Code for detecting visual concepts in images.
- Create directory, checkout code, caffe, coco-hooks
git clone [email protected]:s-gupta/visual-concepts.git code
git clone [email protected]:pdollar/coco.git code/coco
- Make caffe and pycaffe
git clone [email protected]:s-gupta/caffe.git code/caffe
cd code/caffe
git checkout mil
make -j 16
make pycaffe
cd ../../
- Get the COCO images, caffe imagenet models, pretrained models on COCO.
# Get the COCO splits, ground truth
wget ftp://ftp.cs.berkeley.edu/pub/projects/vision/im2cap-cvpr15b/data.tgz && tar -xf data.tgz
# Get the COCO images
cd code
# Download and unzip the images
bash scripts/script_download_coco.sh
# The code assumes images to be stored heirarchically.
# This python scripts does the required copying.
PYTHONPATH='.' python scripts/script_download_coco.py
# Get the caffe imagenet models
wget ftp://ftp.cs.berkeley.edu/pub/projects/vision/im2cap-cvpr15b/caffe-data.tgz && tar -xf caffe-data.tgz
# Get the pretrained models
wget ftp://ftp.cs.berkeley.edu/pub/projects/vision/im2cap-cvpr15b/trained-coco.tgz && tar -xf trained-coco.tgz
cd code
and execute relevant commands from the file scripts/scripts_all.py