import pandas as pd

import numpy as np
import os
from glob import glob
import random
import matplotlib.pylab as plt
../input/breast-histopathology-images
mypaths=[]
for name in glob('../input/breast-histopathology-
images/IDC_regular_ps50_idx5/*',recursive=True):
mypaths.append(name)
print(mypaths[:5])
['../input/breast-histopathology-images/IDC_regular_ps50_idx5/10295',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/10304',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/12868',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/10274',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/12818']
len(mypaths)
279
mypaths[:5]
['../input/breast-histopathology-images/IDC_regular_ps50_idx5/10295',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/10304',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/12868',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/10274',
'../input/breast-histopathology-images/IDC_regular_ps50_idx5/12818']
mp=mypaths[60:100]
imagePatches=[]
for i in mp:
imagePatches+=glob(i+'/*/*.png', recursive=True)
len(imagePatches)
41604
# Two arrays holding images by class type

class0 = [] # 0 = no cancer
class1 = [] # 1 = cancer

for filename in imagePatches:

if filename.endswith("class0.png"):
class0.append(filename)
else:
class1.append(filename)
print(len(class0))
print("hi")
len(class1)
29338
hi
12266
sampled_class0 = random.sample(class0, len(class1))
sampled_class1 = random.sample(class1, len(class1))
len(sampled_class1)
12266
print(len(sampled_class0))
12266
from matplotlib.image import imread
import cv2

def get_image_arrays(data, label):

c=0
img_arrays = []
 for i in data:
if i.endswith('.png'):
img = cv2.imread(i ,cv2.IMREAD_GRAYSCALE)
img_sized = cv2.resize(img, (50, 50), interpolation=cv2.INTER_LINEAR)
img_arrays.append([img_sized, label])
c+=1
return img_arrays,c
class0_array,c0 = get_image_arrays(sampled_class0, 0)
class1_array,c1 = get_image_arrays(sampled_class1[:int(len(sampled_class1)/2)], 1)
print('done')
done
print(c0)
print(c1)
12266
6133
class0_array[0]
[array([[123, 147, 148, ..., 123, 117, 109],
[105, 119, 148, ..., 120, 107, 126],
[110, 127, 106, ..., 144, 160, 185],
...,
[232, 219, 213, ..., 132, 161, 164],
[230, 230, 215, ..., 141, 160, 167],
[229, 231, 227, ..., 129, 136, 148]], dtype=uint8),
0]

combined_data = np.concatenate((class0_array, class1_array))

#random.seed(200)
#random.shuffle(combined_data)
<string>:6: VisibleDeprecationWarning: Creating an ndarray from ragged nested
sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different
lengths or shapes) is deprecated. If you meant to do this, you must specify
'dtype=object' when creating the ndarray
len(combined_data)
18399
combined_data[0]
array([array([[123, 147, 148, ..., 123, 117, 109],
[105, 119, 148, ..., 120, 107, 126],
[110, 127, 106, ..., 144, 160, 185],
...,
[232, 219, 213, ..., 132, 161, 164],
[230, 230, 215, ..., 141, 160, 167],
[229, 231, 227, ..., 129, 136, 148]], dtype=uint8),
0], dtype=object)
X = []
y = []

for features,label in combined_data:

X.append(features)
y.append(label)
c0=0
c1=0
for i in y:
if i==0:
c0+=1
else:
c1+=1
print(c0)
print(c1)
12266
6133
len(y)
18399
X[400]
array([[185, 165, 172, ..., 183, 177, 178],
[158, 168, 176, ..., 169, 178, 193],
[153, 177, 172, ..., 176, 161, 192],
...,
[177, 173, 169, ..., 174, 176, 168],
[164, 199, 174, ..., 187, 172, 165],
[171, 177, 185, ..., 186, 181, 169]], dtype=uint8)
y[40]
0
X = np.array(X).reshape(-1, 50*50)
X.shape
(18399, 2500)
from sklearn.model_selection import train_test_split
from keras.utils.np_utils import to_categorical
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.40,
random_state=70)
y_train = to_categorical(y_train)
y_test = to_categorical(y_test)
print(X_train.shape, X_test.shape, y_train.shape, y_test.shape)
(11039, 2500) (7360, 2500) (11039, 2) (7360, 2)
cn=0
cp=0
for i in y_test:
if i[0]>i[1]:
cn+=1
else:
cp+=1
print(cp)
print(cn)
2432
4928
X_train[0].shape
(2500,)
import tensorflow as tf
from tensorflow import keras
len(X_train)
11039
#plt.matshow(X_train[0])
y_test[0]
array([1., 0.], dtype=float32)
X_train[0]
array([238, 234, 227, ..., 132, 110, 169], dtype=uint8)
model=keras.Sequential([
keras.layers.Dense(1250,input_shape=(2500,),activation='sigmoid'),
keras.layers.Dense(625,input_shape=(1250,),activation='sigmoid'),
keras.layers.Dense(300,input_shape=(625,),activation='sigmoid'),
keras.layers.Dense(150,input_shape=(300,),activation='sigmoid'),
keras.layers.Dense(75,input_shape=(150,),activation='sigmoid'),
keras.layers.Dense(37,input_shape=(75,),activation='sigmoid'),
keras.layers.Dense(18,input_shape=(37,),activation='sigmoid'),
keras.layers.Dense(9,input_shape=(18,),activation='sigmoid'),
keras.layers.Dense(4,input_shape=(9,),activation='sigmoid'),
keras.layers.Dense(2,input_shape=(4,),activation='sigmoid')
])
model.compile(
 optimizer=tf.optimizers.Adam(lr=0.01),
loss='binary_crossentropy',
metrics=['accuracy','Recall','Precision']
)
model.summary()
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense (Dense) (None, 1250) 3126250
_________________________________________________________________
dense_1 (Dense) (None, 625) 781875
_________________________________________________________________
dense_2 (Dense) (None, 300) 187800
_________________________________________________________________
dense_3 (Dense) (None, 150) 45150
_________________________________________________________________
dense_4 (Dense) (None, 75) 11325
_________________________________________________________________
dense_5 (Dense) (None, 37) 2812
_________________________________________________________________
dense_6 (Dense) (None, 18) 684
_________________________________________________________________
dense_7 (Dense) (None, 9) 171
_________________________________________________________________
dense_8 (Dense) (None, 4) 40
_________________________________________________________________
dense_9 (Dense) (None, 2) 10
=================================================================
Total params: 4,156,117
Trainable params: 4,156,117
Non-trainable params: 0
_________________________________________________________________
history=model.fit(X_train,y_train, validation_data=(X_test, y_test),epochs=10)
Epoch 1/10
345/345 [==============================] - 13s 33ms/step - loss: 0.6407 - accuracy:
0.6631 - recall: 0.6428 - precision: 0.6618 - val_loss: 0.6350 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 2/10
345/345 [==============================] - 10s 29ms/step - loss: 0.6400 - accuracy:
0.6625 - recall: 0.6625 - precision: 0.6625 - val_loss: 0.6345 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 3/10
345/345 [==============================] - 10s 29ms/step - loss: 0.6390 - accuracy:
0.6640 - recall: 0.6640 - precision: 0.6640 - val_loss: 0.6348 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 5/10
345/345 [==============================] - 10s 29ms/step - loss: 0.6428 - accuracy:
0.6589 - recall: 0.6589 - precision: 0.6589 - val_loss: 0.6346 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 6/10
345/345 [==============================] - 10s 30ms/step - loss: 0.6397 - accuracy:
0.6624 - recall: 0.6624 - precision: 0.6624 - val_loss: 0.6345 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 7/10
345/345 [==============================] - 10s 30ms/step - loss: 0.6394 - accuracy:
0.6628 - recall: 0.6628 - precision: 0.6628 - val_loss: 0.6345 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 8/10
345/345 [==============================] - 10s 30ms/step - loss: 0.6416 - accuracy:
0.6596 - recall: 0.6596 - precision: 0.6596 - val_loss: 0.6346 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 9/10
345/345 [==============================] - 10s 30ms/step - loss: 0.6396 - accuracy:
0.6630 - recall: 0.6630 - precision: 0.6630 - val_loss: 0.6346 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
Epoch 10/10
345/345 [==============================] - 10s 30ms/step - loss: 0.6398 - accuracy:
0.6627 - recall: 0.6627 - precision: 0.6627 - val_loss: 0.6351 - val_accuracy:
0.6696 - val_recall: 0.6696 - val_precision: 0.6696
#X_train=X_train/255
#X_test=X_test/255
#model.fit(X_train,y_train,epochs=100)
e=model.evaluate(X_test,y_test)
230/230 [==============================] - 2s 9ms/step - loss: 0.6351 - accuracy:
0.6696 - recall: 0.6696 - precision: 0.6696
y_test
array([[1., 0.],
[1., 0.],
[1., 0.],
...,
[1., 0.],
[1., 0.],
[1., 0.]], dtype=float32)
y_predicted=model.predict(X_test)
X_test[0]
array([144, 148, 137, ..., 118, 126, 136], dtype=uint8)
print(model.summary())
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense (Dense) (None, 1250) 3126250
_________________________________________________________________
dense_1 (Dense) (None, 625) 781875
_________________________________________________________________
dense_2 (Dense) (None, 300) 187800
_________________________________________________________________
dense_3 (Dense) (None, 150) 45150
_________________________________________________________________
dense_4 (Dense) (None, 75) 11325
_________________________________________________________________
dense_5 (Dense) (None, 37) 2812
_________________________________________________________________
dense_6 (Dense) (None, 18) 684
_________________________________________________________________
dense_7 (Dense) (None, 9) 171
_________________________________________________________________
dense_8 (Dense) (None, 4) 40
_________________________________________________________________
dense_9 (Dense) (None, 2) 10
=================================================================
Total params: 4,156,117
Trainable params: 4,156,117
Non-trainable params: 0
_________________________________________________________________
None
import matplotlib.pyplot as plt
#plotting the Accuracy of test and training sets
plt.plot(history.history['accuracy'])
plt.plot(history.history['val_accuracy'])
plt.title('Model Accuracy')
plt.ylabel('accuracy')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()

plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model Loss')
plt.ylabel('loss')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()

y_pred=model.predict(X_test)
Y_pred=[]
for i in y_pred:
if i[0]>i[1]:
Y_pred.append(0)
else:
Y_pred.append(1)
Y_test=[]
for i in y_test:
if i[0]>i[1]:
Y_test.append(0)
else:
Y_test.append(1)
from sklearn.metrics import classification_report, confusion_matrix
print('Confusion Matrix')
print(confusion_matrix(Y_test, Y_pred))
print('Classification Report')
print(classification_report(Y_test, Y_pred, target_names=['Negative','Positive']))
Confusion Matrix
[[4928 0]
[2432 0]]
Classification Report
precision recall f1-score support

Negative 0.67 1.00 0.80 4928

Positive 0.00 0.00 0.00 2432

accuracy 0.67 7360

macro avg 0.33 0.50 0.40 7360
weighted avg 0.45 0.67 0.54 7360

/opt/conda/lib/python3.7/site-packages/sklearn/metrics/_classification.py:1245:
UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0
in labels with no predicted samples. Use `zero_division` parameter to control this
behavior.
_warn_prf(average, modifier, msg_start, len(result))
/opt/conda/lib/python3.7/site-packages/sklearn/metrics/_classification.py:1245:
UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0
in labels with no predicted samples. Use `zero_division` parameter to control this
behavior.
_warn_prf(average, modifier, msg_start, len(result))
/opt/conda/lib/python3.7/site-packages/sklearn/metrics/_classification.py:1245:
UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0
in labels with no predicted samples. Use `zero_division` parameter to control this
behavior.
_warn_prf(average, modifier, msg_start, len(result))
confusionmatrix=confusion_matrix(Y_test, Y_pred)
classes=[0,1]
con_mat_df = pd.DataFrame(confusion_matrix(Y_test, Y_pred),
index = classes,
columns = classes)
con_mat_df
0 1
0 4928 0
1 2432 0
import seaborn as sns
figure = plt.figure(figsize=(6, 6))
sns.heatmap(con_mat_df, annot=True,cmap=plt.cm.cool,fmt='d')
plt.tight_layout()
plt.ylabel('Actual')
plt.xlabel('Predicted')
plt.show()

l=[[1501,88],[125,176]]
cm=np.array(l)
cm
array([[1501, 88],
[ 125, 176]])
classes=[0,1]
con_mat_df = pd.DataFrame(cm,
index = classes,
columns = classes)
import seaborn as sns
figure = plt.figure(figsize=(6, 6))
sns.heatmap(con_mat_df, annot=True,cmap=plt.cm.cool,fmt='d')
plt.tight_layout()
plt.ylabel('Actual')
plt.xlabel('Predicted')
plt.show()