From inside the experiments folder, it suffices to run the following command

julia experiments.jl

To install, clone this repository via

git clone https://github.com/dprn/WCvsLHE.git

In order to import the package, do as usual

import Pkg
Pkg.activate(path_to_WCvsLHE_folder)

using WCvsLHE

The only exported functions are wc (implementing the WC contrast enhancement) and lhe (for the Local Histogram Equalization). There is a special type to store parameters, Params, which should be used to call both these functions. It is defined as follows:

struct Params
    σμ :: Float64 # std deviation of local mean average
    σw :: Float64 # std deviation of interaction kernel
    λ :: Float64 # attachment to the data
    M :: Float64 # the interaction weight is ν=1/(2M)
end

The functions wc and lhe take the following parameters:

• I0: Input image
• p: Parameters

Moreover, they have the following keyword arguments:

• model: Can be either :planar for the non-oriented version or :cortical for the oriented one (defaults to :planar)
• θs: Number of orientations to consider in case of a :cortical algorithm (defaults to 30)
• α: Sigmoid parameter, used only in wc (defaults at 5)
• verbose: If true the algorithm returns a Result instance, which saves the result, the number of iterations, and the relative tolerance attained. If false we only return the final image. (Defaults at false)
• max_iter: Maximal number of iterations.
• tolerance: Tolerance at which to stop.

using Images
img = load("myimage.png")

p = Params(2, 5, .7, .7)

wc(img, p, algo_type = :planar)
lhe(img, p, algo_type = :planar)
wc(img, p, algo_type = :cortical)
lhe(img, p, algo_type = :cortical)