/MaximaFinderC

Maxima finding algorithm in ImageJ/Fiji re-implemented in C++

Primary LanguageC++

MaximaFinderC: Maxima finding algorithm in ImageJ/Fiji re-implemented in C++

This is a re-implementation of the java plugin written by Michael Schmid and Wayne Rasband for ImageJ. The original java code source can be found in: https://imagej.nih.gov/ij/developer/source/ij/plugin/filter/MaximumFinder.java.html.

This script is simply a re-implementation of the ImageJ find maxima algorithm writing with pure C++ code. Therefore cython compiler is dropped for protential integration which written in C++.

The outcome is exactly consistent with the ImageJ/Fiji, and the calculation speed is in the same level of MaximaFinder, which use cython compiler to realize the same function.

How To Call

The re-implementation of the maxima finding function can either be called through C++ and Python for any possible usage.

For C++:

  • Directly run ./cfindmaxima.cpp under the root directory.

For Python:

  • Run ./main.py under the root directory.

  1. Use dll to call

    The interface exposed to 'ctypes.cdll.LoadLibrary' is called 'find_maxima', therefore, use command:

    g++ -o cfindmaxima.so -shared -fPIC cfindmaxima.cpp

    under the root directory to compile ./cfindmaxima.cpp, which is the same file for c++ calls.

    ⚠️ This method is unable to back pass array of any type, as a result, maxima points founded can only be persisted in ./points.csv files, which is in the same form of ImageJ/Fiji output type: List.

  2. (Recommend) Use pybind11 to call

    pybind11 offer the possibility to pass np.array type to c++ function, and back pass the np.array type outcome for postprocess.

    To compile it, use:

    cd ./pybind11_call/build/
cmake ../../  # or use CMake to compile
make

    the cfindmaxima.cp38-win_amd64.pyd will be generated under ./pybind11_call/build/. Then use the same python environment in ./main.py and call as:

    from pybind11_call.build import cfindmaxima
x, y = cfindmaxima.find_maxima(phantom, 10)

    x, y is in the np.array type.

The main function used in python is the same as 'analyzeAndMarkMaxima' in ./cfindmaxima.cpp, different interface lead to different calls.