clustering

clustering.py

# pip install scikit-learn

import numpy as np
from sklearn import metrics
from sklearn.cluster import DBSCAN
import matplotlib.pyplot as plt
from sklearn.datasets import make_blobs
from sklearn.preprocessing import StandardScaler

centers = [[1, 1], [-1, -1], [1, -1], [-1,1]] #* (640,480)
X, labels_true = make_blobs(
    n_samples=750, centers=centers, cluster_std=0.2, random_state=0
)

X = StandardScaler().fit_transform(X)




db = DBSCAN(eps=0.3, min_samples=10).fit(X)
labels = db.labels_

# Number of clusters in labels, ignoring noise if present.
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
n_noise_ = list(labels).count(-1)

print("Estimated number of clusters: %d" % n_clusters_)
print("Estimated number of noise points: %d" % n_noise_)

plt.scatter(X[:, 0], X[:, 1])
plt.show()

fingers.txt

306 140
304 140
148 223
233 380
357 161
170 238
153 283
407 215
304 140
146 220
169 237
395 251
356 160
240 392
410 214
306 139
231 378
407 214
139 226
330 122
397 251
396 253
142 224
427 276
197 383
152 284

import numpy as np
from sklearn.cluster import DBSCAN
import matplotlib.pyplot as plt
from matplotlib.patches import Circle

points =  [
    [306, 140],
    [304, 140],
    [148, 223],
    [233, 380],
    [357, 161],
    [170, 238],
    [153, 283],
    [407, 215],
    [304, 140],
    [146, 220],
    [169, 237],
    [395, 251],
    [356, 160],
    [240, 392],
    [410, 214],
    [306, 139],
    [231, 378],
    [407, 214],
    [139, 226],
    [330, 122],
    [397, 251],
    [396, 253],
    [142, 224],
    [427, 276],
    [197, 383],
    [152, 284]
]


X = np.array(points)

canvas_width = 640
canvas_height = 480

dbscan = DBSCAN(eps=60, min_samples=3)
dbscan.fit(X)

labels = dbscan.labels_

# fig, ax = plt.subplots(figsize=(canvas_width / 100, canvas_height / 100))

# ax.set_aspect('equal', adjustable='box')

# ax.scatter(X[:, 0], X[:, 1], c=labels, cmap='viridis', alpha=1)

# for i, point in enumerate(X):
#     if True: # or labels[i] == -1:
#         eps_circle = Circle((point[0], point[1]), dbscan.eps, color='gray', fill=False)
#         ax.add_patch(eps_circle)

# ax.set_xlim(0, canvas_width)
# ax.set_ylim(0, canvas_height)

# ax.set_title(f'DBSCAN Clustering (eps={dbscan.eps})')
# ax.set_xlabel('X-axis')
# ax.set_ylabel('Y-axis')

# plt.show()

cluster_points = {}

cluster_points = {cluster: [point for point_idx, point in enumerate(points) if labels[point_idx] == cluster] for cluster in set(labels)}

for cluster, points in cluster_points.items():
    print(f"Cluster {cluster}: {points}")