net
FGVC-Aircraft	Aircraft	100	6,667	3,333	93.0	93.28	inception_mixed_6e
CUB-200-2011	Bird	200	5,994	5,794	89.4	88.28	inception_mixed_6e
Stanford Cars	Car	196	8,144	8,041	94.5	94.38	inception_mixed_6e
Stanford Dogs	Dog	120	12,000	8,580	92.2	89.66	inception_mixed_7c

Usage

WS-DAN

This repo contains WS-DAN with feature extractors including VGG19('vgg19', 'vgg19_bn'), ResNet34/50/101/152('resnet34', 'resnet50', 'resnet101', 'resnet152'), and Inception_v3('inception_mixed_6e', 'inception_mixed_7c') in PyTorch form, see ./models/wsdan.py.

net = WSDAN(num_classes=num_classes, M=num_attentions, net='inception_mixed_6e', pretrained=True)
net = WSDAN(num_classes=num_classes, M=num_attentions, net='inception_mixed_7c', pretrained=True)
net = WSDAN(num_classes=num_classes, M=num_attentions, net='vgg19_bn', pretrained=True)
net = WSDAN(num_classes=num_classes, M=num_attentions, net='resnet50', pretrained=True)

Dataset Directory

• FGVC-Aircraft (Aircraft)

  -/FGVC-Aircraft/data/
                  └─── images
                          └─── 0034309.jpg
                          └─── 0034958.jpg
                          └─── ...
                  └─── variants.txt
                  └─── images_variant_trainval.txt
                  └─── images_variant_test.txt
  

• CUB-200-2011 (Bird)

  -/CUB-200-2011
          └─── images.txt
          └─── image_class_labels.txt
          └─── train_test_split.txt
          └─── images
                  └─── 001.Black_footed_Albatross
                          └─── Black_Footed_Albatross_0001_796111.jpg
                          └─── ...
                  └─── 002.Laysan_Albatross
                  └─── ...
  

• Stanford Cars (Car)

  -/StanfordCars
        └─── cars_test
                  └─── 00001.jpg
                  └─── 00002.jpg
                  └─── ...
        └─── cars_train
                  └─── 00001.jpg
                  └─── 00002.jpg
                  └─── ...
        └─── devkit
                  └─── cars_train_annos.mat
        └─── cars_test_annos_withlabels.mat
  

• Stanford Dogs (Dog)

  -/StanfordDogs
        └─── Images
            └─── n02085620-Chihuahua
                    └─── n02085620_10074.jpg
                    └─── ...
            └─── n02085782-Japanese_spaniel
                    └─── ...
        └─── train_list.mat
        └─── test_list.mat
  

Run

1. git clone this repo.

2. Prepare data and modify DATAPATH in datasets/<abcd>_dataset.py.

3. Set configurations in config.py (Training Config, Model Config, Dataset/Path Config):

   tag = 'aircraft'  # 'aircraft', 'bird', 'car', or 'dog'

4. $ nohup python3 train.py > progress.bar & for training.

5. $ tail -f progress.bar to see training process (tqdm package is required. Other logs are written in <config.save_dir>/train.log).

6. Set configurations in config.py (Eval Config) and run $ python3 eval.py for evaluation and visualization.

Attention Maps Visualization

Code in eval.py helps generate attention maps. (Image, Heat Attention Map, Image x Attention Map)