oguzakif's picture
init repo
d4b77ac
|
raw
history blame
545 Bytes

Preprocessing VID(Object detection from video)

Large Scale Visual Recognition Challenge 2015 (ILSVRC2015)

Download dataset (86GB)

wget http://bvisionweb1.cs.unc.edu/ilsvrc2015/ILSVRC2015_VID.tar.gz
tar -xzvf ./ILSVRC2015_VID.tar.gz
mv ILSVRC2015/Annotations/VID/val ILSVRC2015/Annotations/VID/train/
mv ILSVRC2015/Data/VID/val ILSVRC2015/Data/VID/train/

Crop & Generate data info (20 min)

python parse_vid.py

#python par_crop.py [crop_size] [num_threads]
python par_crop.py 511 12
python gen_json.py