classifier study based on caffe

Deps

caffe: https://github.com/BVLC/caffe.git
commits: iea455eb29393ebe6de9f14e88bfce9eae74edf6d

Usage

prepare datasets
create the related soft links to caffe
1_dataset.py: generate train and test dataset lists, and write class name label list
2_images_to_lmdb.sh: convert images to lmdb format
3_compute_image_mean.sh: get image mean values
modify the prototxt files, and prepare to train and test

Outputs

data
models

Test result