/ACoL_pytorch

Adversarial Complementary Learning for Weakly Supervised Object Localization Pytorch reproducing

Primary LanguagePython

Adversarial Complementary Learning for Weakly Supervised Object Localization

This is PyTorch implementation of Adversarial Complementary Learning for Weakly Supervised Object Localization

Prerequisites

  • Python 3.6+
  • Pytorch ( >= 1.1)
  • Python bindings for OpenCV
  • tqdm
  • tensorboardX

Data Preparation

Execution

Train
git clone https://github.com/halbielee/ACoL_pytorch.git
wget http://www.vision.caltech.edu/visipedia-data/CUB-200-2011/CUB_200_2011.tgz
tar -xvf CUB_200_2011.tgz
cd ACoL_pytorch
bash script/train.sh
Evaluate
# First download trained model from the link
bash script/evaluate.sh

Performance

We provide the performance with trained model.

Dataset Method Acc1 Acc5 Top1_LOC(0.15/0.2) GT-known condition
CUB-200 acol 76.46 92.44 49.78 / 46.06 59.46 batch 32, lr 0.001, wd 1e-4, 40/150, thr 0.7

Qualitative Image

image