def __init__(self):
super(MainWindow,self).__init__()
# 初始化参数
self.result = [0, 0]
# 初始化UI
self.setupUi(self)
self.center()
# 初始化画板
self.paintBoard = PaintBoard(self, Size = QSize(224, 224), Fill = QColor(0,0,0,255))
self.paintBoard.setPenColor(QColor(255,255,255,255))
self.dArea_Layout.addWidget(self.paintBoard)
self.clearDataArea()
def __init__(self):
# 初始化继承的父类(Qmainwindow)
super(MainWindow,self).__init__()
# 初始化参数
self.mode = MODE_MNIST
self.result = [0, 0]
# 初始化UI
self.setupUi(self)
self.center()
# 初始化画板，可以用鼠标来输入数据的
# 将画板变为gui画面的一部分。
self.paintBoard = PaintBoard(self, Size = QSize(224, 224), Fill = QColor(0,0,0,0))
self.paintBoard.setPenColor(QColor(0,0,0,0))
self.dArea_Layout.addWidget(self.paintBoard) # gui界面中加入paintboard这个部分
self.imgName = ''
self.clearDataArea() # 清除数据区域
png = QtGui.QPixmap('mnist-master\mnist-master\hj.jpg').scaled(160, 150)
self.lbDataArea1.setPixmap(png)
png1 = QtGui.QPixmap('mnist-master\mnist-master\hk.jpg').scaled(160, 115)
self.lbDataArea2.setPixmap(png1)
class MainWindow(QMainWindow, Ui_MainWindow):
def __init__(self):
super(MainWindow, self).__init__()
# 初始化参数
self.mode = MODE_MNIST
self.result = [0, 0]
# 初始化UI
self.setupUi(self)
self.center()
# 初始化画板
self.paintBoard = PaintBoard(self,
Size=QSize(224, 224),
Fill=QColor(0, 0, 0, 0))
self.paintBoard.setPenColor(QColor(0, 0, 0, 0))
self.dArea_Layout.addWidget(self.paintBoard)
self.clearDataArea()
# 窗口居中
def center(self):
# 获得窗口
framePos = self.frameGeometry()
# 获得屏幕中心点
scPos = QDesktopWidget().availableGeometry().center()
# 显示到屏幕中心
framePos.moveCenter(scPos)
self.move(framePos.topLeft())
# 窗口关闭事件
def closeEvent(self, event):
reply = QMessageBox.question(self, 'Message', "Are you sure to quit?",
QMessageBox.Yes | QMessageBox.No,
QMessageBox.Yes)
if reply == QMessageBox.Yes:
event.accept()
else:
event.ignore()
# 清除数据待输入区
def clearDataArea(self):
self.paintBoard.Clear()
self.lbDataArea.clear()
self.lbResult.clear()
self.lbCofidence.clear()
self.result = [0, 0]
"""
回调函数
"""
# 模式下拉列表回调
def cbBox_Mode_Callback(self, text):
if text == '1：MINIST随机抽取':
self.mode = MODE_MNIST
self.clearDataArea()
self.pbtGetMnist.setEnabled(True)
self.paintBoard.setBoardFill(QColor(0, 0, 0, 0))
self.paintBoard.setPenColor(QColor(0, 0, 0, 0))
elif text == '2：鼠标手写输入':
self.mode = MODE_WRITE
self.clearDataArea()
self.pbtGetMnist.setEnabled(False)
# 更改背景
self.paintBoard.setBoardFill(QColor(0, 0, 0, 255))
self.paintBoard.setPenColor(QColor(255, 255, 255, 255))
# 数据清除
def pbtClear_Callback(self):
self.clearDataArea()
# 识别
def pbtPredict_Callback(self):
__img, img_array = [], [
] # 将图像统一从qimage->pil image -> np.array [1, 1, 28, 28]
# 获取qimage格式图像
if self.mode == MODE_MNIST:
