为什么我得到 AttributeError？
Why did I get AttributeError?
我尝试更改原始代码中的几行，但是当我尝试 运行 时，我收到错误提示“AttributeError：模块 'PngImageFile' 没有属性 'shape'。
但是，当 运行ning 原始代码时，我没有问题。我应该怎么做才能消除修改后的代码中的这个错误？
原代码如下：
from skimage.feature import hog, local_binary_pattern
from skimage.transform import pyramid_gaussian
from skimage.io import imread
import joblib
from sklearn.preprocessing import LabelEncoder
from sklearn import svm
from sklearn.metrics import classification_report
from sklearn.model_selection import train_test_split
from skimage import color
from imutils.object_detection import non_max_suppression
import imutils
import numpy as np
import argparse
import cv2
import os
import glob
from sklearn import metrics
from PIL import Image # This will be used to read/modify images (can be done via OpenCV too)
from numpy import *
# define parameters of HOG feature extraction
orientations = 9
pixels_per_cell = (8, 8)
cells_per_block = (2, 2)
threshold = .3
# define path to images:
dataset_path = r"C:\Users\user\Documents\FYP\Dataset FYP\Train New\Contrast" # path of our dataset
# compute HOG features and label them:
for category in category_im_listing: #this loop enables reading the files in the pos_im_listing variable one by one
im_listing = os.listdir(dataset_path + "/" + category)
num_im = size(im_listing)
print("There are " + str(num_im) + " images in category " + str(count + 1))
for file in im_listing:
img = Image.open(dataset_path + "/" + category + "/" + file) # open the file
img = img.resize((150,150))
gray = img.convert('L') # convert the image into single channel i.e. RGB to grayscale
# calculate HOG for positive features
fd = hog(gray, orientations, pixels_per_cell, cells_per_block, block_norm='L2', feature_vector=True)# fd= feature descriptor
data.append(fd)
labels.append(count)
count = count + 1
这是我修改但出错的代码：
from skimage.feature import hog, local_binary_pattern
from skimage.transform import pyramid_gaussian
from skimage.io import imread
import joblib
from sklearn.preprocessing import LabelEncoder
from sklearn import svm
from sklearn.metrics import classification_report
from sklearn.model_selection import train_test_split
from skimage import color
from imutils.object_detection import non_max_suppression
import imutils
import numpy as np
import imageio
from ipynb.fs.full.anna_phog import anna_phog
import argparse
import cv2
import os
import glob
from sklearn import metrics
from PIL import Image # This will be used to read/modify images (can be done via OpenCV too)
from numpy import *
image_path = r"C:\Users\user\Documents\FYP\me\train"
#MODIFIED
S = 8
angle = 360
Level = 3
roi = [1,225,1,300]
save=True
# read the image files:
category_im_listing = os.listdir(image_path) # it will read all the files in the path
num_category_im = size(category_im_listing) # simply states the total no. of category
print("There are " + str(num_category_im) + " categories") # prints the number value of the no.of categories dataset
data= []
labels = []
count = 0
# Modified : I deleted the resize and gray line, and change the fd line into p because I want to extract PHOG instead of HOG
for category in category_im_listing: #this loop enables reading the files in the pos_im_listing variable one by one
im_listing = os.listdir(image_path + "/" + category)
num_im = size(im_listing)
print("There are " + str(num_im) + " images in category " + str(count + 1))
for file in im_listing:
img = Image.open(image_path + "/" + category + "/" + file) # open the file
