Classifying Grayscale Images using Pycaffe

If you have trained a model with 1-dimensional gray image, and want to classify another gray image, the following is the hack worked for me:

  1. copy the offical in $CAFFE_ROOT/python/
  2. specify input_dim as 1, 1, x, x in deploy.prototxt
  3. change all call to in to, False) because if you do not specify the second parameter as False, True will be used by default, the meaning of the second parameter is to tell load_image whether the image is color or gray, if it’s in color, then the returned image will have shape (width, height, 3) or (width, height, 4) depending on whether the alpha channel exists. If you specify False, the shape will be (width, height, 1) as you want.
  4. specify –channel_swap ‘0’ in python because this value is to reorder RGB to BGR, let’s say we have an image im, im is in numpy array format, and im.shape = (10, 10, 3), then caffe will do im = im[:, :, channel_wap] to swap channels, if you do not specify --channel_swap, it will be "2,1,0" by default, then in caffe, im = im[:, :, [2, 1, 0]], but the gray image’s shape is really (10, 10, 1) (if you follow the 2 step), so an index out of bounds exception will be raised. So just specify ‘0’ to --channel_swap, then caffe will run im = im[:, :, [0]], that’s fine.

then just use the official

Here is the gist of and worked for me.

Hope it will work for you too.

Installing frontalization 0.1.3: Face Frontalization in Unconstrained Images using MATLAB R2015b on Ubuntu 16.04

Library source:

The code uses the following dependencies. You MUST have these installed and available on the MATLAB path:

1. calib-1.0.1 function available from:
Installation: unzip calib.1.0.1, rename it to calib under frontaliztion home directory. Then rename calib_cv2.4.mexa64 to calib.mexa64

2. Facial feature detection functions. The code provides examples
of frontalization using different facial landmark detection methods. Currently supported are:
– SDM (default, used in paper; We don’t use this at the moment ),
– The facial feature detector of Zhu and Ramanan (We don’t use this at the moment)
DLIB detector (Our chosen method) .
– Any sparse (five-point) facial landmark detector. (We don’t use this at the moment)

3. OpenCV required by calib for calibration routines and some of the

detectors for cascase classifiers (We have already discussed about OpenCV installation in other blog posts. Check those.)

Frontalization set up:

1. Setup Dlib: Download from

tar jxvf dlib-19.1.tar.bz2
 cd dlib-19.1/
 cd examples/
cd build/
 cmake ..
 cmake --build . --config Release

2. Install dlib dependency (if required):

sudo apt-get install libboost-python1.58.0

3. Open demo.m

change line 86 from :
 detector = 'SDM'; to detector = 'dlib';

4. Open facial_feature_detection.m

Go to case ‘dlib’

change line 106 to following:
 Model3D = load('model3Ddlib'); % reference 3D points corresponding to dlib detections
 Model3D = Model3D.model_dlib;
and change line 111 to following:
 fidu_XY = load('dlib_xy.mat'); % load detections performed by Python script on current image
 fidu_XY = reshape(fidu_XY.lmarks,68,2);

5. Now open

Comment out line 7:
 #from Utils import HOME
Add the following two lines at the end: (Change image list as you like)
 lmarks, bboxes = get_landmarks(['test.jpg'])
 savemat('dlib_xy.mat', {'lmarks':lmarks})

6. Run the python file, this will create the dlib_xy.mat file

8. Now run demo.m , to see the frontalization demo result.

Installing OpenFace: an open source facial behavior analysis toolkit

Installation System: Ubuntu 16.04

1. Installing dependencies:

sudo apt-get update
sudo apt-get install build-essential
sudo apt-get install llvm
sudo gedit /etc/apt/sources.list
deb xenial main restricted
deb xenial main universe
sudo apt-get update
sudo apt-get install clang-3.7 libc++-dev libc++abi-dev
sudo apt-get install cmake
sudo apt-get install libopenblas-dev liblapack-dev
sudo apt-get install git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev
sudo apt-get install python-dev python-numpy libtbb2 libtbb-dev libjpeg-dev libpng-dev libtiff-dev libjasper-dev libdc1394-22-dev checkinstall
sudo unzip
cd opencv-3.1.0
mkdir build
cd build/
make -j4
sudo make install
sudo apt-get install libboost-all-dev