__img = self.lbDataArea.pixmap() # label内若无图像返回None
if __img == None: # 无图像则用纯黑代替
# __img = QImage(224, 224, QImage.Format_Grayscale8)
__img = Image.ImageQt(
Image.fromarray(np.uint8(np.zeros([224, 224]))))
else:
__img = __img.toImage()
elif self.mode == MODE_WRITE:
__img = self.paintBoard.getContentAsQImage()
# 转换成pil image类型处理
pil_img = Image.fromqimage(__img)
pil_img = pil_img.resize((28, 28), Image.ANTIALIAS)
# pil_img.save('test.png')
img_array = np.array(pil_img.convert('L')).reshape(1, 1, 28,
28) / 255.0
# img_array = np.where(img_array>0.5, 1, 0)
# reshape成网络输入类型
__result = network.predict(img_array) # shape:[1, 10]
# print (__result)
# 将预测结果使用softmax输出
__result = softmax(__result)
self.result[0] = np.argmax(__result) # 预测的数字
self.result[1] = __result[0, self.result[0]] # 置信度
self.lbResult.setText("%d" % (self.result[0]))
self.lbCofidence.setText("%.8f" % (self.result[1]))
# 随机抽取
def pbtGetMnist_Callback(self):
self.clearDataArea()
# 随机抽取一张测试集图片，放大后显示
img = x_test[np.random.randint(0, 9999)] # shape:[1,28,28]
img = img.reshape(28, 28) # shape:[28,28]
img = img * 0xff # 恢复灰度值大小
pil_img = Image.fromarray(np.uint8(img))
pil_img = pil_img.resize((224, 224)) # 图像放大显示
# 将pil图像转换成qimage类型
qimage = Image.ImageQt(pil_img)
# 将qimage类型图像显示在label
pix = QPixmap.fromImage(qimage)
self.lbDataArea.setPixmap(pix)
class MainWindow(QMainWindow,Ui_MainWindow):
def __init__(self):
# 初始化继承的父类(Qmainwindow)
super(MainWindow,self).__init__()
# 初始化参数
self.mode = MODE_MNIST
self.result = [0, 0]
# 初始化UI
self.setupUi(self)
self.center()
# 初始化画板，可以用鼠标来输入数据的
# 将画板变为gui画面的一部分。
self.paintBoard = PaintBoard(self, Size = QSize(224, 224), Fill = QColor(0,0,0,0))
self.paintBoard.setPenColor(QColor(0,0,0,0))
self.dArea_Layout.addWidget(self.paintBoard) # gui界面中加入paintboard这个部分
self.imgName = ''
self.clearDataArea() # 清除数据区域
png = QtGui.QPixmap('mnist-master\mnist-master\hj.jpg').scaled(160, 150)
self.lbDataArea1.setPixmap(png)
png1 = QtGui.QPixmap('mnist-master\mnist-master\hk.jpg').scaled(160, 115)
self.lbDataArea2.setPixmap(png1)
# 窗口居中，这次设置的是那个整体的窗口
def center(self):
# 获得窗口
framePos = self.frameGeometry()
# 获得屏幕中心点
scPos = QDesktopWidget().availableGeometry().center()
# 显示到屏幕中心,将窗口设置在屏幕中心
framePos.moveCenter(scPos)
self.move(framePos.topLeft())
# 窗口关闭事件，是否关闭窗口，点击右上角关闭窗口
def closeEvent(self, event):
reply = QMessageBox.question(self, 'Message',
"Are you sure to quit?", QMessageBox.Yes |
QMessageBox.No, QMessageBox.Yes)
if reply == QMessageBox.Yes:
event.accept()
else:
event.ignore()
# 清除数据待输入区
def clearDataArea(self):
self.paintBoard.Clear()#清除画板
self.lbDataArea.clear()
self.lbResult.clear() # 确认的数字
self.lbCofidence.clear() # softmax的值
self.result = [0, 0] # 一个用来存确认数字，一个用来存softmax值