# calculate PHOG for positive features
p = anna_phog(img, bin, angle, Level, roi) #MODIFIED
data.append(p)
labels.append(count)
count = count + 1
这是我在修改后的代码中遇到的错误：
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-76-9e9360ca5082> in <module>
8 img = Image.open(image_path + "/" + category + "/" + file) # open the file
9 # calculate HOG for positive features
---> 10 p = anna_phog(img, bin, angle, Level, roi)# fd= feature descriptor
11 data.append(p)
12 labels.append(count)
~\Desktop\FYP\anna_phog.ipynb in anna_phog(Img, bin, angle, L, roi)
11 "import cv2\n",
12 "import matplotlib.pyplot as plt"
---> 13 ]
14 },
15 {
AttributeError: 'PngImageFile' object has no attribute 'shape'
新错误（更新）：
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-11-00f72c361b79> in <module>
8 img=cv2.imread(image_path + "/" + category + "/" + file) # open the file
9 # calculate PHOG for positive features
---> 10 p = anna_phog(img, bin, angle, Level, roi) #MODIFIED
11 data.append(p)
12 labels.append(count)
~\Desktop\FYP\anna_phog.ipynb in anna_phog(Img, bin, angle, L, roi)
35 " E = cv2.Canny(G,lower,upper) #high and low treshold\n",
36 " GradientX, GradientY = np.gradient(G)\n",
---> 37 " GradientYY = np.gradient(GradientY, axis=1)\n",
38 " \n",
39 " Gr = np.sqrt(np.square(GradientX)+np.square(GradientY))\n",
~\Desktop\FYP\anna_phog.ipynb in anna_BinMatrix(A, E, G, angle, bin)
56 " p = anna_PhogDescriptor(bh_roi,bv_roi,L,bin)\n",
57 " return p"
---> 58 ]
59 },
60 {
TypeError: unsupported operand type(s) for /: 'int' and 'builtin_function_or_method'
我在其他门户网站上看到了anna_phog。
问题是因为此函数需要 numpy array 但您使用 pillow Image.open() 读取图像并且您必须将 img 转换为 numpy array
img = np.asarry(img)
编辑：
还有其他问题。
anna_phog 转换颜色 BGR2GRAY 因为 cv2 在 BGR 中给出颜色而不是 RGB - 但是 pillow 在 [=] 中给出颜色20=]。所以最好使用 cv2 来读取图像。
我尝试更改原始代码中的几行，但是当我尝试 运行 时，我收到错误提示“AttributeError：模块 'PngImageFile' 没有属性 'shape'。 但是，当 运行ning 原始代码时，我没有问题。我应该怎么做才能消除修改后的代码中的这个错误？
原代码如下：
from skimage.feature import hog, local_binary_pattern
from skimage.transform import pyramid_gaussian
from skimage.io import imread
import joblib
from sklearn.preprocessing import LabelEncoder
from sklearn import svm
from sklearn.metrics import classification_report
from sklearn.model_selection import train_test_split
from skimage import color
from imutils.object_detection import non_max_suppression
import imutils
import numpy as np
import argparse
import cv2
import os
import glob
from sklearn import metrics
from PIL import Image # This will be used to read/modify images (can be done via OpenCV too)
from numpy import *
# define parameters of HOG feature extraction
orientations = 9
pixels_per_cell = (8, 8)
cells_per_block = (2, 2)
threshold = .3
# define path to images:
dataset_path = r"C:\Users\user\Documents\FYP\Dataset FYP\Train New\Contrast" # path of our dataset
# compute HOG features and label them:
for category in category_im_listing: #this loop enables reading the files in the pos_im_listing variable one by one
im_listing = os.listdir(dataset_path + "/" + category)
num_im = size(im_listing)
print("There are " + str(num_im) + " images in category " + str(count + 1))
for file in im_listing:
img = Image.open(dataset_path + "/" + category + "/" + file) # open the file
img = img.resize((150,150))
gray = img.convert('L') # convert the image into single channel i.e. RGB to grayscale
# calculate HOG for positive features
fd = hog(gray, orientations, pixels_per_cell, cells_per_block, block_norm='L2', feature_vector=True)# fd= feature descriptor
data.append(fd)