2. Installing OpenFace

git clone
cd OpenFace/
mkdir build

cd build/


3. Using from MATLAB  (Error Resolution: version `CXXABI_1.3.8′ not found)

To solve this issue, I found setting the LD_LIBARARY_PATH from the script works for me. Just add the following at the beginning of the script (change the path according to your system):

setenv('LD_LIBRARY_PATH', '/usr/lib/x86_64-linux-gnu/');

Sample script for running FaceLandmarkImg

from subprocess import call

# FaceLandmarkImg
# Single image analysis
# -f <filename> the image file being input, can have multiple -f flags
# -of <filename> location of output file for landmark points, gaze and action units
# -op <filename> location of output file for 3D landmark points and head pose
# -gaze indicate that gaze estimation should be performed
# -oi <filename> location of output image with landmarks
# -root <dir> the root directory so -f, -of, -op, and -oi can be specified relative to it
# -inroot <dir> the input root directory so -f can be specified relative to it
# -outroot <dir> the root directory so -of, -op, and -oi can be specified relative to it
# Batch image analysis
# -fdir <directory> - runs landmark detection on all images (.jpg and .png) in a directory, if the directory contains
# .txt files (image_name.txt) with bounding box (min_x min_y max_x max_y), it will use those for initialisation
# -ofdir <directory> directory where detected landmarks, gaze, and action units should be written
# -oidir <directory> directory where images with detected landmarks should be stored
# -opdir <directory> directory where pose files are output (3D landmarks in images together with head pose and gaze)

exe = "../build/bin/FaceLandmarkImg"

# f_param = './OpenFace/image_sequence/001.jpg'
# of_param = './OpenFace/python/img_output/001.txt'
# op_param = './OpenFace/python/img_output/001_3d.txt'
# oi_param = './OpenFace/python/img_output/001.jpg'
# call([exe, "-f", f_param, "-of", of_param, "-op", op_param, "-oi", oi_param])

fdir_param = './OpenFace/image_sequence/'
ofdir_param = './OpenFace/python/imgseq_output'
oidir_param = ofdir_param
opdir_param = ofdir_param

call([exe, "-fdir", fdir_param, "-ofdir", ofdir_param, "-oidir", oidir_param, "-opdir", opdir_param, "-wild"])


Installing Nvidia DIGITS on Ubuntu 16.04.1

(I am assuming caffe and pycaffe are already successfully installed. If not, check my previous post on that.)

1. Install dependencies

sudo apt-get install --no-install-recommends git graphviz gunicorn python-dev python-flask python-gevent python-h5py python-numpy python-pil python-protobuf python-scipy

2. Download and install source

git clone $DIGITS_HOME

3. Install python packages [Upgrade pip if needed]

sudo pip install -r $DIGITS_HOME/requirements.txt

4. Set an environment variable in ~/.bashrc so DIGITS knows where Caffe is installed:

export CAFFE_HOME=${HOME}/caffe

Remeber DIGITS will look for caffe binaries in ${HOME}/caffe/build/tools/ directory, so make sure the binaries are installed in that manner.

5. Start DIGITS server


6. The default location of the web app is


Installing Caffe with Cuda 7.5 on Ubuntu 16.04.1 (Nvidia GeForce GTX 960M GPU)

There are three main steps:

Step 1: Installing the Nvidia graphics driver

Step 2: Installing Cuda 7.5

Step 3: Installing Caffe, pycaffe

Step 1: Installing the Nvidia graphics driver


sudo apt-get purge nvidia*
sudo apt-get purge bumblebee* primus
sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt-get update
sudo apt-get install nvidia-352 nvidia-prime
sudo add-apt-repository -r ppa:bumblebee/stable

