人臉表情識別 深度神經網絡 python實現 簡單模型 fer2013數據集

參考網址：https://sefiks.com/2018/01/01/facial-expression-recognition-with-keras/

1.數據集介紹及處理：

（1)　　數據集Fer2013下載地址為：https://www.kaggle.com/c/challenges-in-representation-learning-facial-expression-recognition-challenge/data

　　該數據集中每張圖片的像素為48*48，該數據集用excel讀取后顯示的格式如下圖所示：

              

第一列為標簽（也即為什么表情），第二列為像素值，第三列是代表該圖片是訓練集還是測試集，已經給你打亂了。只需要用即可

(2)pandas讀取數據集　　

import numpy as np 
import pandas as pd

data = pd.read_csv('data/fer2013/fer2013.csv')
num_of_instances = len(data) #獲取數據集的數量
print("數據集的數量為：",num_of_instances)

pixels = data['pixels']
emotions = data['emotion']
usages = data['Usage']

（3）分離訓練集和測試集

num_classes = 7   #表情的類別數目
x_train,y_train,x_test,y_test = [],[],[],[]

for emotion,img,usage in zip(emotions,pixels,usages):    
    try: 
        emotion = keras.utils.to_categorical(emotion,num_classes)   # 獨熱向量編碼
        val = img.split(" ")
        pixels = np.array(val,'float32')
        
        if(usage == 'Training'):
            x_train.append(pixels)
            y_train.append(emotion)
        elif(usage == 'PublicTest'):
            x_test.append(pixels)
            y_test.append(emotion)
    except:
        print("",end="")

（4）把數據集轉換為numpy數組格式，方便后續處理

x_train = np.array(x_train)
y_train = np.array(y_train)
x_train = x_train.reshape(-1,48,48,1)
x_test = np.array(x_test)
y_test = np.array(y_test)
x_test = x_test.reshape(-1,48,48,1)

(5)顯示其中的前4張圖片

import matplotlib.pyplot as plt
%matplotlib inline

for i in range(4): 
    plt.subplot(221+i)
    plt.gray()
    plt.imshow(x_train[i].reshape([48,48]))

 

 2. 創建網絡 進行訓練和測試

from keras.models import Sequential
from keras.layers import Conv2D,MaxPool2D,Activation,Dropout,Flatten,Dense
from keras.optimizers import Adam
from keras.preprocessing.image import ImageDataGenerator

batch_size = 8
epochs = 20

model = Sequential()

#第一層卷積層
model.add(Conv2D(input_shape=(48,48,1),filters=32,kernel_size=3,padding='same',activation='relu'))
model.add(Conv2D(filters=32,kernel_size=3,padding='same',activation='relu'))
model.add(MaxPool2D(pool_size=2, strides=2))

#第二層卷積層
model.add(Conv2D(filters=64,kernel_size=3,padding='same',activation='relu'))
model.add(Conv2D(filters=64,kernel_size=3,padding='same',activation='relu'))
model.add(MaxPool2D(pool_size=2, strides=2))

#第三層卷積層
model.add(Conv2D(filters=128,kernel_size=3,padding='same',activation='relu'))
model.add(Conv2D(filters=128,kernel_size=3,padding='same',activation='relu'))
model.add(MaxPool2D(pool_size=2, strides=2))

model.add(Flatten())

#全連接層
model.add(Dense(64,activation = 'relu'))
model.add(Dropout(0.5))
model.add(Dense(7,activation = 'softmax'))

#進行訓練
model.compile(loss = 'categorical_crossentropy',optimizer = Adam(),metrics=['accuracy'])
model.fit(x_train,y_train,batch_size=batch_size,epochs=epochs)


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])

這是一種通用識別架構，由於我的電腦配置不行，程序正在訓練，不再貼運行結果。可自行修改網絡架構。

程序中需要注意的地方：同時遍歷多個數組或列表時，可用zip()函數進行遍歷。