labels.append(count)
count = count + 1
这是我修改但出错的代码：
from skimage.feature import hog, local_binary_pattern
from skimage.transform import pyramid_gaussian
from skimage.io import imread
import joblib
from sklearn.preprocessing import LabelEncoder
from sklearn import svm
from sklearn.metrics import classification_report
from sklearn.model_selection import train_test_split
from skimage import color
from imutils.object_detection import non_max_suppression
import imutils
import numpy as np
import imageio
from ipynb.fs.full.anna_phog import anna_phog
import argparse
import cv2
import os
import glob
from sklearn import metrics
from PIL import Image # This will be used to read/modify images (can be done via OpenCV too)
from numpy import *
image_path = r"C:\Users\user\Documents\FYP\me\train"
#MODIFIED
S = 8
angle = 360
Level = 3
roi = [1,225,1,300]
save=True
# read the image files:
category_im_listing = os.listdir(image_path) # it will read all the files in the path
num_category_im = size(category_im_listing) # simply states the total no. of category
print("There are " + str(num_category_im) + " categories") # prints the number value of the no.of categories dataset
data= []
labels = []
count = 0
# Modified : I deleted the resize and gray line, and change the fd line into p because I want to extract PHOG instead of HOG
for category in category_im_listing: #this loop enables reading the files in the pos_im_listing variable one by one
im_listing = os.listdir(image_path + "/" + category)
num_im = size(im_listing)
print("There are " + str(num_im) + " images in category " + str(count + 1))
for file in im_listing:
img = Image.open(image_path + "/" + category + "/" + file) # open the file
# calculate PHOG for positive features
p = anna_phog(img, bin, angle, Level, roi) #MODIFIED
data.append(p)
labels.append(count)
count = count + 1
这是我在修改后的代码中遇到的错误：
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-76-9e9360ca5082> in <module>
8 img = Image.open(image_path + "/" + category + "/" + file) # open the file
9 # calculate HOG for positive features
---> 10 p = anna_phog(img, bin, angle, Level, roi)# fd= feature descriptor
11 data.append(p)
12 labels.append(count)
~\Desktop\FYP\anna_phog.ipynb in anna_phog(Img, bin, angle, L, roi)
11 "import cv2\n",
12 "import matplotlib.pyplot as plt"
---> 13 ]
14 },
15 {
AttributeError: 'PngImageFile' object has no attribute 'shape'
新错误（更新）：
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-11-00f72c361b79> in <module>
8 img=cv2.imread(image_path + "/" + category + "/" + file) # open the file
9 # calculate PHOG for positive features
---> 10 p = anna_phog(img, bin, angle, Level, roi) #MODIFIED
11 data.append(p)
12 labels.append(count)
~\Desktop\FYP\anna_phog.ipynb in anna_phog(Img, bin, angle, L, roi)
35 " E = cv2.Canny(G,lower,upper) #high and low treshold\n",
36 " GradientX, GradientY = np.gradient(G)\n",
---> 37 " GradientYY = np.gradient(GradientY, axis=1)\n",
38 " \n",
39 " Gr = np.sqrt(np.square(GradientX)+np.square(GradientY))\n",
~\Desktop\FYP\anna_phog.ipynb in anna_BinMatrix(A, E, G, angle, bin)
56 " p = anna_PhogDescriptor(bh_roi,bv_roi,L,bin)\n",
57 " return p"
---> 58 ]
59 },
60 {
TypeError: unsupported operand type(s) for /: 'int' and 'builtin_function_or_method'
我在其他门户网站上看到了anna_phog。
问题是因为此函数需要 numpy array 但您使用 pillow Image.open() 读取图像并且您必须将 img 转换为 numpy array
img = np.asarry(img)
编辑：
还有其他问题。
anna_phog 转换颜色 BGR2GRAY 因为 cv2 在 BGR 中给出颜色而不是 RGB - 但是 pillow 在 [=] 中给出颜色20=]。所以最好使用 cv2 来读取图像。