caffe convert_image

http://stackoverflow.com/questions/31427094/guide-to-use-convert-imageset-cpp

Prepare your data

Images: put all images in a folder (I'll call it here 

/path/to/jpegs/

).
Labels: create a text file (e.g., 

/path/to/labels/train.txt

)
with a line per input image . For example:

img_0000.jpeg 1

img_0001.jpeg 0

img_0002.jpeg 0

In this example the first image is labeled 

1

 while
the other two are labeled 

0

.

Note that label starting from zero, and continuous (连续的)

Convert the dataset

Run the binary in shell

~$ GLOG_logtostderr=1 $CAFFE_ROOT/build/tools/convert_imageset \
--resize_height=200 --resize_width=200 --shuffle  \
/path/to/jpegs/ \
/path/to/labels/train.txt \
/path/to/lmdb/train_lmdb

Command line explained:

GLOG_logtostderr

 flag
is set to 1 before calling 

convert_imageset

 indicates
the logging mechanism to redirect log messages to stderr.

--resize_height

 and 

--resize_width

 resize all input
images to same size 

200x200

.

--shuffle

 randomly
change the order of images and does not preserve the order in the 

/path/to/labels/train.txt

 file.
Following are the path to the images folder, the labels text file and the output name. Note that the output name should not exist prior to calling 

convert_imageset

 otherwise
you'll get a scary error message.

Other flags that might be useful:

--backend

 -
allows you to choose between an 

lmdb

 dataset
or 

levelDB

.

--gray

 -
convert all images to gray scale.

--help

 -
shows some help, see all relevant flags under Flags from tools/convert_imageset.cpp
https://github.com/BVLC/caffe/issues/1538#
Difference between "DataLayer" and "ImageDataLayer"