# -*- coding: utf-8 -*- """ Created on Sat Apr 20 14:16:19 2019 @author: qinzhen """ import helper as hlp #Step1 产生数据 #参数 rad = 10 thk = 5 sep = 5 #产生数据 X, y = hlp.generatedata(rad, thk, sep, 100) knn = hlp.KNeighborsClassifier_(1) knn.fit(X, y) hlp.draw(X, y, knn) knn = hlp.KNeighborsClassifier_(3) knn.fit(X, y) hlp.draw(X, y, knn)
#!/usr/bin/env python
# -*- coding=utf-8 -*-
from config import VIDEO_ID, NUM_WINNERS
from helper import get_commenters, draw, get_beijing_time
if __name__ == '__main__':
print("北京时间: {}".format(get_beijing_time()))
with open("cookies.txt", "r") as f:
raw = f.read().strip()
cookies = dict(cookies_are=raw)
followed, not_followed = get_commenters(VIDEO_ID, cookies)
pool = set(followed.keys())
winners = draw(pool, NUM_WINNERS)
print("*************************************")
print("抽奖结果: ")
for i in range(len(winners)):
print("第{}位获奖观众: {} - https://space.bilibili.com/{}".format(
i + 1, followed[winners[i]], winners[i]))
# -*- coding: utf-8 -*-
"""
Created on Thu May 2 09:25:37 2019
@author: qinzhen
"""
import numpy as np
import helper as hlp
X = np.array([[0, 0], [0, 1], [5, 5]]).astype("float")
y = np.array([1, 1, -1])
#(a)
knn = hlp.KNeighborsClassifier_(1)
knn.fit(X, y)
hlp.draw(X, y, knn)
#(b)
scaler = hlp.StandardScaler_()
X1 = scaler.fit_transform(X)
knn = hlp.KNeighborsClassifier_(1)
knn.fit(X1, y)
hlp.draw(X, y, knn, flag=2, preprocess=scaler)
#(c)
pca = hlp.PCA_(1)
pca.fit(X)
X2 = pca.transform(X)
knn = hlp.KNeighborsClassifier_(1)
knn.fit(X2, y)
pool = set([i for i in range(TEST_POOL_SIZE)])
def test_config_params():
if TEST_POOL_SIZE <= 0 or TEST_WINNERS <= 0 or TEST_EXPERIMENTS <= 0:
print("测试参数设定错误，各项测试参数均应大于零")
raise ValueError
if TEST_POOL_SIZE < TEST_WINNERS:
print("测试参数设定错误，抽奖人数应小于等于奖池人数")
raise ValueError
if __name__ == '__main__':
try:
test_config_params()
except ValueError:
sys.exit(1)
print("测试开始...")
print("将进行{}次实验，每次抽取{}名获奖观众".format(TEST_EXPERIMENTS, TEST_WINNERS))
mean = TEST_EXPERIMENTS * TEST_WINNERS // TEST_POOL_SIZE
print("理想状况下，每位观众被抽中次数应接近{}".format(mean))
results = {}
for i in range(TEST_EXPERIMENTS):
result = draw(pool, TEST_WINNERS)
for key in result:
results[key] = results.get(key, 0) + 1
print("实验结束，结果如下：")
print("观众最少被抽中{}次，最多被抽中{}次".format(min(results.values()),
max(results.values())))
"""
Created on Sat Apr 20 12:48:23 2019
@author: qinzhen
"""
#(a)
import numpy as np
import helper as hlp
X = np.array([[1, 0], [0, 1], [0, -1], [-1, 0], [0, 2], [0, -2], [-2, 0]])
y = np.array([-1, -1, -1, -1, 1, 1, 1])
knn = hlp.KNeighborsClassifier_(1)
knn.fit(X, y)
hlp.draw(X, y, knn)
knn = hlp.KNeighborsClassifier_(3)
knn.fit(X, y)
hlp.draw(X, y, knn)
#(b)
#特征转换
Z = np.c_[np.sqrt(X[:, 0]**2 + X[:, 1]**2),
np.arctan(X[:, 1] / (X[:, 0] + 10**(-8)))]
knn = hlp.KNeighborsClassifier_(1)
knn.fit(Z, y)
hlp.draw(X, y, knn, flag=0)
knn = hlp.KNeighborsClassifier_(3)
knn.fit(Z, y)
print("You had this char before")
for item in hangedPortion:
if (len(item) > 0):
print(item)
else:
choosed.append(char)
if (char in word):
indexList = helper.find(word, char)
for i in indexList:
tempGuess[i] = char
helper.newLine()
if ('_' not in tempGuess):
helper.clear('You won!!!')
print('The word was ' + word)
helper.newLine()
helper.draw(notHanged)
helper.newLine()
break
else:
helper.clear()
status(2)
for item in hangedPortion:
if (len(item) > 0):
print(item)
else:
wrongs -= 1
helper.clear()
status(1)
print('Wrong. You are left {} shots.'.format(wrongs))
if (len(hangedgrad) > 1):
hangedPortion.append(hangedgrad.pop())
def main():
# cap_img = cv2.imread('Images/Test_Markers/7.jpeg')
camera_parameters, dist = he.camera_calibration('Images/Calibration')
dir_name = os.getcwd()
obj = OBJ(os.path.join(dir_name, 'Models/fox/fox.obj'), swapyz=True)
model_list = he.create_list('Images')
model_keypoints = he.find_model_keypoints(model_list)
# print('b')
if not args.problem:
cap = cv2.VideoCapture(0)
while True:
# read the current frame
ret, frame = cap.read()
if not ret:
print("Unable to capture video")
return
cnt = None
temp = he.find_marker(he.convert2grayscale(frame), model_list,
model_keypoints, camera_parameters, dist)
if temp is not None:
cnt, homographies, models = temp
# compute Homography if enough matches are found
if cnt is not None:
# if a valid homography matrix was found render cube on model plane
for x, h, m in zip(cnt, homographies, models):
try:
projection = he.projection_matrix(camera_parameters, h)
frame = he.render(frame, obj, projection, m, False)
# rect = cnt[0]
# pseudo_model = h.crop_img(frame.copy(), rect)
# r_p, c_p = pseudo_model.shape[0], pseudo_model.shape[1]
# pts = [x[0] for x in rect]
# pts2 = [[0, 0], [0, r_p], [c_p, r_p], [c_p, 0]]
# projection = cv2.getPerspectiveTransform(pts, pts2)
break
except:
pass
# draw first 10 matches.
if args.rectangle:
if cnt is not None:
for x in cnt:
x = he.convert_to_ndarray(x)
cv2.drawContours(frame, [x], -1, (0, 255, 0), 3)
points_dum = np.array([[[0, 0, 0]], [[100, 0, 0]],
[[0, 100, 0]], [[0, 0,
100]]])
points_dum = np.float32(
[np.float32(p) for p in points_dum])
# print(axis)
newpts = cv2.perspectiveTransform(
points_dum, projection)
cap_img = he.draw(
frame, newpts[0][0],
[newpts[1][0], newpts[2][0], newpts[3][0]])
break
# show result
else:
print('No marker found')
cv2.imshow('frame', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
elif args.problem:
# print('c')
no = input('Enter image no. to process: ')
# img nos -
cap_img = cv2.imread('Images/Test_Markers/Prob2/' + str(no) + '.jpeg')
camera_parameters, dist = he.camera_calibration('Images/Calibration')
model_list = he.create_list('Images')
model_keypoints = he.find_model_keypoints(model_list)
cnt = None
temp = he.find_marker(he.convert2grayscale(cap_img), model_list,
model_keypoints, camera_parameters, dist)
if temp is not None:
cnt, homographies, models = temp
cnt = [cnt[1], cnt[0]]
homographies = [homographies[1], homographies[0]]
models = [models[1], models[0]]
if cnt is not None:
if (len(cnt) == 2):
ro1, co1 = models[0].shape
ro2, co2 = models[1].shape
proj1 = he.projection_matrix(camera_parameters,
homographies[0])
proj2 = he.projection_matrix(camera_parameters,
homographies[1])
points1 = np.array([[[co1 / 2, ro1 / 2, 0]]])
points1 = np.float32([np.float32(p) for p in points1])
origin1 = np.array(
cv2.perspectiveTransform(points1, proj1)[0][0])
points2 = np.array([[[0, 0, 0]], [[100, 0, 0]], [[0, 100, 0]],
[[0, 0, 100]]])
points2 = np.float32([np.float32(p) for p in points2])
axes2 = cv2.perspectiveTransform(points2, proj2)
origin2 = axes2[0][0]
ptz = axes2[3][0]
######
if args.rectangle:
projection = proj1
points_dum = np.array([[[0, 0, 0]], [[100, 0, 0]],
[[0, 100, 0]], [[0, 0, 100]]])
points_dum = np.float32(
[np.float32(p) for p in points_dum])
newpts = cv2.perspectiveTransform(points_dum, projection)
cap_img = he.draw(
cap_img, newpts[0][0],
[newpts[1][0], newpts[2][0], newpts[3][0]])
projection = proj2
points_dum = np.array([[[0, 0, 0]], [[100, 0, 0]],
[[0, 100, 0]], [[0, 0, 100]]])
points_dum = np.float32(
[np.float32(p) for p in points_dum])
newpts = cv2.perspectiveTransform(points_dum, projection)
cap_img = he.draw(
cap_img, newpts[0][0],
[newpts[1][0], newpts[2][0], newpts[3][0]])
######
origin2 = np.array(origin2)
origin1_x = origin1[0]
origin1_y = origin1[1]
origin2_x = origin2[0]
ptz_x = ptz[0]
origin2_y = origin2[1]
ptz_y = ptz[1]
moving_slope = (ptz_y - origin2_y) / (ptz_x - origin2_x)
# moving_slope = (origin2_y-origin1_y)/(origin2_x-origin1_x)
moving_positive = origin1 + STEP_SIZE * \
np.array([1, moving_slope])
moving_negative = origin1 - STEP_SIZE * \
np.array([1, moving_slope])
distance_pos = np.linalg.norm(origin2 - moving_positive)
distance_neg = np.linalg.norm(origin2 - moving_negative)
if (distance_pos < distance_neg):
counter = 0
old_distance = math.inf
while True:
print('Slope:' + str(moving_slope))
frame = cap_img.copy()
moving_positive = origin1 + counter * STEP_SIZE * \
np.array([1, moving_slope])
new_distance = np.linalg.norm(origin2 -
moving_positive)
print('new:' + str(new_distance))
print('old:' + str(old_distance))
if (new_distance > old_distance):
print('break')
break
old_distance = new_distance
counter += 1
shift = moving_positive - origin1
frame = he.render(frame, obj, proj1, models[0], False,
shift[0], shift[1])
print(shift)
cv2.imshow('Moving_obj', frame)
cv2.waitKey(1)
else:
counter = 0
old_distance = math.inf
while True:
print('Slope:' + str(moving_slope))
frame = cap_img.copy()
moving_negative = origin1 - counter * STEP_SIZE * \
np.array([1, moving_slope])
new_distance = np.linalg.norm(origin2 -
moving_negative)
print('new:' + str(new_distance))
print('old:' + str(old_distance))
if (new_distance > old_distance):
print('break')
break
old_distance = new_distance
counter += 1
shift = moving_negative - origin1
frame = he.render(frame, obj, proj1, models[0], False,
shift[0], shift[1])
print(shift)
cv2.imshow('Moving_obj', frame)
cv2.waitKey(1)
else:
print('Less than 2 Markers Found')
cv2.destroyAllWindows()
return 0
y_cd.append(-1)
else:
break
return np.array(X_cd), np.array(y_cd)
#### (b)
n = 1000
X, y = hlp.Data(n, 1, scale=0.5)
plt.scatter(X[:, 0], X[:, 1], edgecolor='k', c=y)
plt.show()
#原始数据的分类结果
nn = hlp.KNeighborsClassifier_(1)
nn.fit(X, y)
hlp.draw(X, y, nn, n=500, flag=1)
#### CNN1
#condense data的分类结果
X_cd, y_cd = CNN1(X, y)
nn = hlp.KNeighborsClassifier_(1)
nn.fit(X_cd, y_cd)
hlp.draw(X_cd, y_cd, nn, n=500, flag=1)
#### CNN2
#condense data的分类结果
X_cd, y_cd = CNN2(X, y)
nn = hlp.KNeighborsClassifier_(1)
nn.fit(X_cd, y_cd)
hlp.draw(X_cd, y_cd, nn, n=500, flag=1)
'''
y_test[y_test != 1] = -1
#### (b)
#训练模型
nn = hlp.KNeighborsClassifier_(3)
nn.fit(X_train, y_train)
y_test_pred = nn.predict(X_test)
y_train_pred = nn.predict(X_train)
#计算错误率
ein = np.mean(y_train != y_train_pred)
etest = np.mean(y_test != y_test_pred)
print("Ein of 3-NN is {}".format(ein))
print("Etest of 3-NN is {}".format(etest))
#作图
hlp.draw(X_train, y_train, nn, n=500, flag=1)
#### (c)
X_cd, y_cd = hlp.CNN(X_train, y_train, k=3)
#训练模型
nn = hlp.KNeighborsClassifier_(3)
nn.fit(X_cd, y_cd)
y_test_pred = nn.predict(X_test)
y_train_pred = nn.predict(X_train)
#计算错误率
ein = np.mean(y_train != y_train_pred)
etest = np.mean(y_test != y_test_pred)
print("Ein of C-NN is {}".format(ein))
print("Etest of C-NN is {}".format(etest))
#作图