teratail
回答率
85.48%
質問するログイン新規登録
トップに関する質問画像データの特徴量を抽出し機械学習用のデータセットを作りたい

編集履歴

質問編集履歴

3

文章表現

2021/04/13 21:25

投稿

shishi_maru440
shishi_maru440

スコア38

test CHANGED
@@ -1 +1 @@
1
- 画像データをnumpy配列にし機械学習用のデータセットを作りたい
1
+ 画像データの特徴量を抽出し機械学習用のデータセットを作りたい
test CHANGED
@@ -4,9 +4,9 @@
4
4
 
5
5
  自前の画像データから機械学習用のデータセットを作成し学習させたいのですが、
6
6
 
7
- 下記エラーにより実行できません。画像データを配列にするときの方法が不適切では
7
+ モデルをfitさせた時に下記エラーにより実行できません。画像データを配列にする
8
-
8
+
9
- と考えているのですが、解決方法をご存知の方ご教示ください。
9
+ ときの方法が不適切ではと考えているのですが、解決方法をご存知の方ご教示ください。
10
10
 
11
11
 
12
12
 
@@ -52,8 +52,6 @@
52
52
 
53
53
 
54
54
 
55
- #1.anotation_results_form_check.ipynbで作成したdata2.csvを読み込み
56
-
57
55
  df2 = pd.read_csv("data2.csv")
58
56
 
59
57
  df2.shape
@@ -108,7 +106,7 @@
108
106
 
109
107
 
110
108
 
111
- #targetにsmileのカテゴリを代入
109
+ **#targetにsmileのカテゴリを代入
112
110
 
113
111
  smiles =df_join["smile"]
114
112
 
@@ -122,7 +120,7 @@
122
120
 
123
121
  target.append(data)
124
122
 
125
- y = np.array(target)
123
+ y = np.array(target)**
126
124
 
127
125
  y
128
126
 
@@ -162,6 +160,26 @@
162
160
 
163
161
  x = np.array(photo_array)
164
162
 
163
+ x[0]
164
+
165
+ ```
166
+
167
+ ```OUT
168
+
169
+ array([[217, 217, 217, ..., 188, 196, 203],
170
+
171
+ [217, 217, 217, ..., 188, 196, 203],
172
+
173
+ [217, 217, 217, ..., 188, 196, 204],
174
+
175
+ ...,
176
+
177
+ [ 4, 4, 4, ..., 44, 43, 42],
178
+
179
+ [ 4, 4, 4, ..., 44, 43, 42],
180
+
181
+ [ 4, 4, 4, ..., 44, 43, 41]], dtype=uint8)
182
+
165
183
  ```
166
184
 
167
185
  ```

2

文法の修正

2021/04/13 21:25

投稿

shishi_maru440
shishi_maru440

スコア38

test CHANGED
@@ -1 +1 @@
1
- 画像データから機械学習用のデータセットを作る時にエラーが発生(setting an array element with a sequence.)
1
+ 画像データをnumpy配列にし機械学習用のデータセットを作りたい
test CHANGED
File without changes

1

不具合内容を変更

2021/04/13 14:41

投稿

shishi_maru440
shishi_maru440

スコア38

test CHANGED
@@ -1 +1 @@
1
- 自前の画像データから機械学習用のデータセットを作りたい
1
+ 画像データから機械学習用のデータセットを作る時にエラーが発生(setting an array element with a sequence.)
test CHANGED
@@ -20,206 +20,308 @@
20
20
 
