def run(self):
worstCluster = self.dataset
while len(self.means) < self.k:
if isinstance(worstCluster, Dataset):
worstDataset = worstCluster
elif isinstance(worstCluster, Mean):
worstDataset = worstCluster.coveredDataset
bisection = Kmeans(dataset=worstDataset,
k=2,
trials=self.trials,
maxRounds=self.trials,
key=self.key)
bisection.run()
bisectionSolution = bisection.getBestSolution()
self.means += bisectionSolution.means
worstCluster = max(self.means,
key=lambda m: m.getMeanSquaredError())
# if the number of means is not enouth remove the worst cluster
# found to bisect it in the next iteration.
if len(self.means) < self.k:
self.means.remove(worstCluster)
self.setMeanSquaredError()
def get_colors_from_image(filepath, number_of_colors):
# print 'processing '+filepath
kmeans = Kmeans(number_of_colors, 6, 5, 200) # default 6,5,200
im = Image.open(filepath)
colors = kmeans.run(im)
im.close()
return colors
def get_color_from_image(filepath):
# print 'processing '+filepath
kmeans = Kmeans(1)
im = Image.open(filepath)
color = kmeans.run(im)[0]
im.close()
return color
def run(self):
self.bisecting = [
BisectingKmeans(dataset=self.dataset,
k=i,
trials=self.trials,
maxRounds=self.maxRounds) for i in self.krange
]
self.normal = [
Kmeans(dataset=self.dataset,
k=i,
trials=self.trials,
maxRounds=self.maxRounds) for i in self.krange
]
for i, j in enumerate(self.krange):
self.bisecting[i].run()
self.normal[i].run()
