Code for C-MIL: Continuation Multiple Instance Learning for Weakly Supervised Object Detection.

New: Pytorch version of C-MIL is avalable at here. Thanks for the contribution of shenyunhang.

• Ubuntu 16.04 LTS
• NVIDIA V100 + CUDA9.0 + CuDNN7.0
• Torch7

1. Install the dependencies

   cd ./C-MIL
   export DIR=$(pwd) 
   
   luarocks install hdf5 matio protobuf rapidjson loadcaffe xml
   
   cd $DIR/libs/functions
   sh install.sh
   
   cd $DIR/layers
   luarocks make

2. Download dataset, proposals and ImageNet pre-trained model

   Download VOC2007 from: https://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar https://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar

   Download proposals from: https://dl.dropboxusercontent.com/s/orrt7o6bp6ae0tc/selective_search_data.tgz

   Download VGGF from: https://www.robots.ox.ac.uk/~vgg/software/deep_eval/releases/bvlc/VGG_CNN_F.caffemodel https://gist.githubusercontent.com/ksimonyan/a32c9063ec8e1118221a/raw/6a3b8af023bae65669a4ceccd7331a5e7767aa4e/VGG_CNN_F_deploy.prototxt

   mkdir $DIR/data    
   mkdir $DIR/output

   The data folder has the following structure:

   $C-MIL/data/datasets/VOCdevkit_2007/
   $C-MIL/data/datasets/VOCdevkit_2007/VOCcode
   $C-MIL/data/datasets/VOCdevkit_2007/VOC2007
   $C-MIL/data/datasets/VOCdevkit_2007/...
   $C-MIL/data/datasets/proposals/
   $C-MIL/data/models/
   $C-MIL/data/results/

3. Train, test and evaluate

   cd $DIR
   # train
   th train_cmil.lua 0 SSW
   # test
   th test_cmil.lua 0 SSW
   # evaluate
   th detection_mAP.lua 0 SSW output/path/to/scorefiles/score_test_epoch20.h5 2

Acknowledgements will be added later.