def __init__(self,
num_hidden_units=256,
global_pooling='avg',
regularizer_val=0.0001):
# Network parameters
self._num_hidden_units = num_hidden_units
self._regularizer_value = regularizer_val
self._regularizer = regularizers.l2(regularizer_val)
self._global_pooling = global_pooling
def performCrossValidation(x_data, y_data):
_metrics = metrics.metrics(DIMENSIONS, 2)
for seed in range(ROUNDS):
np.random.seed(seed)
kfold = KFold(n_splits=KFOLDS, shuffle=True, random_state=seed)
for train, test in kfold.split(x_data, y_data):
from keras import backend as K
K.clear_session()
classes_train = pd.concat(
[y_data[train], pd.get_dummies(y_data[train])],
axis=1).drop(['Prior_Emotion'], axis=1)
classes_test = pd.concat(
[y_data[test], pd.get_dummies(y_data[test])],
axis=1).drop(['Prior_Emotion'], axis=1)
input_shape = sentence_enISEAR.shape[1] # feature count
model = Sequential()
model.add(
Dense(7,
input_shape=(input_shape, ),
activity_regularizer=regularizers.l2(0.01)))
model.add(Activation('sigmoid'))
model.compile(loss='binary_crossentropy',
metrics=['accuracy'],
optimizer='adam')
model.fit(sentence_enISEAR[train],
vectors[train],
batch_size=BATCH_SIZE,
epochs=EPOCHS,
verbose=VERBOSITY,
class_weight=class_weight)
vectors_test = vectors[test]
ids_test = ids[test]
weights = [0.5, 0.5, 0.5, 0.5, 0.5, 0.5, 0.5]
preds = model.predict(sentence_enISEAR[test])
results = []
for row in range(len(preds)):
res = []
for dim in range(len(DIMENSIONS)):
value = preds[row][dim]
if (value >= weights[dim]):
value = 1
else:
value = 0
res.append(value)
results.append(res)
predicted = np.array(results)
_metrics.evaluateFold(predicted, vectors_test)
_metrics.showFinalResults(args.format)
def action(): #此函数是控制整个采样，训练， 匹配测试的过程。
global filename, total_EEG_data_number_for_test
global step, directly_load_model_flag, openfile_name, directly_load_filename
filename = ui.lineEdit.text() #写入文件名
if step == 3: # 匹配测试
disable_Gui()
ui.label.setText("匹配测试中...")
ui.label.repaint()
ui.pushButton.setText("匹配测试中...")
ui.pushButton.repaint()
if directly_load_model_flag == False:
model = load_model(filename + '_key.h5')
if directly_load_model_flag == True:
model = load_model(directly_load_filename)
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect((server_address, 5204)) #5204 为 OPENBCI GUI的缺省服务器段的发送端口
EEG_data_for_test = np.zeros(
[total_EEG_data_number_for_test,
total_EEG_Features]) # EEG_data_for_test为采样的测试EEG帧。
temp_total_EEG_data_number_for_test = total_EEG_data_number_for_test
total_EEG_data_number_for_test = 1 #把total_EEG_data_number_for_test赋值给temp_total_EEG_data_number_for_test后，total_EEG_data_number_for_test变为1，只是为了GUI的进度条显示测试的进度%。
match_counter = 0 # 用于计数EEG通过测试的个数。
for times in range(
temp_total_EEG_data_number_for_test): #这个大循环，内嵌一个1的小循环。
for k in range(total_EEG_data_number_for_test):
EEG_data_one_line = (s.recv(1024).decode('utf-8')).split(
'A') ####按照字符"A"来截断每一条EEG数据，分割成24小份
for i in range(total_EEG_Features):
if len(EEG_data_one_line
) == 25: #这个判断是为了避免有时候读取EEG时候，遇到换行符丢失的现象。
EEG_data_for_test[k][i] = float(EEG_data_one_line[i])
else:
EEG_data_for_test[k][i] = EEG_data_for_test[k - 1][i]
test = model.predict(EEG_data_for_test, verbose=1)
for k in range(total_EEG_data_number_for_test):
if test[k] >= match_triger:
match_counter += 1
ui.progressBar.setProperty(
"value",
(times + 1) / temp_total_EEG_data_number_for_test * 100)
result_text = "匹配率为" + str(
match_counter / temp_total_EEG_data_number_for_test *
100) + "%"
ui.label.setText(result_text)
ui.label.repaint()
ui.lcdNumber.display(match_counter /
temp_total_EEG_data_number_for_test * 100)
QApplication.processEvents() #用于PyQt界面的刷新，保证流畅程度。
result_text = "测试结束，最终匹配率为" + str(
match_counter / temp_total_EEG_data_number_for_test * 100) + "%"
ui.label.setText(result_text)
ui.label.repaint()
total_EEG_data_number_for_test = temp_total_EEG_data_number_for_test #重新恢复全局变量total_EEG_data_number_for_test的值。
ui.pushButton.setText("开始匹配测试")
ui.pushButton.repaint()
enable_Gui()
if step == 2: #机器学习
disable_Gui()
ui.label.setText("开始机器学习你的脑纹。")
ui.label.repaint()
ui.pushButton.setText("2-设置中...")
