/frok-server

face recognition odnoklassniki - server side

Primary LanguageC++

frok-server

face recognition odnoklassniki - server side

How to install in Linux:

Requirements:

  1. Ubuntu 12.04 or higher (Linux Mint is made of the same core as ubuntu 12.04 so everything will be similar for Mint)
    Installing Ubuntu:

    1. Download ubuntu from http://www.ubuntu.com/download/desktop
      2a) VirtualBox:
      2.1. Download Oracle VirtualBox from https://www.virtualbox.org/wiki/Downloads
      2.2. Follow instruction http://www.psychocats.net/ubuntu/virtualbox
      2b) PC:
      2.1. Make bootable usb or dvd. Tool for making bootablie usb - http://unetbootin.net/download-unetbootin/
      2.2. Make some free space for installation (RAM size * 2 (for swap area) + ~20GB)
      how to "diskmanagement in Windows 7" http://airstream.pro/2011/09/23/how-to-use-disk-management-on-windows7/
      2.3. Boot ubuntu image (Check your BIOS->BOOT settings for booting usb)
      2.4. Follow installation tips. Choose 2 * RAM size of free space for swap area and rest of free space for being mounted to "/"
      2.5. Installation complete. Use google for help.

  2. compilation tools 4.8.2 or higher
    sudo apt-get update
    sudo apt-get dist-upgrade #should be used for "fresh" Ubuntu 14.04 Desktop only
    sudo apt-get install gcc
    sudo apt-get install g++

  3. git 1.9.1 or higher
    sudo apt-get install git

  4. (!!!FOR DEVELOPERS ONLY!!!) QT creator 4.8 or higher (I use 4.8 on Linux Mint and 5.3 on Ubuntu)
    3.1. Download "Qt Online Installer for Linux" from http://qt-project.org/downloads
    3.2. Run downloaded script, choose all tools and gcc only installation (we don't need cross-platform compilers). Destination folder should be /opt
    3.3. Run "/opt/Qt/Tools/QtCretor/bin/qtcreator" to launch qtcreator

  5. cmake
    sudo apt-get install cmake

  6. GTK+2.0 or higher
    apt-get install libgtk2.0-dev

  7. opencv-2.4.9
    3.1. download .zip from http://sourceforge.net/projects/opencvlibrary/files/opencv-unix/2.4.9/opencv-2.4.9.zip/download
    3.2. Unpack archieve to /opt/
    sudo unzip opencv-2.4.9.zip -d /opt/
    3.3. Make opencv from sources:
    cd /opt/opencv-2.4.9/
    mkdir static
    sudo cmake -D CMAKE_BUILD_TYPE=RELEASE -D BUILD_SHARED_LIBS=NO -D CMAKE_INSTALL_PREFIX=/opt/opencv-2.4.9/static -D WITH_QT=NO ./
    sudo make
    sudo make install
    3.4. Set opencv paths
    gedit ~/.bashrc
    add 'export LD_LIBRARY_PATH=/opt/opencv-2.4.9/static/lib/:$LD_LIBRARY_PATH' to the end of file, save and exit
    reboot

Issues:
1. undefined reference to symbol 'pthread_mutexattr_settype@@GLIBC_2.2.5'
Solution:
change '-L/usr/lib/x86_64-linux-gnu' in *.mk files to your linux-gnu path
Use this command to see your linux-gnu path:
find /usr/lib | grep linux-gnu
2. When running from QtCreatod: error while oading shared libraries: libopencv_contrib.so ...
Solution:
Open Projects tab->build->Build Environment (on the bottom)
add environment variable:
name - LD_LIBRARY_PATH, value - /opt/opencv-2.4.9/install/lib/

Build steps: [TBD] more info + QtCreator info

  1. in terminal go to FaceDetection library cd ./build for details: make help for debug: make CFG=debug build make CFG=debug clean make CFG=debug rebuild for release: make CFG=release build etc for debug withoud daemon mode: make CFG=debug CCFLAG+=-DNO_DAEMON build etc