Step 2: Installing Cuda 7.5


1. Download from


sudo apt-get purge nvidia-cuda*
sudo sh --override
sudo reboot

Step 3: Installing Caffe, pycaffe

1. Enter the following commands

sudo apt-get update

sudo apt-get upgrade

sudo apt-get install -y build-essential cmake git pkg-config

sudo apt-get install -y libprotobuf-dev libleveldb-dev libsnappy-dev libhdf5-serial-dev protobuf-compiler

sudo apt-get install -y libatlas-base-dev

sudo apt-get install -y --no-install-recommends libboost-all-dev

sudo apt-get install -y libgflags-dev libgoogle-glog-dev liblmdb-dev

# (Python general)
sudo apt-get install -y python-pip

# (Python 2.7 development files)
sudo apt-get install -y python-dev
sudo apt-get install -y python-numpy python-scipy

# (OpenCV 2.4)
sudo apt-get install -y libopencv-dev

2. clone caffe repository

 git clone 

3. enter caffe folder and copy the Makefile.config

 cp Makefile.config.example Makefile.config 

4. Edit Makefile.config as follows

PYTHON_INCLUDE := /usr/include/python2.7 /usr/lib/python2.7/dist-packages/numpy/core/include


INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial

LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu /usr/lib/x86_64-linux-gnu/hdf5/serial

5. Apply the following commands (for Ubuntu 16.04)

find . -type f -exec sed -i -e 's^&quot;hdf5.h&quot;^&quot;hdf5/serial/hdf5.h&quot;^g' -e 's^&quot;hdf5_hl.h&quot;^&quot;hdf5/serial/hdf5_hl.h&quot;^g' '{}' \;

cd /usr/lib/x86_64-linux-gnu

sudo ln -s

sudo ln -s

6. Installing required python packages

cd python

for req in $(cat requirements.txt); do pip install $req; done

7. Go back to caffe installation folder (~/caffe e.g.), and run the installation commands:

make all

make test

make runtest

make pycaffe

8. In order to make the Python work with Caffe, open the file ~/.bashrc for editing in your favorite text editor. There, add the following line at the end of file:

export PYTHONPATH=/path/to/caffe-master/python:$PYTHONPATH

9. Your binaries will reside in .build_release folder


ERROR Resolution: “error — unsupported GNU version! gcc versions later than 5.3 are not supported!”

comment the #error line in file /usr/local/cuda/include/host_config.h

#if __GNUC__ > 5 || (__GNUC__ == 5 && __GNUC_MINOR__ > 3)

//#error — unsupported GNU version! gcc versions later than 5.3 are not supported!

#endif /* __GNUC__ > 5 || (__GNUC__ == 5 && __GNUC_MINOR__ > 1) */

ERROR Resolution: “/usr/include/string.h: In function ‘void* __mempcpy_inline(void*, const void*, size_t)’:
/usr/include/string.h:652:42: error: ‘memcpy’ was not declared in this scope
return (char *) memcpy (__dest, __src, __n) + __n;
Makefile:588: recipe for target ‘.build_release/cuda/src/caffe/util/im2col.o’ failed
make: *** [.build_release/cuda/src/caffe/util/im2col.o] Error 1

If you are compiling with make then edit Makefile and replace the line
NVCCFLAGS += -ccbin=$(CXX) -Xcompiler -fPIC $(COMMON_FLAGS)

Compiling congealReal: Unsupervised joint alignment of complex images

The c++ code provided for the project ‘Unsupervised joint alignment of complex images’ at is very useful application for unsupervised image alignment. But the Makefile is now outdated, but we can fix it by modifying it as following,

IFLAGS = `pkg-config --cflags opencv` -O3
LFLAGS = `pkg-config opencv --libs`

all: congealReal funnelReal

rm congealReal funnelReal

congealReal: congealReal.cpp
g++ $(IFLAGS) -o congealReal congealReal.cpp $(LFLAGS)

funnelReal: funnelReal.cpp
g++ $(IFLAGS) -o funnelReal funnelReal.cpp $(LFLAGS)

