mc2s 0.1

Package to apply MCR-ALS with or without Chan-vese constraint on CARS images

MC2S

MC2S provides a script to apply MCR-ALS and projection using least squares regression based on CARS data package.

Datasets

Datasets can be provided on request at [e-mail]

Usage

To launch the main script :

carsdata

You can also provide directly the configuration file with :

carsdata -j path/to/configuration/file.json

Configuration files are in JSon format, examples of configuration are available in the configs folder. Configuration attributes are the same as objects constructors in the source code. Hence, mc2.json contains:

{
  "data" : ["Path/to/file"],
  "analyzer" : 
  {
      "MCR" :
      {
        "output_dim" : 5,
        "guesser" :
        {
          "Simplisma" :
          {
            "simp_err" : 5
          }
        },
        "c_regr" : 
        {
          "NNLS" : {}
        },
        "c_constr" : 
        {
          "ChanVeseConstraint":
          {
          "nu": 0,
          "lambda1": 1,
          "lambda2": 1,
          "mu": 0.35
          },
          "NormConstraint" :
          {
            "axis" : 1
          }
        },
        "st_regr" : 
        {
          "NNLS" : {}
        },
        "st_constr" : {}
      }
  },
  "vspan" : [
    {
      "begin" : 3180,
      "end" : 3200,
      "color" : "cyan"
    },
    {
      "begin" : 3046,
      "end" : 3066,
      "color" : "green"
    },
    {
      "begin" : 2997,
      "end" : 3017,
      "color" : "red"
    },
    {
      "begin" : 2910,
      "end" : 2930,
      "color" : "green"
    },
    {
      "begin" : 2834,
      "end" : 2854,
      "color" : "red"
    }
  ],
  "spectra_colors" : [
    "darkmagenta",
    "mediumvioletred",
    "navy",
    "teal",
    "saddlebrown"
  ]
}