All crop images from http://www.vision.caltech.edu/visipedia/CUB-200-2011.html If the CUB200 dataset is helpful to you, please cite this official website.
-
Download the tgz files of images and segmentations link from the official website.
Or you can use
download.sh
script to do it. Try:$ ./download.sh
(Make sure you have installed
wget
andtar
before you execute the script.) -
Extract tgz files to
CUB_200_2011
folder, or any path you want.Now check the following python packages required:
- os
- numpy
- argparse
- scipy
and run
center_crop.py
to process the raw images. For example:$ python3 center_crop.py -dd CUB_200_2011/ -os '80,80' -od 'center_crop/'
Try
python3 center_crop.py -h
to see more information how to use. -
All crop-images and crop-segmentations saved in
<dataset_path>/center_crop/
.