module KMedians

open System.Collections.Generic
open System.Drawing

open Emgu.CV
open Emgu.CV.Structure

type Result = {
    fg: Image<Gray, byte>
    median_bg: float
    median_fg: float
    d_fg: Image<Gray, float32> } // Euclidean distances of the foreground to median_fg.

let kmedians (img: Image<Gray, float32>) : Result =
    let nbIteration = 3
    let w = img.Width
    let h = img.Height

    let min = ref [| 0.0 |]
    let minLocation = ref <| [| Point() |]
    let max = ref [| 0.0 |]
    let maxLocation = ref <| [| Point() |]
    img.MinMax(min, max, minLocation, maxLocation)

    let mutable median_bg = (!max).[0] - ((!max).[0] - (!min).[0]) / 4.0
    let mutable median_fg = (!min).[0] + ((!max).[0] - (!min).[0]) / 4.0
    use mutable d_bg = new Image<Gray, float32>(img.Size)
    let mutable d_fg = new Image<Gray, float32>(img.Size)
    let mutable fg = new Image<Gray, byte>(img.Size)

    for i in 1 .. nbIteration do
        d_bg <- img.AbsDiff(Gray(median_bg))
        d_fg <- img.AbsDiff(Gray(median_fg))

        CvInvoke.Compare(d_fg, d_bg, fg, CvEnum.CmpType.LessThan)

        let bg_values = List<float>()
        let fg_values = List<float>()

        for i in 0 .. h - 1 do
            for j in 0 .. w - 1 do
                if fg.Data.[i, j, 0] > 0uy
                then fg_values.Add(float img.Data.[i, j, 0])
                else bg_values.Add(float img.Data.[i, j, 0])

        median_bg <- MathNet.Numerics.Statistics.Statistics.Median(bg_values)
        median_fg <- MathNet.Numerics.Statistics.Statistics.Median(fg_values)

    { fg = fg; median_bg = median_bg; median_fg = median_fg; d_fg = d_fg }