以fer2013为数据集,利用cnn的python人脸表情识别代码
时间: 2023-07-30 13:06:52 浏览: 74
好的,以下是使用fer2013数据集和CNN模型进行人脸表情识别的Python代码:
```python
import pandas as pd
import numpy as np
import keras
from keras.utils import np_utils
from keras.models import Sequential
from keras.layers import Conv2D, MaxPooling2D, Dropout, Flatten, Dense
# 加载数据集
data = pd.read_csv('/path/to/fer2013.csv')
# 将像素值转换为图像
pixels = data['pixels'].tolist()
images = []
for pixel_sequence in pixels:
image = [int(pixel) for pixel in pixel_sequence.split(' ')]
image = np.asarray(image).reshape(48, 48)
images.append(image)
images = np.asarray(images)
images = np.expand_dims(images, -1)
# 将标签转换为独热编码
labels = np_utils.to_categorical(data['emotion'])
# 划分训练集、验证集和测试集
train_images, val_images, test_images = images[:25000], images[25000:30000], images[30000:]
train_labels, val_labels, test_labels = labels[:25000], labels[25000:30000], labels[30000:]
# 构建CNN模型
model = Sequential()
model.add(Conv2D(32, (3, 3), activation='relu', input_shape=(48, 48, 1)))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Conv2D(128, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Conv2D(128, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(1024, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(7, activation='softmax'))
# 编译模型
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
# 训练模型
model.fit(train_images, train_labels, batch_size=64, epochs=50, validation_data=(val_images, val_labels))
# 在测试集上评估模型
score = model.evaluate(test_images, test_labels, verbose=0)
print('Test loss:', score[0])
print('Test accuracy:', score[1])
```
这个代码使用了Keras库来构建CNN模型。首先,我们将fer2013数据集加载到Pandas DataFrame中,并将像素值转换为图像。然后,我们将标签转换为独热编码,并将数据集划分为训练集、验证集和测试集。接下来,我们构建CNN模型,并编译它。最后,我们在训练集和验证集上训练模型,并在测试集上评估模型的性能。
请注意,这只是一个简单的示例代码,您可以根据需要进行修改和优化。