"""
回调函数
"""
# 模式下拉列表回调，两种模式
def cbBox_Mode_Callback(self, text):
if text == '1：MINIST随机抽取':
print(text)
self.mode = MODE_MNIST
self.clearDataArea()
self.pbtGetMnist.setEnabled(True)
self.pbtjiazai.setEnabled(False)
self.paintBoard.setBoardFill(QColor(0,0,0,0))
self.paintBoard.setPenColor(QColor(0,0,0,0))
elif text == '2：鼠标手写输入':
#print(text)
self.mode = MODE_WRITE
self.clearDataArea()
self.pbtGetMnist.setEnabled(False)
self.pbtjiazai.setEnabled(False)
# 更改背景
self.paintBoard.setBoardFill(QColor(0,0,0,255))
self.paintBoard.setPenColor(QColor(255,255,255,255))
elif text == '3：手写图片识别':
#print('111')
self.mode = MODE_PICTURE
self.clearDataArea()
self.pbtGetMnist.setEnabled(False)
self.pbtjiazai.setEnabled(True)
# 更改背景
self.paintBoard.setBoardFill(QColor(0,0,0,0))
self.paintBoard.setPenColor(QColor(0,0,0,0))
# def get_picture(self):
# 数据清除，点击数据清除区域
def pbtClear_Callback(self):
self.clearDataArea()
# 识别
def pbtPredict_Callback(self):
__img, img_array =[],[] # 将图像统一从qimage->pil image -> np.array [1, 1, 28, 28]
# 获取qimage格式图像
if self.mode == MODE_MNIST: #随机抽取
__img = self.lbDataArea.pixmap() # label内若无图像返回None
if __img == None: # 无图像则用纯黑代替
# __img = QImage(224, 224, QImage.Format_Grayscale8)
__img = ImageQt.ImageQt(Image.fromarray(np.uint8(np.zeros([224,224]))))
else: __img = __img.toImage()
pil_img = ImageQt.fromqimage(__img)
pil_img = pil_img.resize((28, 28), Image.ANTIALIAS)
img_array = np.array(pil_img.convert('L')).reshape(1,1,28, 28) / 255.0
elif self.mode == MODE_WRITE:
__img = self.paintBoard.getContentAsQImage()
pil_img = ImageQt.fromqimage(__img)
pil_img = pil_img.resize((28, 28), Image.ANTIALIAS)
pil_img.save('test.png')
img_array = np.array(pil_img.convert('L')).reshape(1,1,28, 28) / 255.0
elif self.mode == MODE_PICTURE:
pic = image_prepare(self.imgName) # 得到28*28的数组
img_array = pic.reshape(1,1,28, 28) / 255.0 # 转换成神经网络要求的输入格式
'''
关于这部分有一点需要说明：
开始通过image_prepare得到预处理的图片之后，最开始是变成一维数组，然后一直想着变成二维数组并进行灰度反转。
为什么会这样想呢，因为我开始在变成reshape(1,1,28, 28)格式时采用的是网上的另一种方法，以至于格式基本是没有变化的。
其实直接变成一维数组然后进行上面的转换也是一样可以的，这个部分的失误浪费了很长时间！以后debug时需要多注意。
'''
# reshape成网络输入类型
__result = network.predict(img_array) # shape:[1, 10]
# print (__result)
# 将预测结果使用softmax输出，得到输出结果
__result = softmax(__result)
self.result[0] = np.argmax(__result) # 预测的数字
self.result[1] = __result[0, self.result[0]] # 置信度
# 结果显示在界面上
self.lbResult.setText("%d" % (self.result[0]))
self.lbCofidence.setText("%.8f" % (self.result[1]))
# 随机抽取
def pbtGetMnist_Callback(self):
self.clearDataArea()
# 随机抽取一张测试集图片，放大后显示，x_text为前面读取的数据集
# 随机从10000个照片中选择一张。
img = x_test[np.random.randint(0, 9999)] # shape:[1,28,28]
img = img.reshape(28, 28) # shape:[28,28]
img = img * 0xff # 恢复灰度值大小
