implemented k means

main.py

@@ -1,4 +1,70 @@
 from skimage.io import imread
+from scipy.stats import moment
+import numpy as np
+
+def find_nearest_point(data, target):
+    idx = np.array([calc_distance(p, target) for p in data]).argmin()
+    return data[idx]
+
+
+def centeroidnp(arr):
+    length = arr.shape[0]
+    sum_x = np.sum(arr[:, 0])
+    sum_y = np.sum(arr[:, 1])
+    sum_z = np.sum(arr[:, 2])
+    return sum_x/length, sum_y/length, sum_z/length
+
+
+def calc_distance(x, y):
+    return np.absolute(np.linalg.norm(x - y))
+
+
+def k_means(data, count):
+    # Pick n random points to start
+    index = np.random.choice(data.shape[0], count, replace=False)
+    means = data[index]
+    data = np.delete(data, index, axis=0)
+
+    distance_delta = 100
+    means_distance = 0
+    while distance_delta > 0.1:
+        print(f"new iteration, distance moved: {distance_delta}")
+        # Initialize cluster map
+        clusters = {}
+        for m in means:
+            clusters[str(m)] = []
+
+        # Find closest mean to each point
+        for point in data:
+            closest = find_nearest_point(means, point)
+            clusters[str(closest)].append(point)
+
+        # Find the centeroid of each mean
+        new_means = []
+        previous_distance = means_distance
+        means_distance = 0
+        for mean in means:
+            mean_key = str(mean)
+            # Clean up the results a little bit
+            clusters[mean_key] = np.stack(clusters[str(mean)])
+            # Calculate new mean
+            raw_mean = centeroidnp(clusters[mean_key])
+            nearest_mean_point = find_nearest_point(data, raw_mean)
+            means_distance = means_distance + calc_distance(mean, nearest_mean_point)
+            new_means.append(nearest_mean_point)
+        means_distance = means_distance / float(count)
+        distance_delta = abs(previous_distance - means_distance)
+        means = np.stack(new_means)
+    print(means)
+
+
+
+
 
 im = imread("image.png")
-print(im)
+starting_resolution = im.shape
+raw_pixels = im.reshape(-1, 3)
+
+colors = [[45, 85, 255], [0, 181, 204], [243, 225, 107]]
+
+k_means(raw_pixels, len(colors))