21
21
  TypeError Traceback (most recent call last)
22
22
 
23
+ TypeError: only size-1 arrays can be converted to Python scalars
24
+
25
+
26
+
27
+ The above exception was the direct cause of the following exception:
28
+
29
+
30
+
31
+ ValueError Traceback (most recent call last)
32
+
33
+ ValueError: setting an array element with a sequence.
34
+
35
+ ```
36
+
37
+
38
+
39
+ ### 該当のソースコード
40
+
41
+
42
+
43
+ ```python
44
+
45
+ import pandas as pd
46
+
47
+ import numpy as np
48
+
49
+ import glob
50
+
51
+ import cv2
52
+
53
+
54
+
55
+ #1.anotation_results_form_check.ipynbで作成したdata2.csvを読み込み
56
+
57
+ df2 = pd.read_csv("data2.csv")
58
+
59
+ df2.shape
60
+
61
+
62
+
63
+
64
+
65
+ #file pathの設定(フォルダ内の画像データを読み込み 画像データは6486枚)
66
+
67
+ files = glob.glob("/Users/a440/Desktop/happy_images/*")
68
+
69
+ len(files)
70
+
71
+
72
+
73
+ #filesのデータフレームを作成
74
+
75
+ df_files = pd.DataFrame({"data":files})
76
+
77
+
78
+
79
+ #dataのformatを合わせる
80
+
81
+ df2["data"] = "/Users/a440/Desktop/happy_images/" + df2["data"] + ".jpg"
82
+
83
+ df2 = df2.drop("Unnamed: 0", axis=1)
84
+
85
+ df2.to_csv("data2-2.csv")
86
+
87
+
88
+
89
+ #df2とdf_filesをdata列でマージする
90
+
91
+ df_join = pd.merge(df2, df_files, how="inner",on="data",indicator=True)
92
+
93
+ df_join.shape
94
+
95
+ ```
96
+
97
+ ```OUT
98
+
99
+ 6441, 3
100
+
101
+ ```
102
+
103
+ ```
104
+
105
+
106
+
107
+ df_join.to_csv("df_join.csv")
108
+
109
+
110
+
111
+ #targetにsmileのカテゴリを代入
112
+
113
+ smiles =df_join["smile"]
114
+
115
+ target = []
116
+
117
+
118
+
119
+ for smile in smiles:
120
+
121
+ data = np.asarray(smile)
122
+
123
+ target.append(data)
124
+
125
+ y = np.array(target)
126
+
127
+ y
128
+
129
+ ```
130
+
131
+ ```OUT
132
+
133
+ array([1, 3, 1, ..., 3, 3, 3])
134
+
135
+ ```
136
+
137
+ ```
138
+
139
+ #photo_array に画像データの配列を代入
140
+
141
+ titles = df_join["data"]
142
+
143
+
144
+
145
+
146
+
147
+ image_size = 50
148
+
149
+
150
+
151
+ photo_array = []
152
+
153
+
154
+
155
+ for title in titles:
156
+
157
+ bgr = cv2.imread(title, cv2.IMREAD_GRAYSCALE)
158
+
159
+ bgr = np.asarray(bgr)
160
+
161
+ photo_array.append(bgr)
162
+
163
+ x = np.array(photo_array)
164
+
165
+ ```
166
+
167
+ ```
168
+
169
+ from sklearn import linear_model
170
+
171
+ clf = linear_model.LogisticRegression()
172
+
173
+
174
+
175
+ from sklearn.model_selection import StratifiedKFold
176
+
177
+ ss = StratifiedKFold(n_splits=10,
178
+
179
+ shuffle=True)
180
+
181
+
182
+
183
+ for train_index, test_index in ss.split(x, y):
184
+
185
+ x_train, x_test = x[train_index], x[test_index]
186
+
187
+ y_train, y_test = y[train_index], y[test_index]
188
+
189
+
190
+
191
+ clf.fit(x_train, y_train)
192
+
193
+ print(clf.score(x_test, y_test))
194
+
195
+ ```
196
+
197
+ ```エラーメッセージ
198
+
199
+ ---------------------------------------------------------------------------
200
+
201
+ TypeError Traceback (most recent call last)
202
+
203
+ TypeError: only size-1 arrays can be converted to Python scalars
204
+
205
+
206
+
207
+ The above exception was the direct cause of the following exception:
208
+
209
+
210
+
211
+ ValueError Traceback (most recent call last)
212
+
23
- <ipython-input-14-70537278e5a3> in <module>
213
+ <ipython-input-15-305a5bf690e8> in <module>
24
-
25
- 7
214
+
26
-
27
- 8 for train_index, test_index in ss.split(photo_array, target):
28
-
29
- ----> 9 x_train, x_test = photo_array[train_index], photo_array[test_index]
30
-
31
- 10 y_train, y_test = target[train_index], target[test_index]
215
+ 10 y_train, y_test = y[train_index], y[test_index]
32
216
 
