An imageclipper intended for using when creating OpenCV haar cascade sample images.
This is a fork of the imageclipper program initially written by Naotoshi Seo.
This fork was made initially for myself to be able to compile the program on OSX. Since I used CMake it was easy to make one build that works on all platforms.
From the original description:
It is often required to crop images manually fast for computer vision researchers to gather training and testing image sets, e.g., cropping faces in images.
This simple multi-platform (Windows, OSX and Linux were verified) software helps you to clip images manually fast.
Install requirements using your favorite package system or download the sources yourself.
$ sudo apt-get install cmake libopencv-dev libboost-all-dev
$ sudo port install cmake boost opencv
And then build:
$ git clone $ cd imageclipper $ mkdir build && cd build $ cmake .. # This should work in most cases.
(Optional) Or specifying your own builds of OpenCV or Boost.
$ cmake -DOpenCV_DIR=/path/to/opencv/build/ -DBOOST_ROOT=/path/to/boost ..
$ cmake --build . # Builds the program.
Has been tested with this verison on Windows 7, but should work with whatever you like: Boost 1.55.0 32-bit MSVC12 The default unpack location is c:\local\boost155_0, if you change it you need to know the path when building.
Download and unpack OpenCV and remember the path.
From the git bash console: (use cmd.exe if you prefer)
$ git clone $ cd imageclipper $ mkdir build && cd build $ cmake -DOpenCV_DIR=/path/to/opencv/build/ -DBOOST_ROOT=/c/local/boost_1_55_0/ .. $ cmake --build . --config Release $ start imageclipper.sln # If you want to open in Visual Studio instead.
The executable can then be found under build/bin/Release/imageclipper.exe