From: Tomas Musil <tomik.musil@gmail.com>
Date: Fri, 18 Jul 2014 15:35:52 +0000 (+0200)
Subject: scale colors
X-Git-Url: http://git.tomasm.cz/imago.git/commitdiff_plain/f19938820e152bbd68f2ed3f8aa3f7d7f182ca06

scale colors
---

diff --git a/src/intrsc.py b/src/intrsc.py
index 68b18a0..73de142 100644
--- a/src/intrsc.py
+++ b/src/intrsc.py
@@ -57,13 +57,13 @@ def board(image, lines, show_all, do_something):
     board_raw = sum(board_raw, [])
 
     ### Show color distribution
-    luma = [s[0] for s in board_raw]
-    saturation = [s[1] for s in board_raw]
 
     if show_all:
         import matplotlib.pyplot as pyplot
         import Image
         fig = pyplot.figure(figsize=(8, 6))
+        luma = [s[0] for s in board_raw]
+        saturation = [s[1] for s in board_raw]
         pyplot.scatter(luma, saturation, 
                        color=[(s[2][0]/255.,
                                s[2][1]/255.,
@@ -77,7 +77,15 @@ def board(image, lines, show_all, do_something):
         image_p = Image.fromstring('RGB', size, buff, 'raw')
         do_something(image_p, "color distribution")
 
-    clusters = k_means.cluster(3, 2,zip(zip(luma, saturation), range(len(luma))),
+    max_s0 = max(s[0] for s in board_raw)
+    min_s0 = min(s[0] for s in board_raw)
+    norm_s0 = lambda x: (x - min_s0) / (max_s0 - min_s0)
+    max_s1 = max(s[1] for s in board_raw)
+    min_s1 = min(s[1] for s in board_raw)
+    norm_s1 = lambda x: (x - min_s1) / (max_s1 - min_s1)
+    color_data = [(norm_s0(s[0]), norm_s1(s[1])) for s in board_raw]
+
+    clusters = k_means.cluster(3, 2,zip(color_data, range(len(color_data))),
                                [[0., 0.5], [0.5, 0.5], [1., 0.5]])
 
     if show_all:
diff --git a/src/k_means.py b/src/k_means.py
index f429824..119f576 100644
--- a/src/k_means.py
+++ b/src/k_means.py
@@ -4,11 +4,11 @@ import random
 
 def cluster(k, d, data, i_centers=None):
     """Find *k* clusters on *d* dimensional *data*."""
-    borders = [(min(p[0][i] for p in data), max(p[0][i] for p in data))
-               for i in range(d) ]
     if i_centers:
         old_centers = i_centers
     else:
+        borders = [(min(p[0][i] for p in data), max(p[0][i] for p in data))
+               for i in range(d)]
         old_centers = [[(h - l) * random.random() + l for (l, h) in borders]
                for _ in range(k)]
     clusters, centers = next_step(old_centers, data)