33
217
  11
34
218
 
35
-
36
-
37
- TypeError: only integer scalar arrays can be converted to a scalar index
38
-
39
-
40
-
41
- ```
42
-
43
-
44
-
45
- ### 該当のソースコード
46
-
47
-
48
-
49
- ```python
50
-
51
- import pandas as pd
52
-
53
- import numpy as np
54
-
55
- import glob
56
-
57
- import cv2
58
-
59
-
60
-
61
- #1.anotation_results_form_check.ipynbで作成したdata2.csvを読み込み
62
-
63
- df2 = pd.read_csv("data2.csv")
64
-
65
- df2.shape
66
-
67
-
68
-
69
-
70
-
71
- #file pathの設定(フォルダ内の画像データを読み込み 画像データは6486枚)
72
-
73
- files = glob.glob("/Users/a440/Desktop/happy_images/*")
74
-
75
- len(files)
76
-
77
-
78
-
79
- #filesのデータフレームを作成
80
-
81
- df_files = pd.DataFrame({"data":files})
82
-
83
-
84
-
85
- #dataのformatを合わせる
86
-
87
- df2["data"] = "/Users/a440/Desktop/happy_images/" + df2["data"] + ".jpg"
88
-
89
- df2 = df2.drop("Unnamed: 0", axis=1)
90
-
91
- df2.to_csv("data2-2.csv")
92
-
93
-
94
-
95
- #df2とdf_filesをdata列でマージする
96
-
97
- df_join = pd.merge(df2, df_files, how="inner",on="data",indicator=True)
98
-
99
- df_join.shape
100
-
101
- ```
102
-
103
- ```OUT
104
-
105
- 6441, 3
106
-
107
- ```
108
-
109
- ```
110
-
111
-
112
-
113
- df_join.to_csv("df_join.csv")
114
-
115
-
116
-
117
- #targetにsmileのカテゴリを代入
118
-
119
- target = df_join["smile"]
120
-
121
- target
122
-
123
- ```
124
-
125
- ```OUT
126
-
127
- 0 1
128
-
129
- 1 3
130
-
131
- 2 1
132
-
133
- 3 3
134
-
135
- 4 2
136
-
137
- ..
138
-
139
- 6436 3
140
-
141
- 6437 3
142
-
143
- 6438 3
144
-
145
- 6439 3
146
-
147
- 6440 3
148
-
149
- Name: smile, Length: 6441, dtype: int64
150
-
151
- ```
152
-
153
- ```
154
-
155
-
156
-
157
- titles = df_join["data"]
158
-
159
-
160
-
161
- photo_array = []
162
-
163
-
164
-
165
- for title in titles:
166
-
167
- bgr = cv2.imread(title, cv2.IMREAD_GRAYSCALE)
168
-
169
- bgr = np.array(bgr)
170
-
171
- photo_array.append(bgr)
172
-
173
-
174
-
175
- photo_array[0]
176
-
177
- ```
178
-
179
- ```OUT
180
-
181
- array([[217, 217, 217, ..., 188, 196, 203],
182
-
183
- [217, 217, 217, ..., 188, 196, 203],
184
-
185
- [217, 217, 217, ..., 188, 196, 204],
186
-
187
- ...,
188
-
189
- [ 4, 4, 4, ..., 44, 43, 42],
190
-
191
- [ 4, 4, 4, ..., 44, 43, 42],
192
-
193
- [ 4, 4, 4, ..., 44, 43, 41]], dtype=uint8)
194
-
195
- ```
196
-
197
- ```
198
-
199
- from sklearn import linear_model
200
-
201
- clf = linear_model.LogisticRegression()
202
-
203
-
204
-
205
- from sklearn.model_selection import StratifiedKFold
206
-
207
- ss = StratifiedKFold(n_splits=10,
208
-
209
- shuffle=True)
210
-
211
-
212
-
213
- for train_index, test_index in ss.split(photo_array, target):
214
-
215
- x_train, x_test = photo_array[train_index], photo_array[test_index]
216
-
217
- y_train, y_test = target[train_index], target[test_index]
218
-
219
-
220
-
221
- clf.fit(x_train, y_train)
222
-
223
- print(clf.score(x_test, y_test))
224
-
225
- ```
219
