Discriminative Localization with Image-Level Annotation

Pytorch implementation of "Learning Deep Features for Discriminative Localization"

B. Zhou, A. Khosla, A. Lapedriza, A. Oliva, and A. Torralba Learning Deep Features for Discriminative Localization. Computer Vision and Pattern Recognition (CVPR), 2016. [PDF][Project Page]

Contents

1. Basic installation
2. Data and Pretrain Model
3. Demo
4. Beyond the demo:Training and Testing
5. Results of Action40 Dataset

Basic installation

Requirements for pytorch (see: pytorch installation instuctions)

Data and Pretrain Model

• Action40 datasets
  • action_40_dataset.tar
  • class_id.json
• Pretrain Models
  • vgg_16_converted_from_caffe.pth
  • action_40_pretrain_model.pth

Demo

After sucessfully completing Basic installation and Data and Pretrain Model, you will be ready to run the demo.

cd $CAM_ROOT
./demo.py

Beyond the demo: Training and Testing

Train:

set train_flag = True in action40_config.py , then run train_action40.py.

Test:

set train_flag = False in action40_config.py , then run test_action40.py

Results