ui.pushButton.repaint()
######################################### 开始训练 ######################################
####################### 读取代训练的EEG数据############################
f = open(filename + 'EEG.txt', 'r') #读取代训练的EEG数据
All_EEG_data_lines = f.readlines()
EEG_data_A = np.zeros([total_EEG_data_number, total_EEG_Features])
for k in range(total_EEG_data_number):
EEG_data_one_line = (All_EEG_data_lines[k].split('A')
) ####按照字符"A"来截断每一条EEG数据，分割成24小份
for i in range(total_EEG_Features):
if len(EEG_data_one_line
) == 25: #这个判断是为了避免有时候读取EEG时候，遇到换行符丢失的现象。
EEG_data_A[k][i] = float(EEG_data_one_line[i])
else:
EEG_data_A[k][i] = EEG_data_A[k - 1][i]
f.close()
###############读取代训练的EEG数据完毕######################
###############开始读取代陪训的Random 1/2 EEG数据######################
f = open('random_EEG_1.txt', 'r') #这是用于参与训练的他人EEG
All_EEG_data_lines = f.readlines()
EEG_data_B = np.zeros([total_EEG_data_number, total_EEG_Features])
for k in range(total_EEG_data_number):
EEG_data_one_line = (All_EEG_data_lines[k].split('A')
) ####按照字符"A"来截断每一条EEG数据，分割成24小份
for i in range(total_EEG_Features):
if len(EEG_data_one_line
) == 25: #这个判断是为了避免有时候读取EEG时候，遇到换行符丢失的现象。
EEG_data_B[k][i] = float(EEG_data_one_line[i])
else:
EEG_data_B[k][i] = EEG_data_B[k - 1][i]
f.close()
###############读取random_EEG_1完毕######################
f = open('random_EEG_2.txt', 'r') #这也是用于参与训练的他人EEG
All_EEG_data_lines = f.readlines()
EEG_data_C = np.zeros([total_EEG_data_number, total_EEG_Features])
for k in range(total_EEG_data_number):
EEG_data_one_line = (All_EEG_data_lines[k].split('A')
) ####按照字符"A"来截断每一条EEG数据，分割成24小份
for i in range(total_EEG_Features):
if len(EEG_data_one_line
) == 25: #这个判断是为了避免有时候读取EEG时候，遇到换行符丢失的现象。
EEG_data_C[k][i] = float(EEG_data_one_line[i])
else:
EEG_data_C[k][i] = EEG_data_C[k - 1][i]
f.close()
###############读取random_EEG_2完毕######################
##################读取代陪训的Random 1/2 EEG数据######################
########################开始合成总的数据样本（包括待训练数据，以及两个陪训数据#############################
y0 = np.ones([total_EEG_data_number, 1])
y1 = np.zeros([total_EEG_data_number, 1])
y2 = np.zeros([total_EEG_data_number, 1])
x = np.vstack((EEG_data_A, EEG_data_B, EEG_data_C))
y = np.vstack((y0, y1, y2))
########################开始搭建神经网络#############################
model = Sequential()
model.add(Dense(240, input_shape=(24, )))
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(
Dense(240,
kernel_regularizer=regularizers.l2(0.01),
bias_regularizer=regularizers.l2(0.01)))
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(
Dense(240,
kernel_regularizer=regularizers.l2(0.01),
bias_regularizer=regularizers.l2(0.01)))
model.add(Activation('relu'))
model.add(Dropout(0.2))
model.add(Dense(1))
model.add(Activation('sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer='rmsprop',
metrics=['accuracy'])
########################神经网络搭建完毕#############################
################这个tb，是为了使用TensorBoard########################
tb = TensorBoard(
log_dir='./logs', # log 目录
histogram_freq=0, # 按照何等频率（epoch）来计算直方图，0为不计算
batch_size=32, # 用多大量的数据计算直方图
write_graph=True, # 是否存储网络结构图
write_grads=False, # 是否可视化梯度直方图
write_images=False, # 是否可视化参数
embeddings_freq=0,
embeddings_layer_names=None,
embeddings_metadata=None)
# 在命令行，先conda activate envs，然后进入本代码所在的目录，然后用 tensorboard --logdir=logs/ 来看log
# 然后打开chrome浏览器，输入http://localhost:6006/ 来查看
# 如果出现tensorboard错误，那么需要修改 ...\lib\site-packages\tensorboard\manager.py，其中keras环境下的这个文件，我已经修改好了。
########################开始训练#############################
for i in range(int(100 / 100)): #这个for，是1次性训练，是为了在TensorBoard中查看训练情况。
#for i in range(100): #这个for，只是为了进度条的显示，所以分成 100次来训练。
model.fit(x,
y,
epochs=int(training_times / 100 * 100),
batch_size=training_batch_size,
verbose=1,
shuffle=True,
callbacks=[tb]) #这一行带callbacks，是为了使用TensorBoard
#model.fit(x, y, epochs=int(training_times/100), batch_size=training_batch_size,verbose = 1,shuffle=True) #这一行带callbacks，所以无法使用TensorBoard
ui.progressBar.setProperty("value", i + 1)
QApplication.processEvents() #用于PyQt界面的刷新，保证流畅程度。
model.save(filename + '_key.h5')
########################训练完毕，并保存训练好的神经网络#############################
ui.label.setText("你的脑纹锁设置成功！")
ui.label.repaint()
ui.pushButton.setText("开始匹配测试")
ui.pushButton.repaint()
ui.progressBar.setProperty("value", 0)
step = 3
ui.label_2.setText("目前载入的是" + filename + "的已经训练好的脑纹")
ui.label_2.repaint()
enable_Gui()
if step == 1: #### 录制脑电波
disable_Gui()
ui.label.setText("开始录制你的脑纹信息，请保持不动。")
ui.label.update()
ui.pushButton.setText("1-录制脑纹中...")
ui.pushButton.repaint()
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
s.connect((server_address, 5204))
f = open(filename + 'EEG.txt', 'w')
for k in range(total_EEG_data_number):
EEG_data = (s.recv(1024).decode('utf-8'))
EEG_data_to_write = EEG_data + '\r\n'
f.write(EEG_data_to_write)
ui.progressBar.setProperty("value",
k / total_EEG_data_number * 100 + 2)
QApplication.processEvents() #用于PyQt界面的刷新，保证流畅程度。
f.close()
step = 2
ui.label_2.update()
ui.label.setText("接下来神经网络开始学习并设置你的脑纹。")
ui.label.update()
ui.pushButton.setText("2-开始机器学习")
ui.pushButton.repaint()
enable_Gui()
if step == 0:
ui.label.setText("返回第一步，请输入你的名字。")
ui.label.update()
ui.pushButton.setText("1-重新开始录制脑纹")
ui.pushButton.repaint()
ui.label_2.setText("目前没有载入任何已经训练好的脑纹。")
ui.label_2.repaint()
step = 1