Use float32 to reduce memory footprint.

[master-thesis.git] / Parasitemia / Parasitemia / KMedians.fs

diff --git a/Parasitemia/Parasitemia/KMedians.fs b/Parasitemia/Parasitemia/KMedians.fs
index 9b5d50f..f7f2e54 100644 (file)
--- a/Parasitemia/Parasitemia/KMedians.fs
+++ b/Parasitemia/Parasitemia/KMedians.fs
@@ -12,7 +12,7 @@ type Result = {
     median_fg: float
     d_fg: Image<Gray, float32> } // Euclidean distances of the foreground to median_fg.
 
-let kmedians (img: Image<Gray, float32>) (fgFactor: float) : Result =
+let kmedians (img: Image<Gray, float32>) : Result =
     let nbIteration = 3
     let w = img.Width
     let h = img.Height
@@ -30,8 +30,9 @@ let kmedians (img: Image<Gray, float32>) (fgFactor: float) : Result =
     let mutable fg = new Image<Gray, byte>(img.Size)
 
     for i in 1 .. nbIteration do
-        CvInvoke.Pow(img - median_bg, 2.0, d_bg)
-        CvInvoke.Pow(img - median_fg, 2.0, d_fg)
+        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>()
@@ -46,8 +47,6 @@ let kmedians (img: Image<Gray, float32>) (fgFactor: float) : Result =
         median_bg <- MathNet.Numerics.Statistics.Statistics.Median(bg_values)
         median_fg <- MathNet.Numerics.Statistics.Statistics.Median(fg_values)
 
-    CvInvoke.Sqrt(d_fg, d_fg)
-
     { fg = fg; median_bg = median_bg; median_fg = median_fg; d_fg = d_fg }