pil_img = Image.fromarray(np.uint8(img))
pil_img = pil_img.resize((224, 224)) # 图像放大显示，和绘图窗口大小一致
# 将pil图像转换成qimage类型
qimage = ImageQt.ImageQt(pil_img)
# 将qimage类型图像显示在label
pix = QPixmap.fromImage(qimage)
self.lbDataArea.setPixmap(pix)
def pbtjiazai_Callback(self):
# 打开文件路径
# 设置文件扩展名过滤,注意用双分号间隔
self.imgName, imgType = QFileDialog.getOpenFileName(self, "打开图片", "", " *.png;;*.jpg;;*.jpeg;;*.bmp;;All Files (*)")
print(self.imgName) # imgName就是图片的路径，为了之后在识别的时候也可以用,写成了self.imgName,self代表创建的实例本身
png = QtGui.QPixmap(self.imgName).scaled(240, 240)
# 显示图片并设置图片格式
self.lbDataArea.setPixmap(png)
class MainWindow(QMainWindow,Ui_MainWindow):
def __init__(self):
super(MainWindow,self).__init__()
# 初始化参数
self.result = [0, 0]
# 初始化UI
self.setupUi(self)
self.center()
# 初始化画板
self.paintBoard = PaintBoard(self, Size = QSize(224, 224), Fill = QColor(0,0,0,255))
self.paintBoard.setPenColor(QColor(255,255,255,255))
self.dArea_Layout.addWidget(self.paintBoard)
self.clearDataArea()
# 窗口居中
def center(self):
# 获得窗口
framePos = self.frameGeometry()
# 获得屏幕中心点
scPos = QDesktopWidget().availableGeometry().center()
# 显示到屏幕中心
framePos.moveCenter(scPos)
self.move(framePos.topLeft())
# 窗口关闭事件
def closeEvent(self, event):
reply = QMessageBox.question(self, 'Message',
"你确定要退出吗?", QMessageBox.Yes |
QMessageBox.No, QMessageBox.Yes)
if reply == QMessageBox.Yes:
event.accept()
else:
event.ignore()
# 清除数据待输入区
def clearDataArea(self):
self.paintBoard.Clear()
self.lbDataArea.clear()
self.lbResult.clear()
self.lbCofidence.clear()
self.result = [0, 0]
#mnist.train.images是一个55000 * 784维的矩阵, mnist.train.labels是一个55000 * 10维的矩阵
# 数据清除
def pbtClear_Callback(self):
self.clearDataArea()
Gui = MainWindow()
Gui.show()
# 识别
def pbtPredict_Callback(self):
img, img_array =[],[] # 将图像统一从qimage->pil image -> np.array [1, 1, 28, 28]
# 获取qimage格式图像
img = self.paintBoard.getContentAsQImage()
# 转换成pil image类型处理
pil_img = ImageQt.fromqimage(img)
pil_img = pil_img.resize((28, 28), Image.ANTIALIAS)
pil_img.save('test.png')
pil_img = pil_img.convert('L')
tv = list(pil_img.getdata())
tva = [(255-x)*1.0/255.0 for x in tv]
#定义初始化权值函数
def weight_variable(shape):
initial=tf.truncated_normal(shape,stddev=0.1)
return tf.Variable(initial)
#定义初始化偏置函数
def bias_variable(shape):
initial=tf.constant(0.1,shape=shape)
return tf.Variable(initial)
#卷积层
def conv2d(input,filter):
return tf.nn.conv2d(input,filter,strides=[1,1,1,1],padding='SAME')
#池化层
def max_pool_2x2(value):
return tf.nn.max_pool(value,ksize=[1,2,2,1],strides=[1,2,2,1],padding='SAME')
keep_prob=tf.placeholder(tf.float32)
#p_keep_conv = tf.placeholder(tf.float32) # 卷积层的dropout概率
#p_keep_hidden = tf.placeholder(tf.float32)# 全连接层的dropout概率
#输入层
