INSTALL

install Matio library (https://sourceforge.net/projects/matio/)
	Matio is a C library to read/save files in .mat format
	If Matlab and zlib are installed run the following commands:
	./configure
	sudo make install
	Sometimes Matio lacks executable permission for configure after unzipping, fix by running
	chmod u+x configure
	May also want to install using sudo checkinstall to create a .deb file to facilitate uninstallation

Install OpenCV 3.0 (optional)
	if at runtime there is a GTK error about GTK2 and GTK3 symbols conflicting
	recompile opencv with -DWITH_GTK=OFF

CMake command to use with this OpenCV guide: http://www.pyimagesearch.com/2016/10/24/ubuntu-16-04-how-to-install-opencv/
 cmake -D CMAKE_BUILD_TYPE=RELEASE \
 	-D WITH_GTK=OFF -D WITH_GTK_2_X=OFF -D WITH_QT=ON \
 	-D BUILD_JPEG=ON \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D INSTALL_PYTHON_EXAMPLES=OFF \
    -D INSTALL_C_EXAMPLES=OFF ..    
    -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-3.3.0/modules \
    -D PYTHON_EXECUTABLE=~/.virtualenvs/cv/bin/python \
    -D BUILD_EXAMPLES=ON ..
    
These flags -D WITH_GTK=OFF -D WITH_GTK_2_X=OFF -D WITH_QT=ON prevent OpenCV against linking to libprotobuf, in case there is 
an older version of the library installed via apt-get

Install qt5 and qttools5-dev-tools (qt designer in this package)

Install cmake >3.1 since it has better support for CUDA and automatic qt5 ui building
If cmake cannot find Qt5 libraries, add the following line to .bashrc with the
path to qt5 install directory:
export CMAKE_PREFIX_PATH=/usr/lib/x86_64-linux-gnu/qt5/:$CMAKE_PREFIX_PATH;


if caffe throws segfault when trying to use Python layer CustomData:

	install a newer version of protobuf > 3.1
	follow instructions here:
	https://github.com/google/protobuf/blob/master/src/README.md
	install protobuf for python3
		sudo install python3-pip
		 pip3 install --user --upgrade protobuf==3.1.0.post1
			
	Then recompile caffe with python3 enabled
	if while compiling caffe -lboost-python3 is not found
	create a symbolic link from the libboost-python-py34.so located in the /usr/lib/x86_64-linux-gnu/
		sudo ln -s /usr/lib/x86_64-linux-gnu/libboost_python-py35.so /usr/lib/x86_64-linux-gnu/libboost_python3.so
	
	sudo pip3 install scikit-image


If at any point during the compilation you get this error
	fatal error: caffe/proto/caffe.pb.h: No such file or directory
Run the following commands in the caffe root directory:
	protoc src/caffe/proto/caffe.proto --cpp_out=.
	mkdir include/caffe/proto
	mv src/caffe/proto/caffe.pb.h include/caffe/proto
	
	
If there is out of memory error at runtime
install cuDNN v5 (May 27, 2016) for CUDA 8.0
	download https://developer.nvidia.com/rdp/cudnn-archive
	unpack archive and copy cudnn header file from include/ directory to /usr/local/include/
	and libraries from lib64 directory to /usr/local/lib (or any other location, which is on the PATH)
recompile caffe with USE_CUDNN flag 
	
	
add caffe python library and OpenSALICON folder to PYTHONPATH
	export PYTHONPATH=/home/yulia/Documents/caffe/python:/home/yulia/Documents/ONR-TD/fixation_code/FixControlNew/contrib/OpenSALICON/:$PYTHONPATH
	

install gtest library
sudo apt-get install libgtest-dev 
this installs the source, build it by executing these commands:
cd /usr/src/gtest
sudo cmake CMakeLists.txt
sudo make

install fftw3 library
sudo apt-get install libfftw3-dev

Building FixControl using cmake:
to enable OpenSALICON support, build with option WITH_SALICON
with no options, the original version of FixControl with face templates will be built

mkdir build
cd build
cmake -DWITH_SALICON=ON ../
make

set CUDA_VISIBLE_DEVICES environment variable for cuda, also works for running two instances of FixControl on two GPUs


To use with DeepGazeII
install tensorflow
pip3 install tensorflow
pip3 install tensorflow-gpu
pip3 install pandas
pip3 install seaborn