Run steps: FrokAgentApp needs some root capabilities so it should be run with sudo run './FrokAgentApp' for FrokAgentApp parameters usage First parameter is #CPU that agent will obtain (-1 to run on all available CPUs) Second is port number (agent can be accessed via any ipV4 address on local machine on spicified port) example: "sudo ./FrokAgentApp 3 27015" This will start FrokAgentApp that will obtain CPU number 3 and port number 27015 example: "sudo ./FrokAgentApp -1 3000" This will start 'N' FrokAgentApps where N is number of available CPUs. First agent will obtain #CPU1 and port 3000, second #CPU2 and port 3001 etc Running FrokAgentApp in release with sudo will create /etc/frok/frok.conf with some default parameters (they have to be changed with your system configs) and /var/log/frok.log

Configuration:

  1. Create file /etc/frok/frok.conf and add config parameters (see example below)

    sudo touch /etc/frok/frok.conf sudo chmod 664 /var/log/frok.log sudo vi /etc/frok/frok.conf

Example of frok.conf (user = zda):

OUTPUT_FILE = /var/log/frok.log PHOTO_BASE_PATH = /home/zda/faces/ TARGET_PHOTOS_PATH = /home/zda/faces/

  1. Create log file: sudo touch /var/log/frok.log sudo chmod 666 /var/log/frok.log

How to install in Windows:

  1. Donwload and install to C:\ opencv from http://sourceforge.net/projects/opencvlibrary/files/opencv-win/2.4.8/opencv-2.4.8.exe/download
  2. Download dll-installer https://www.dropbox.com/s/ue6ub71k53tvxwi/opencv_inst.rar
  3. Launch dll-installer as admin (MS Visual studiio 2008+ must be installed before launching) with following parameters: <32/64 version of Windows> Example: D:\dll-installer D:\ 64 13
  4. Pull this repository.
  5. Check following properties of project: C/C++: Additional Include Directories = C:\opencv\build\include\opencv; C:\opencv\build\include; Linker: Additional Library Directories = C:\opencv\build\x64\vc12\lib;C:\opencv\build\include\opencv;C:\opencv\build\include; Additional Dependances = opencv_core248d.lib opencv_imgproc248d.lib opencv_highgui248d.lib opencv_video248d.lib opencv_objdetect248d.lib opencv_calib3d248d.lib opencv_features2d248d.lib opencv_nonfree248d.lib opencv_flann248d.lib opencv_legacy248d.lib kernel32.lib user32.lib gdi32.lib winspool.lib comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib uuid.lib odbc32.lib odbccp32.lib Import Library = opencv_core248d.lib opencv_imgproc248d.lib opencv_highgui248d.lib opencv_video248d.lib opencv_objdetect248d.lib opencv_calib3d248d.lib opencv_features2d248d.lib opencv_nonfree248d.lib opencv_flann248d.lib opencv_legacy248d.lib $(TargetDir)$(TargetName).lib,user32.lib,kernel32.lib,user32.lib,gdi32.lib,winspool.lib,comdlg32.lib,advapi32.lib,shell32.lib,ole32.lib,oleaut32.lib,uuid.lib,odbc32.lib,odbccp32.lib P.S. note that Additional Library Directories should consider path to libraries with current configuration (C:\opencv\build\x86... for 64 bin, C:\opencv\build\x86 for 32)
  6. Download and install Doxygen http://www.stack.nl/~dimitri/doxygen/download.html
  7. Run as admin in cmd "REG ADD HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Policies\System /v EnableLUA /t REG_DWORD /d 0 /f"
  8. Указать пути к папкам где будут сохраняться фотки и модели юзеров в файле FaceDetectionLib/activities.h Например: define ID_PATH "Z:\frok\" define TARGET_PATH "Z:\frok\1\"
  9. Build project and see FaceDetection_manual for additional information