#定义两个placeholder
x=tf.placeholder(tf.float32,[None,784]) #28*28
y=tf.placeholder(tf.float32,[None,10])
#改变x的格式转为4维的向量[batch,in_hight,in_width,in_channels]
x_image=tf.reshape(x,[-1,28,28,1])
w0 = weight_variable([625,10]) #FC 128 inputs,10 outputs
w4 = weight_variable([128*4*4,625]) # FC 128 * 4 * 4 inputs, 625 outputs
w1 = weight_variable([625,128]) # FC 625 inputs, 128 outputs (labels)
#卷积、激励、池化操作
#初始化第一个卷积层的权值和偏置
W_conv1=weight_variable([3,3,1,32]) #3*3的采样窗口，32个卷积核从1个平面抽取特征
b_conv1=bias_variable([32])
#把x_image和权值向量进行卷积，再加上偏置值，然后应用于relu激活函数
h_conv1=tf.nn.relu(conv2d(x_image,W_conv1))
h_pool1=max_pool_2x2(h_conv1) #进行max_pooling 池化层
h_drop1=tf.nn.dropout(h_pool1,keep_prob)
#初始化第二个卷积层的权值和偏置
W_conv2=weight_variable([3,3,32,64]) #3*3的采样窗口，64个卷积核从32个平面抽取特征
b_conv2=bias_variable([64])
#把第一个池化层结果和权值向量进行卷积，再加上偏置值，然后应用于relu激活函数
h_conv2=tf.nn.relu(conv2d(h_pool1,W_conv2))
h_pool2=max_pool_2x2(h_conv2) #池化层
h_drop2=tf.nn.dropout(h_pool2,keep_prob)
#初始化第三个卷积层的权值和偏置
W_conv3=weight_variable([3,3,64,128]) #3*3的采样窗口，128个卷积核从64个平面抽取特征
b_conv3=bias_variable([128])
#把第二个池化层结果和权值向量进行卷积，再加上偏置值，然后应用于relu激活函数
h_conv3=tf.nn.relu(conv2d(h_pool2,W_conv3))
h_pool3=max_pool_2x2(h_conv3)
h_drop3=tf.nn.dropout(h_pool3,keep_prob)
#合并所有的feature map
h_feature = tf.reshape(h_drop3,[-1,w4.get_shape().as_list()[0]])
h_f = tf.nn.dropout(h_feature,keep_prob)
#全连接层
h_fc1 = tf.nn.relu(tf.matmul(h_f,w4))
h_fc2 = tf.nn.dropout(h_fc1,keep_prob)
pyx = tf.matmul(h_fc2,w0)
#交叉熵代价函数
cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(labels=y,logits=pyx))
#使用AdamOptimizer进行优化
train_step = tf.train.RMSPropOptimizer(0.001, 0.9).minimize(cross_entropy)
#结果放在bool集中
correct_prediction=tf.equal(tf.argmax(pyx,1),tf.argmax(y,1))
accuracy=tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
saver = tf.train.Saver()
#创建会话
with tf.Session() as sess:
sess.run(tf.global_variables_initializer()) #初始化变量
saver.restore(sess,model_file)
#测试
prediction = tf.argmax(pyx,1)
predint=prediction.eval(feed_dict={x: [tva],keep_prob: 1.0}, session=sess)
print(predint[0])
#acc=sess.run(accuracy,feed_dict={x:[tva],keep_prob:1.0})
# img_array = np.where(img_array>0.5, 1, 0)
# reshape成网络输入类型
# shape:[1, 10]
# print (__result)
# 将预测结果使用softmax输出
self.result[0] = predint[0] # 预测的数字
#self.result[1] = __result[0, self.result[0]] # 置信度
self.lbResult.setText("%d" % (self.result[0]))
class MainWindow(QMainWindow, Ui_MainWindow):
def __init__(self):
super(MainWindow, self).__init__()
# 初始化参数
self.mode = MODE_MNIST
self.result = [0, 0]
# 初始化UI
self.setupUi(self)
self.center()
# 初始化画板
self.paintBoard = PaintBoard(self,
Size=QSize(224, 224),
Fill=QColor(0, 0, 0, 0))
self.paintBoard.setPenColor(QColor(0, 0, 0, 0))
