MENU
Edit
In Place Editing Advanced Editing

Face completion with a multi-output estimators

This example shows the use of multi-output estimator to complete images. The goal is to predict the lower half of a face given its upper half.

The first column of images shows true faces. The next columns illustrate how extremely randomized trees, k nearest neighbors, linear regression and ridge regression complete the lower half of those faces.

../_images/sphx_glr_plot_multioutput_face_completion_001.png

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
print(__doc__)
 
import numpy as np
import matplotlib.pyplot as plt
 
from sklearn.datasets import fetch_olivetti_faces
from sklearn.utils.validation import check_random_state
 
from sklearn.ensemble import ExtraTreesRegressor
from sklearn.neighbors import KNeighborsRegressor
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import RidgeCV
 
# Load the faces datasets
data = fetch_olivetti_faces()
targets = data.target
 
data = data.images.reshape((len(data.images), -1))
train = data[targets < 30]
test = data[targets >= 30]  # Test on independent people
 
# Test on a subset of people
n_faces = 5
rng = check_random_state(4)
face_ids = rng.randint(test.shape[0], size=(n_faces, ))
test = test[face_ids, :]
 
n_pixels = data.shape[1]
X_train = train[:, :np.ceil(0.5 * n_pixels)]  # Upper half of the faces
y_train = train[:, np.floor(0.5 * n_pixels):]  # Lower half of the faces
X_test = test[:, :np.ceil(0.5 * n_pixels)]
y_test = test[:, np.floor(0.5 * n_pixels):]
 
# Fit estimators
ESTIMATORS = {
    "Extra trees": ExtraTreesRegressor(n_estimators=10, max_features=32,
                                       random_state=0),
    "K-nn": KNeighborsRegressor(),
    "Linear regression": LinearRegression(),
    "Ridge": RidgeCV(),
}
 
y_test_predict = dict()
for name, estimator in ESTIMATORS.items():
    estimator.fit(X_train, y_train)
    y_test_predict[name] = estimator.predict(X_test)
 
# Plot the completed faces
image_shape = (64, 64)
 
n_cols = 1 + len(ESTIMATORS)
plt.figure(figsize=(2. * n_cols, 2.26 * n_faces))
plt.suptitle("Face completion with multi-output estimators", size=16)
 
for i in range(n_faces):
    true_face = np.hstack((X_test[i], y_test[i]))
 
    if i:
        sub = plt.subplot(n_faces, n_cols, i * n_cols + 1)
    else:
        sub = plt.subplot(n_faces, n_cols, i * n_cols + 1,
                          title="true faces")
 
 
    sub.axis("off")
    sub.imshow(true_face.reshape(image_shape),
               cmap=plt.cm.gray,
               interpolation="nearest")
 
    for j, est in enumerate(sorted(ESTIMATORS)):
        completed_face = np.hstack((X_test[i], y_test_predict[est][i]))
 
        if i:
            sub = plt.subplot(n_faces, n_cols, i * n_cols + 2 + j)
 
        else:
            sub = plt.subplot(n_faces, n_cols, i * n_cols + 2 + j,
                              title=est)
 
        sub.axis("off")
        sub.imshow(completed_face.reshape(image_shape),
                   cmap=plt.cm.gray,
                   interpolation="nearest")
 
plt.show()

Total running time of the script: (0 minutes 4.875 seconds)

Download Python source code: plot_multioutput_face_completion.py
Download IPython notebook: plot_multioutput_face_completion.ipynb
doc_scikit_learn
doc_scikit_learn
2025-01-10 15:47:30
Comments
Leave a Comment

Please login to continue.