+ ---> 12 clf.fit(x_train, y_train)
220
+
221
+ 13 print(clf.score(x_test, y_test))
222
+
223
+ 14
224
+
225
+
226
+
227
+ /opt/anaconda3/lib/python3.7/site-packages/sklearn/linear_model/_logistic.py in fit(self, X, y, sample_weight)
228
+
229
+ 1342 X, y = self._validate_data(X, y, accept_sparse='csr', dtype=_dtype,
230
+
231
+ 1343 order="C",
232
+
233
+ -> 1344 accept_large_sparse=solver != 'liblinear')
234
+
235
+ 1345 check_classification_targets(y)
236
+
237
+ 1346 self.classes_ = np.unique(y)
238
+
239
+
240
+
241
+ /opt/anaconda3/lib/python3.7/site-packages/sklearn/base.py in _validate_data(self, X, y, reset, validate_separately, **check_params)
242
+
243
+ 430 y = check_array(y, **check_y_params)
244
+
245
+ 431 else:
246
+
247
+ --> 432 X, y = check_X_y(X, y, **check_params)
248
+
249
+ 433 out = X, y
250
+
251
+ 434
252
+
253
+
254
+
255
+ /opt/anaconda3/lib/python3.7/site-packages/sklearn/utils/validation.py in inner_f(*args, **kwargs)
256
+
257
+ 70 FutureWarning)
258
+
259
+ 71 kwargs.update({k: arg for k, arg in zip(sig.parameters, args)})
260
+
261
+ ---> 72 return f(**kwargs)
262
+
263
+ 73 return inner_f
264
+
265
+ 74
266
+
267
+
268
+
269
+ /opt/anaconda3/lib/python3.7/site-packages/sklearn/utils/validation.py in check_X_y(X, y, accept_sparse, accept_large_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, multi_output, ensure_min_samples, ensure_min_features, y_numeric, estimator)
270
+
271
+ 800 ensure_min_samples=ensure_min_samples,
272
+
273
+ 801 ensure_min_features=ensure_min_features,
274
+
275
+ --> 802 estimator=estimator)
276
+
277
+ 803 if multi_output:
278
+
279
+ 804 y = check_array(y, accept_sparse='csr', force_all_finite=True,
280
+
281
+
282
+
283
+ /opt/anaconda3/lib/python3.7/site-packages/sklearn/utils/validation.py in inner_f(*args, **kwargs)
284
+
285
+ 70 FutureWarning)
286
+
287
+ 71 kwargs.update({k: arg for k, arg in zip(sig.parameters, args)})
288
+
289
+ ---> 72 return f(**kwargs)
290
+
291
+ 73 return inner_f
292
+
293
+ 74
294
+
295
+
296
+
297
+ /opt/anaconda3/lib/python3.7/site-packages/sklearn/utils/validation.py in check_array(array, accept_sparse, accept_large_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, ensure_min_samples, ensure_min_features, estimator)
298
+
299
+ 596 array = array.astype(dtype, casting="unsafe", copy=False)
300
+
301
+ 597 else:
302
+
303
+ --> 598 array = np.asarray(array, order=order, dtype=dtype)
304
+
305
+ 599 except ComplexWarning:
306
+
307
+ 600 raise ValueError("Complex data not supported\n"
308
+
309
+
310
+
311
+ /opt/anaconda3/lib/python3.7/site-packages/numpy/core/_asarray.py in asarray(a, dtype, order)
312
+
313
+ 81
314
+
315
+ 82 """
316
+
317
+ ---> 83 return array(a, dtype, copy=False, order=order)
318
+
319
+ 84
320
+
321
+ 85
322
+
323
+
324
+
325
+ ValueError: setting an array element with a sequence.
326
+
327
+ ```