Question

我正在使用Keras（TF后端）中的U-Net架构对许多256x256图像执行逐像素的多类分类。我使用数据生成器对输出进行了一次热编码，使输出为256x256x32数组（我有32个不同的类，这些类以像素值表示，它们是256x256“蒙版”图像中从0到31的整数）。

但是，大多数地面真理数组都是空的-换句话说，到目前为止，最普通的类是0。当我训练U-Net时，它似乎过拟合了0类。损耗很低且精度很高，但这仅是因为〜99％的基本事实为0，所以U-Net只会输出一堆0，而我只真正关心其他31个类（例如，可以将其他类别归类为基本事实。

在计算损失函数时，是否有一种方法可以使某些类别比其他类别“权重”更多（如果这样，这种方法是否合适）？我不确定这是我的数据固有的问题还是方法上的问题。这是我的U-Net：

def unet(pretrained_weights = None,input_size = (256,256,1)):
inputs = keras.engine.input_layer.Input(input_size)
conv1 = Conv2D(64, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(inputs)
conv1 = Conv2D(64, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv1)
pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)
conv2 = Conv2D(128, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(pool1)
conv2 = Conv2D(128, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv2)
pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
conv3 = Conv2D(256, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(pool2)
conv3 = Conv2D(256, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv3)
pool3 = MaxPooling2D(pool_size=(2, 2))(conv3)
conv4 = Conv2D(512, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(pool3)
conv4 = Conv2D(512, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv4)
#drop4 = Dropout(0.5)(conv4)
drop4 = SpatialDropout2D(0.5)(conv4)
pool4 = MaxPooling2D(pool_size=(2, 2))(drop4)

conv5 = Conv2D(1024, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(pool4)
conv5 = Conv2D(1024, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv5)
#drop5 = Dropout(0.5)(conv5)
drop5 = SpatialDropout2D(0.5)(conv5)

up6 = Conv2D(512, 2, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(drop5))
merge6 = concatenate([drop4,up6], axis = 3)
conv6 = Conv2D(512, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(merge6)
conv6 = Conv2D(512, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv6)

up7 = Conv2D(256, 2, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(conv6))
merge7 = concatenate([conv3,up7], axis = 3)
conv7 = Conv2D(256, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(merge7)
conv7 = Conv2D(256, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv7)

up8 = Conv2D(128, 2, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(conv7))
merge8 = concatenate([conv2,up8], axis = 3)
conv8 = Conv2D(128, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(merge8)
conv8 = Conv2D(128, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv8)

up9 = Conv2D(64, 2, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(conv8))
merge9 = concatenate([conv1,up9], axis = 3)
conv9 = Conv2D(64, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(merge9)
conv9 = Conv2D(64, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv9)
conv9 = Conv2D(32, 3, activation = 'relu', padding = 'same', kernel_initializer = 'he_normal')(conv9)
conv10 = Conv2D(32, 1, activation = 'softmax')(conv9)
#conv10 = Flatten()(conv10)
#conv10 = Dense(65536, activation = 'softmax')(conv10)
flat10 = Reshape((65536,32))(conv10)
#conv10 = Conv1D(1, 1, activation='linear')(conv10)

model = Model(inputs = inputs, outputs = flat10)

opt = Adam(lr=1e-6,clipvalue=0.01)
model.compile(optimizer = opt, loss = 'categorical_crossentropy', metrics = ['categorical_accuracy'])
#model.compile(optimizer = Adam(lr = 1e-6), loss = 'binary_crossentropy', metrics = ['accuracy'])
#model.compile(optimizer = Adam(lr = 1e-4),

#model.summary()

if(pretrained_weights):

    model.load_weights(pretrained_weights)

return model

请让我知道是否需要更多信息来诊断问题。

Answer 1

处理不平衡类的常见解决方案是对某些类进行加权。在培训期间，通过使用可选的class_weight参数，在Keras中这很容易。

model.fit(x, y, class_weight=class_weight)

您可以自己在字典中定义班级权重：

class_weight = {0: 1, 1: 100}

或者，您可以使用sklearn函数compute_class_weight从数据中自动生成权重。

class_weights = class_weight.compute_class_weight('balanced', np.unique(y), y)

使用深度学习防止在多类别分类中过度适合特定类别

1 个答案: