07/02/2023, 17:25
gỡ lỗi file.py với ARface database
bạn nào rảnh về phần này chỉ giúp mình sửa lỗi với
import numpy as np
import matplotlib.pyplot as plt
from sklearn import linear_model
from sklearn.metrics import accuracy_score
from scipy import misc # for
from sklearn import preprocessing
np.random.seed(1)
path = 'D:\a\ARdatabase\AR_warp_zip\test2' # path to the database đường dẫn đến cơ sở dữ liệu
train_ids = np.arange(1, 26)
test_ids = np.arange(26, 50)
view_ids = np.hstack((np.arange(1, 8), np.arange(14, 21)))
D = 165*120 # original dimension
d = 500 # new dimension
# generate the projection matrix tạo ma trận
ProjectionMatrix = np.random.randn(D, d)
def build_list_fn(pre, img_ids, view_ids):
"""
pre = 'M-' or 'W-'
img_ids: indexes of images
view_ids: indexes of views
"""
list_fn = []
for im_id in img_ids:
for v_id in view_ids:
fn = path + pre + str(im_id).zfill(3) + '-' + \
str(v_id).zfill(2) + '.bmp'
list_fn.append(fn)
return list_fn
def rgb2gray(rgb):
# Y' = 0.299 R + 0.587 G + 0.114 B
return rgb[:,:,0]*.299 + rgb[:, :, 1]*.587 + rgb[:, :, 2]*.114
# feature extraction khai thác tính năng
def vectorize_img(filename):
# load image tải hình ảnh
rgb = misc.imread(filename)
# convert to gray scale chuyên sang tham xám
gray = rgb2gray(rgb)
# vectorization each row is a data point vector hóa
im_vec = gray.reshape(1, D)
return im_vec
def build_data_matrix(img_ids, view_ids):
total_imgs = img_ids.shape[0]*view_ids.shape[0]*2
X_full = np.zeros((total_imgs, D))
y = np.hstack((np.zeros((total_imgs/2, )), np.ones((total_imgs/2, ))
list_fn_m = build_list_fn('M-', img_ids, view_ids)
list_fn_w = build_list_fn('W-', img_ids, view_ids)
list_fn = list_fn_m + list_fn_w
for i in range(len(list_fn)):
X_full[i, :] = vectorize_img(list_fn[i])
X = np.dot(X_full, ProjectionMatrix)
return (X, y)
(X_train_full, y_train) = build_data_matrix(train_ids, view_ids)
x_mean = X_train_full.mean(axis = 0)
x_var = X_train_full.var(axis = 0)
def feature_extraction(X):
return (X - x_mean)/x_var
X_train = feature_extraction(X_train_full)
X_train_full = None ## free this variable
(X_test_full, y_test) = build_data_matrix(test_ids, view_ids)
X_test = feature_extraction(X_test_full)
X_test_full = None
logreg = linear_model.LogisticRegression(C=1e5) # just a big number điwa ra 1 con số
logreg.fit(X_train, y_train)
y_pred = logreg.predict(X_test)
print ("Accuracy: %.2f %%" %(100*accuracy_score(y_test, y_pred)))
Ngô Sỹ Bình
viết 17:26 ngày 07/02/2023
