Я пытаюсь классифицировать МРТ-изображения опухоли головного мозга как нормальные, злокачественные или доброкачественные. Для этого я хочу запустить две нейронные сети в одной программе. Первая сеть классифицирует, является ли МРТ головного мозга опухолевой или не опухолевой. Если опухоль, вторая сеть классифицирует, является ли изображение МРТ мозга злокачественным или доброкачественным. Код выглядит следующим образом:
# Importing the Keras libraries and packages
from keras.models import Sequential
from keras.layers import Conv2D
from keras.layers import MaxPooling2D
from keras.layers import Flatten
from keras.layers import Dense
import sys
from PIL import Image
sys.modules['Image'] = Image
#import PIL as pillow
#from PIL import Image
# Initialising the CNN
classifier = Sequential()
# Step 1 - Convolution
classifier.add(Conv2D(32, (3, 3), input_shape = (100, 100, 3), activation = 'relu'))
# Step 2 - Pooling
classifier.add(MaxPooling2D(pool_size = (2, 2)))
# Adding a second convolutional layer
classifier.add(Conv2D(32, (3, 3), activation = 'relu'))
classifier.add(MaxPooling2D(pool_size = (2, 2)))
# Step 3 - Flattening
classifier.add(Flatten())
# Step 4 - Full connection
classifier.add(Dense(units = 128, activation = 'relu'))
classifier.add(Dense(units = 1, activation = 'sigmoid'))
# Compiling the CNN
classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])
# Part 2 - Fitting the CNN to the images
from keras.preprocessing.image import ImageDataGenerator
train_datagen = ImageDataGenerator(rescale = 1./255,
shear_range = 0.2,
zoom_range = 0.2,
horizontal_flip = True)
test_datagen = ImageDataGenerator(rescale = 1./255)
training_set = train_datagen.flow_from_directory('C:\\Users\\Admin\\Desktop\\tumor_non_tumour\\training',
target_size = (100, 100),
batch_size = 32,
class_mode = 'binary')
test_set = test_datagen.flow_from_directory('C:\\Users\\Admin\\Desktop\\tumor_non_tumour\\testing',
target_size = (100, 100),
batch_size = 32,
class_mode = 'binary')
classifier.fit_generator(training_set,
steps_per_epoch = 10,
epochs = 20,
validation_data = test_set,
validation_steps = 10)
# Part 3 - Making new predictions
import numpy as np
from tkinter import *
from tkinter import filedialog
root = Tk()
from keras.preprocessing import image
test_image = image.load_img(filedialog.askopenfilename( filetypes = ( ("image files" , "*.jpg") , ("all files", "*.*"))), target_size = (100, 100))
test_image = image.img_to_array(test_image)
test_image = np.expand_dims(test_image, axis = 0)
result = classifier.predict(test_image)
training_set.class_indices
if result[0][0] == 0:
prediction = 'normal'
else :
prediction = 'tumorous'
#print(prediction)
if(result[0][0] == 1) :
sys.modules['Image'] = Image
#import PIL as pillow
#from PIL import Image
# Initialising the CNN
classifier = Sequential()
# Step 1 - Convolution
classifier.add(Conv2D(32, (3, 3), input_shape = (100, 100, 3), activation = 'relu'))
# Step 2 - Pooling
classifier.add(MaxPooling2D(pool_size = (2, 2)))
# Adding a second convolutional layer
classifier.add(Conv2D(32, (3, 3), activation = 'relu'))
classifier.add(MaxPooling2D(pool_size = (2, 2)))
# Step 3 - Flattening
classifier.add(Flatten())
# Step 4 - Full connection
classifier.add(Dense(units = 128, activation = 'relu'))
classifier.add(Dense(units = 1, activation = 'sigmoid'))
# Compiling the CNN
classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])
# Part 2 - Fitting the CNN to the images
#from keras.preprocessing.image import ImageDataGenerator
train_datagen = ImageDataGenerator(rescale = 1./255,
shear_range = 0.2,
zoom_range = 0.2,
horizontal_flip = True)
test_datagen = ImageDataGenerator(rescale = 1./255)
training_set = train_datagen.flow_from_directory("C:/Users/Admin/Desktop/malig_benign/training",
target_size = (100, 100),
batch_size = 32,
class_mode = 'binary')
test_set = test_datagen.flow_from_directory('C:/Users/Admin/Desktop/malig_benign/testing',
target_size = (100, 100),
batch_size = 32,
class_mode = 'binary')
classifier.fit_generator(training_set,
steps_per_epoch = 10,
epochs = 10,
validation_data = test_set,
validation_steps = 10)
# Part 3 - Making new predictions
#import numpy as np
#from tkinter import *
#from tkinter import filedialog
#from keras.preprocessing import image
test_image = image.load_img(test_image, target_size = (100, 100))
test_image = image.img_to_array(test_image)
test_image = np.expand_dims(test_image, axis = 0)
result = classifier.predict(test_image)
training_set.class_indices
if result[0][0] == 0:
prediction = 'malignant'
else :
prediction = 'benign'
print (prediction)
Я даже пытался изменить имя переменной во второй нейронной сети, но все еще не получил окончательный результат. Я получаю эту ошибку:
AttributeError Traceback (most recent call last)
~\Anaconda3\envs\tensorflow\lib\site-packages\PIL\Image.py in open(fp, mode)
2546 try:
-> 2547 fp.seek(0)
2548 except (AttributeError, io.UnsupportedOperation):
AttributeError: 'numpy.ndarray' object has no attribute 'seek'
During handling of the above exception, another exception occurred:
AttributeError Traceback (most recent call last)
<ipython-input-2-d9dc0d704b3a> in <module>()
110
111 from keras.preprocessing import image
--> 112 test_image = image.load_img(test_image, target_size = (100, 100))
113 test_image = image.img_to_array(test_image)
114 test_image1= np.expand_dims(test_image, axis = 0)
~\Anaconda3\envs\tensorflow\lib\site-packages\keras\preprocessing\image.py in load_img(path, grayscale, target_size, interpolation)
360 raise ImportError('Could not import PIL.Image. '
361 'The use of `array_to_img` requires PIL.')
--> 362 img = pil_image.open(path)
363 if grayscale:
364 if img.mode != 'L':
~\Anaconda3\envs\tensorflow\lib\site-packages\PIL\Image.py in open(fp, mode)
2547 fp.seek(0)
2548 except (AttributeError, io.UnsupportedOperation):
-> 2549 fp = io.BytesIO(fp.read())
2550 exclusive_fp = True
2551
AttributeError: 'numpy.ndarray' object has no attribute 'read'
Может кто-нибудь рассказать мне, как решить, или какой другой лучший способ реализовать то же самое.