emotion.py

import tensorflow as tf
import h5py
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Conv2D, MaxPooling2D, AveragePooling2D
from tensorflow.keras.layers import Dense, Activation, Dropout, Flatten

from tensorflow.keras.preprocessing import image
from tensorflow.keras.preprocessing.image import ImageDataGenerator

import numpy as np
import matplotlib.pyplot as plt
import os


sess = tf.Session() 
#variables
num_classes = 7 #angry, disgust, fear, happy, sad, surprise, neutral
batch_size = 256
epochs = 5
#------------------------------


# In[3]:


with open("/Users/divalicious/Desktop/Emotional Analysis/challenges-in-representation-learning-facial-expression-recognition-challenge/fer2013/fer2013/fer2013.csv") as f:
    content = f.readlines()
 
lines = np.array(content)
 
num_of_instances = lines.size
print("number of instances: ",num_of_instances)


# In[7]:


x_train, y_train, x_test, y_test = [], [], [], []
 
for i in range(1,num_of_instances):
 #try:
  emotion, img, usage = lines[i].split(",")
 
  val = img.split(" ")
  pixels = np.array(val, 'float32')
 
  emotion = tf.keras.utils.to_categorical(emotion, num_classes)
 
  if 'Training' in usage:
       y_train.append(emotion)
       x_train.append(pixels)
  elif 'PublicTest' in usage:
       y_test.append(emotion)
       x_test.append(pixels)
 #except:
 # print("", end="")


#------------------------------
#data transformation for train and test sets
x_train = np.array(x_train, 'float32')
y_train = np.array(y_train, 'float32')
x_test = np.array(x_test, 'float32')
y_test = np.array(y_test, 'float32')

x_train /= 255 #normalize inputs between [0, 1]
x_test /= 255

x_train = x_train.reshape(x_train.shape[0], 48, 48, 1)
x_train = x_train.astype('float32')
x_test = x_test.reshape(x_test.shape[0], 48, 48, 1)
x_test = x_test.astype('float32')

print(x_train.shape[0], 'train samples')
print(x_test.shape[0], 'test samples')


# In[13]:


model = Sequential()
 
#1st convolution layer
model.add(Conv2D(64, (5, 5), activation='relu', input_shape=(48,48,1)))
model.add(MaxPooling2D(pool_size=(5,5), strides=(2, 2)))
 
#2nd convolution layer
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(AveragePooling2D(pool_size=(3,3), strides=(2, 2)))
 
#3rd convolution layer
model.add(Conv2D(128, (3, 3), activation='relu'))
model.add(Conv2D(128, (3, 3), activation='relu'))
model.add(AveragePooling2D(pool_size=(3,3), strides=(2, 2)))
 
model.add(Flatten())
 
#fully connected neural networks
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.2))
 
model.add(Dense(num_classes, activation='softmax'))


# In[17]:


gen = ImageDataGenerator()
train_generator = gen.flow(x_train, y_train, batch_size=batch_size)
 
model.compile(loss='categorical_crossentropy'
, optimizer=tf.keras.optimizers.Adam()
, metrics=['accuracy']
)
 
#model.summary()

checkpoint_path = "/Users/divalicious/Desktop/cp.ckpt"
checkpoint_dir = os.path.dirname(checkpoint_path)

# Create a callback that saves the model's weights
cp_callback = tf.keras.callbacks.ModelCheckpoint(filepath=checkpoint_path,save_weights_only=True,verbose=1)

model.save_weights(checkpoint_path)

model.fit_generator(train_generator, steps_per_epoch=batch_size, epochs=epochs, callbacks=[cp_callback])

model.save('my_model.h5')
# In[ ]:


train_score = model.evaluate(x_train, y_train, verbose=0)
print('Train loss:', train_score[0])
print('Train accuracy:', 100*train_score[1])
 
test_score = model.evaluate(x_test, y_test, verbose=0)
print('Test loss:', test_score[0])
print('Test accuracy:', 100*test_score[1])


# In[18]:


def emotion_analysis(emotions):
    objects = ('angry', 'disgust', 'fear', 'happy', 'sad', 'surprise', 'neutral')
    y_pos = np.arange(len(objects))

    plt.bar(y_pos, emotions, align='center', alpha=0.5)
    plt.xticks(y_pos, objects)
    plt.ylabel('percentage')
    plt.title('emotion')

    plt.show()



img = tf.keras.preprocessing.image.load_img("/Users/divalicious/Desktop/test.jpeg", grayscale=True, target_size=(48, 48))
 
x = tf.keras.preprocessing.image.img_to_array(img)
x = np.expand_dims(x, axis = 0)
 
x /= 255
 
custom = model.predict(x)
emotion_analysis(custom[0])
 
x = np.array(x, 'float32')
x = x.reshape([48, 48]);
 
plt.gray()
plt.imshow(x)
plt.show()