人脸识别

#第一步数据的加载  2准备train labe  >>进行训练 #3 cnn  4 检测
import tensorflow as tf
import numpy as np
import scipy.io as sio
f=open(r'C:\Users\Administrator\Desktop\火焰识别\人脸识别库\Yale_64x64.mat','rb')#完成人脸数据库的加载
mdict=sio.loadmat(f)#这时候就把网上的数据弄进来了
train_data=mdict['fea']
train_label=mdict['gnd']
train_data=np.random.permutation(train_data)
train_label=np.random.permutation(train_label)#这两步完成数据的打乱
test_data=train_data[0:64]
test_label=train_label[0:64]
np.random.seed(100)
test_data=np.random.permutation(test_data)#%%permutation是进行随机排列
np.random.seed(100)
test_label=np.random.permutation(test_label)  ##到这里就已经完成了数据的分配和加载
#下来完成第二步
train_data=train_data.reshape(train_data.shape[0],64,64,1).astype(np.float32)/255  #n*64*64*1 转换为0-1之间
train_labels_new=np.zeros((165,15))
for i in range(0,165):
    j=int(train_label[i,0])-1  #1-15 >>>0-14  one-hot向量
    train_labels_new[i,j]=1
test_data_input=test_data.reshape(test_data.shape[0],64,64,1).astype(np.float32)/255  #n*64*64*1 转换为0-1之间
test_labels_input=np.zeros((64,15))
for i in range(0,64):
    j=int(train_label[i,0])-1  #1-15 >>>0-14
    test_labels_input[i,j]=1  
data_input=tf.placeholder(tf.float32,[None,64,64,1])
label_input=tf.placeholder(tf.float32,[None,15])
#这时候样本准备已经完成 下来进行一层的卷积神经网络
layer1=tf.layers.conv2d(inputs=data_input,filters=32,kernel_size=2,strides=1,padding='SAME',activation=tf.nn.relu)
layer1_pool=tf.layers.max_pooling2d(layer1,pool_size=2,strides=2)
layer2=tf.reshape(layer1_pool,[-1,32*32*32])
layer2_relu=tf.layers.dense(layer2,1024,tf.nn.relu)
output=tf.layers.dense(layer2_relu,15)

#人工神经网络已经搭建完成
loss=tf.losses.softmax_cross_entropy(onehot_labels=label_input,logits=output)
train=tf.train.GradientDescentOptimizer(0.01).minimize(loss)
accuracy=tf.metrics.accuracy(labels=tf.argmax(label_input,axis=1),predictions=tf.argmax(output,axis=1))[1]

#run 
init=tf.group(tf.global_variables_initializer(),tf.local_variables_initializer())
with tf.Session() as sess:
    sess.run(init)
    for i in range(0,200):
        train_data_input=np.array(train_data)
        train_label_input=np.array(train_labels_new)
        sess.run([train,loss],feed_dict={data_input:train_data_input,label_input:train_label_input})
        acc=sess.run(accuracy,feed_dict={data_input:test_data_input,label_input:test_labels_input})
        print('acc%2.f',acc)