self.dArea_Layout.addWidget(self.paintBoard)
self.clearDataArea()
# 窗口居中
def center(self):
# 获得窗口
framePos = self.frameGeometry()
# 获得屏幕中心点
scPos = QDesktopWidget().availableGeometry().center()
# 显示到屏幕中心
framePos.moveCenter(scPos)
self.move(framePos.topLeft())
# 窗口关闭事件
def closeEvent(self, event):
reply = QMessageBox.question(self, 'Message', "Are you sure to quit?",
QMessageBox.Yes | QMessageBox.No,
QMessageBox.Yes)
if reply == QMessageBox.Yes:
event.accept()
else:
event.ignore()
# 清除数据待输入区
def clearDataArea(self):
self.paintBoard.Clear()
self.lbDataArea.clear()
self.lbResult.clear()
self.lbCofidence.clear()
self.result = [0, 0]
"""
回调函数
"""
# 模式下拉列表回调
def cbBox_Mode_Callback(self, text):
if text == '1：MINIST随机抽取':
self.mode = MODE_MNIST
self.clearDataArea()
self.pbtGetMnist.setEnabled(True)
self.paintBoard.setBoardFill(QColor(0, 0, 0, 0))
self.paintBoard.setPenColor(QColor(0, 0, 0, 0))
elif text == '2：鼠标手写输入':
self.mode = MODE_WRITE
self.clearDataArea()
self.pbtGetMnist.setEnabled(False)
# 更改背景
self.paintBoard.setBoardFill(QColor(0, 0, 0, 255))
self.paintBoard.setPenColor(QColor(255, 255, 255, 255))
# 数据清除
def pbtClear_Callback(self):
self.clearDataArea()
# 识别
def pbtPredict_Callback(self):
__img, img_array = [], [
] # 将图像统一从qimage->pil image -> np.array [1, 1, 28, 28]
# 获取qimage格式图像
if self.mode == MODE_MNIST:
__img = self.lbDataArea.pixmap() # label内若无图像返回None
if __img == None: # 无图像则用纯黑代替
# __img = QImage(224, 224, QImage.Format_Grayscale8)
__img = ImageQt.ImageQt(
Image.fromarray(np.uint8(np.zeros([224, 224]))))
else:
__img = __img.toImage()
elif self.mode == MODE_WRITE:
__img = self.paintBoard.getContentAsQImage()
# 转换成pil image类型处理
pil_img = ImageQt.fromqimage(__img)
pil_img = pil_img.resize((28, 28), Image.ANTIALIAS)
img_array = np.array(pil_img.convert('L')).reshape(1, 1, 28, 28)
img = normalize(img_array)
img = paddle.Tensor(img)
__result = network(img)
argmax__result = paddle.argmax(__result).numpy()
self.result[0] = argmax__result[0] # 置信度
m = F.sigmoid(__result).numpy()
self.result[1] = m[0][self.result[0]]
self.lbResult.setText("%d" % (self.result[0]))
self.lbCofidence.setText("%.8f" % (self.result[1]))
# 随机抽取
def pbtGetMnist_Callback(self):
self.clearDataArea()
# 随机抽取一张测试集图片，放大后显示
path_origin = "data/test"
files = list(
filter(lambda x: x.endswith('.jpg'), os.listdir(path_origin)))
random.shuffle(files)
number = random.randint(0, len(files))
img = Image.open(os.path.join(path_origin, files[number]))
pil_img = Image.fromarray(np.uint8(img))
pil_img = pil_img.resize((224, 224)) # 图像放大显示
# 将pil图像转换成qimage类型
qimage = ImageQt.ImageQt(pil_img)
# 将qimage类型图像显示在label
pix = QPixmap.fromImage(qimage)
self.lbDataArea.setPixmap(pix)