ImagesToLARModel.jl

Installation

Pkg.clone("git://github.com/sadan91/ImagesToLARModel.jl.git")

This module require an installation of python with scipy and numpy for the denoising filter

Use

Data preparation

using(ImagesToLARModel)
prepareData(<JSON-configuration-file-path>)

or:

using(ImagesToLARModel)
prepareData(<Input directory>, <Output directory> [, <crop>, <noise_shape>, <threshold>])

This is an example of a valid JSON configuration file:

{
  "inputDirectory": "Path of the input directory",
  "outputDirectory": "Path of the output directory",
  "crop": List with values for images resizing (they can be extended or cropped),
  "noise_shape": A number which indicates the intensity of the denoising
                 filter (0 if you want to disable denoising),
  "threshold": A number indicating the chosen threshold for data
  "threshold3d": A number indicating the chosen threshold for the
                 three-dimensional filter (0 if you want to disable this filter)
  "zDim": A number indicating the number of images computed at once from the
          three-dimensional filter (0 if you want to take the entire stack)
}

For example we can write:

{ 
    "inputDirectory": "/home/juser/IMAGES/",
    "outputDirectory": "/home/juser/OUTPUT/",
    "crop": [[1,800],[1,600],[1,50]],
    "noise_shape": 0,
    "threshold": 8,
    "threshold3d": 100,
    "zDim":0
}

These are the accepted parameters:

• inputDirectory: Directory containing the stack of images
• outputDirectory: Directory containing the output
• crop: Parameter for images resizing (they can be extended or cropped)
• noise_shape: Intensity of the denoising filter for images (0 if you want to disable it)
• threshold: Set a threshold for raw data. Pixels under that threshold will be set to black, otherwise they will be set to white. If threshold is not specified, segmentation will be done using a clustering algorithm
• threshold3d: set a threshold for the three-dimensional filter
• zDim: set the number of images computed at once from the three-dimensional filter

Data conversion

using(ImagesToLARModel)
convertImagesToLARModel(<JSON-configuration-file-path>)

or:

using(ImagesToLARModel)
convertImagesToLARModel(<Input directory>, <Output directory>, <Border x>, <Border y>, <Border z>[, <DEBUG_LEVEL>, <parallelMerge>])

This is an example of a valid JSON configuration file:

{
  "inputDirectory": "Path of the input directory",
  "outputDirectory": "Path of the output directory",
  "nx": border x,
  "ny": border y,
  "nz": border z,
  "DEBUG_LEVEL": julia Logging level
  "parallelMerge": "true" or "false",
}

For example we can write:

{
    "inputDirectory": "/home/juser/IMAGES/",
    "outputDirectory": "/home/juser/OUTPUT/",
    "nx": 2,
    "ny": 2,
    "nz": 2,
    "DEBUG_LEVEL": 2
}

These are the accepted parameters:

• inputDirectory: Directory containing the stack of images
• outputDirectory: Directory containing the output
• nx, ny, nz: Border dimensions
• DEBUG_LEVEL: Debug level for Julia logger. It can be one of the following:
  • DEBUG (1 for JSON configuration file)
  • INFO (2 for JSON configuration file)
  • WARNING (3 for JSON configuration file)
  • ERROR (4 for JSON configuration file)
  • CRITICAL (5 for JSON configuration file)
• parallelMerge: Choose if you want to merge model files using a distribuite algorithm or not (experimental)