How to enable MatCaffe wrapper for Caffe

1. Go to caffe root directory, for me it is :

 $cd /home/naved/Desktop/caffe-master 

2. Make Matcaffe binaries:

 $ make all matcaffe 

3. Test matcaffe installation:

 $make mattest 

4. Now open ~/.bashrc and check if the following lines are there, otherwise find them in your (it can be different for your case) system and modify accordingly:

export PATH=$PATH:/usr/local/cuda/bin
export LD_LIBRARY_PATH=/usr/local/cuda/lib64/:$LD_LIBRARY_PATH
export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/

# Caffe Root
export CAFFE_ROOT=/home/guo/Desktop/caffe-master

5. Now open matlab FROM TERMINAL ,(NOT using launcher icon! This is important!) e.g.:


6. Run a test code to makes sure it is running properly. Before running the code make sure to add path to the caffe/matlab directory as shown in the example code below:


addpath('/home/guo/Desktop/caffe-master/matlab'); % MAKE SURE TO ADD THIS!

img = imread('ak.png');
img = single(img);

averageImage = [129.1863,104.7624,93.5940] ;

img = cat(3,img(:,:,1)-averageImage(1),...

img = img(:, :, [3, 2, 1]); % convert from RGB to BGR
img = permute(img, [2, 1, 3]); % permute width and height

model = 'VGG_FACE_16_deploy.prototxt';
weights = 'VGG_FACE.caffemodel';
net = caffe.Net(model, weights, 'test'); % create net and load weights

res = net.forward({img});
prob = res{1};

caffe_ft = net.blobs('fc7').get_data();

Installing Caffe and PyCaffe on Ubuntu 14.04

1. Install OpenCV 2.4.11

2. Install general dependencies

sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler sudo apt-get install --no-install-recommends libboost-all-dev

CUDA: Install via the NVIDIA package instead of apt-get to be certain of the library and driver versions. Install the library and latest driver separately; the driver bundled with the library is usually out-of-date. This can be skipped for CPU-only installation.

BLAS: install ATLAS by sudo apt-get install libatlas-base-dev or install OpenBLAS or MKL for better CPU performance.

Python (optional): if you use the default Python you will need to sudo apt-get install the python-dev package to have the Python headers for building the pycaffe interface.

3. remaining dependencies for 14.04

sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev

3. clone and install Caffe
git clone
cd /path/to/caffe/folder

mkdir build 
cd build 
cmake .. 
make all 
make install 
make runtest

4. Pycaffe installation

sudo apt-get install cython

sudo apt-get install python-skimage
sudo apt-get install python-pip
sudo pip install protobuf

make pycaffe

To import the caffe Python module after completing the installation, add the module directory to your $PYTHONPATH by export PYTHONPATH=/path/to/caffe/python:$PYTHONPATH

5. run python terminal and check by typing:
import caffe

OpenCV2 and g++: running your first program

Here’s a sample ‘helloworld’ program which takes an image and displays it:

#include "opencv2/opencv.hpp"

using namespace cv;

int main( int argc, char** argv )
// Read the source file
Mat src;
src = imread("fruits.jpg");

// Show the results
imshow( " ORIGINAL ", src );

return 0;

Save it is ‘helloworld.cpp’ using your favorite editor. Download a ‘fruits.jpg’ file from internet and save it in the same folder.

Comping the program using g++:

1. Use following command to compile:

g++ -Wall helloworld.cpp `pkg-config --libs --cflags opencv` -o helloworld

‘pkg-config –libs –cflags opencv’ adds all the necessary header file path and libraries.

2. run using ‘./helloworld’

OpenCV Installation Issue on Ubuntu: How to fix the ‘include : No such file or directory’ error

To install the latest OpenCV from git we follow the instructions on :

But if you stumble on the ‘include<Eigen/Eigenvalues>: No such file or directory’ error during make here’s how to fix that:

1. Install package

sudo apt-get install libeigen3-dev

2. Then change the cmake command to:


3. do ‘make’ and ‘sudo make install’ as instructed.