Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Sign up
Appearance settings

Commit 7eb8da3

Browse files
committed
预测验证码
1 parent 2bf5051 commit 7eb8da3

File tree

6 files changed

+29
-3
lines changed

6 files changed

+29
-3
lines changed

‎ch46-机器学习-K近邻/2-使用kNN对手写数字OCR.py‎

Lines changed: 6 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -43,9 +43,10 @@
4343
accuracy = correct * 100.0 / result.size
4444
print('准确率', accuracy) # 准确率91.76%
4545

46-
''''''
46+
4747
# save the data
4848
np.savez('knn_data.npz', train=train, train_labels=train_labels,test=test,test_labels=test_labels)
49+
4950
# Now load the data
5051
with np.load('knn_data.npz') as data:
5152
print(data.files)
@@ -56,5 +57,8 @@
5657

5758

5859
#TODO 怎样预测数字?
59-
retval, results=knn.predict(test[3:5])
60+
retval, results=knn.predict(test[1003:1005])
6061
# Docstring: predict(samples[, results[, flags]]) -> retval, results
62+
print(retval, results)#(4.0, array([[ 4.],[ 4.]], dtype=float32))
63+
#对比
64+
cv2.imwrite('test[1005].jpg',test[1005].reshape((20,20)))

‎ch46-机器学习-K近邻/2.png‎

24.7 KB
Loading[フレーム]

‎ch46-机器学习-K近邻/6.png‎

28 KB
Loading[フレーム]
3.83 MB
Binary file not shown.
695 Bytes
Loading[フレーム]

‎ch46-机器学习-K近邻/预测手写数字1.py‎

Lines changed: 23 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -19,6 +19,28 @@
1919
test = data['test']
2020
test_labels = data['test_labels']
2121

22+
print('加载KNN,数据')
2223
knn = cv2.ml.KNearest_create()
2324
knn.train(train, cv2.ml.ROW_SAMPLE, train_labels)
24-
ret, result, neighbours, dist = knn.findNearest(test, k=5)
25+
ret, result, neighbours, dist = knn.findNearest(test, k=5)
26+
27+
# 加载相片
28+
print('加载相片')
29+
img2 = cv2.imread('2.png', 0)
30+
gray2 = cv2.resize(img2, (20, 20))
31+
# gray2=gray2.reshape((400,))
32+
gray21 = gray2.reshape((-1, 400)).astype(np.float32)
33+
34+
img6 = cv2.imread('6.png', 0)
35+
gray6 = cv2.resize(img6, (20, 20))
36+
# gray2=gray2.reshape((400,))
37+
gray61 = gray6.reshape((-1, 400)).astype(np.float32)
38+
39+
g2 = np.append(gray21, gray61)
40+
g3 = g2.reshape((2, 400))
41+
42+
# 预测
43+
retval, results = knn.predict(g3)
44+
print(retval, results) # 不准确
45+
# (0.0, array([[ 0.],
46+
# [ 5.]], dtype=float32))

0 commit comments

Comments
(0)

AltStyle によって変換されたページ (->オリジナル) /