Add files from MNIST neural network

Neural Network - MNIST/Imagem para Array 28x28 MNIST.py

@@ -0,0 +1,48 @@
+from PIL import Image, ImageFilter
+import matplotlib.pyplot as plt
+import numpy as np
+
+def imageprepare(img):
+    """
+    This function returns the pixel values.
+    The imput is an image file location.
+    """
+    im = Image.open(img).convert('L')
+    width = float(im.size[0])
+    height = float(im.size[1])
+    newImage = Image.new('L', (28, 28), (255))  # creates white canvas of 28x28 pixels
+
+    if width > height:  # check which dimension is bigger
+        # Width is bigger. Width becomes 20 pixels.
+        nheight = int(round((20.0 / width * height), 0))  # resize height according to ratio width
+        if (nheight == 0):  # rare case but minimum is 1 pixel
+            nheight = 1
+            # resize and sharpen
+        img = im.resize((20, nheight), Image.ANTIALIAS).filter(ImageFilter.SHARPEN)
+        wtop = int(round(((28 - nheight) / 2), 0))  # calculate horizontal position
+        newImage.paste(img, (4, wtop))  # paste resized image on white canvas
+    else:
+        # Height is bigger. Heigth becomes 20 pixels.
+        nwidth = int(round((20.0 / height * width), 0))  # resize width according to ratio height
+        if (nwidth == 0):  # rare case but minimum is 1 pixel
+            nwidth = 1
+            # resize and sharpen
+        img = im.resize((nwidth, 20), Image.ANTIALIAS).filter(ImageFilter.SHARPEN)
+        wleft = int(round(((28 - nwidth) / 2), 0))  # caculate vertical pozition
+        newImage.paste(img, (wleft, 4))  # paste resized image on white canvas
+
+    # newImage.save("sample.png
+
+    tv = list(newImage.getdata())  # get pixel values
+
+    # normalize pixels to 0 and 1. 0 is pure white, 1 is pure black.
+    tva = [(255 - x) * 1.0 / 255.0 for x in tv]
+    #print(tva)
+    return tva
+
+img_array = imageprepare('./test3.jpeg')#file path here
+# mnist IMAGES are 28x28=784 pixels
+print(len(x))
+img_array_reshaped = np.array(img_array).reshape(28,28)
+
+plt.imshow(img_array_reshaped, cmap='gist_gray')

Neural Network - MNIST/Mnist_DEVPROCESS.py

@@ -0,0 +1,81 @@
+# https://github.com/keras-team/keras/blob/master/examples/mnist_mlp.py
+# https://nextjournal.com/schmudde/ml4a-mnist
+# https://github.com/keras-team/keras/blob/master/examples/mnist_cnn.py (CONVOLUTIONAL)
+# https://github.com/Hvass-Labs/TensorFlow-Tutorials/blob/master/01_Simple_Linear_Model.ipynb (TF)
+
+import keras
+from keras.datasets import mnist
+from keras.models import Sequential
+from keras.layers import Dense, Dropout
+from keras.optimizers import RMSprop
+import matplotlib.pyplot as plt
+import numpy as np
+import random
+
+'''
+TRY:
+-> Loss functions:
+Log loss 
+KL Divergence
+Mean Squared Error
+Categorical_crossentropy (atual)
+
+-> Optimizers:
+RMSprop (atual)
+
+
+'''
+
+
+batch_size = 128
+num_classes = 10 # {0, 1, 2, 3, 4, 5, 6, 7, 8, 9} = 10
+epochs = 20
+
+(x_train, y_train), (x_test, y_test) = mnist.load_data()
+
+plt.figure()
+numbers = np.concatenate([np.concatenate([x_train[i] for i in [int(random.random() * len(x_train)) for i in range(9)]], axis=1) for i in range(9)], axis=0)
+plt.imshow(numbers, cmap='gist_gray', interpolation='none')
+plt.xticks([])
+plt.yticks([])
+plt.xlabel('Alguns números do MNIST')
+plt.show()
+
+'''
+r = int(random.random() * len(x_train))
+numbers2 = np.array([x_train[x] for x in range(r-5, r)])
+plt.xlabel('Alguns números do MNIST, começando de ' + str(r-5))
+plt.imshow(numbers2.reshape(int(numbers2.size/28), 28), cmap='gist_gray')
+'''
+
+x_train = x_train.reshape(60000, 784)
+x_test = x_test.reshape(10000, 784)
+x_train = x_train.astype('float32')
+x_test = x_test.astype('float32')
+
+x_train /= 255 # Cores [0, 255] -> [0, 1]
+x_test /= 255 
+
+print(x_train.shape[0], 'números/arquivos de treino')
+print(x_test.shape[0], 'números/arquivos de teste')
+
+y_train = keras.utils.to_categorical(y_train, num_classes)
+y_test = keras.utils.to_categorical(y_test, num_classes)
+
+classifier = Sequential()
+classifier.add(Dense(512, activation='relu', input_shape=(784,)))
+classifier.add(Dropout(0.2))
+classifier.add(Dense(512, activation='relu'))
+classifier.add(Dropout(0.2))
+classifier.add(Dense(num_classes, activation='softmax'))
+
+#classifier.summary()
+
+classifier.compile(loss='categorical_crossentropy', optimizer=RMSprop(), metrics=['accuracy'])
+
+classifier.fit(x_train, y_train, batch_size=batch_size, epochs=epochs, verbose=1, validation_data=(x_test, y_test))
+
+score = classifier.evaluate(x_test, y_test, verbose=1)
+
+print('Test loss:', score[0])
+print('Test accuracy:', score[1])