Question

我一直在研究分段问题很多天，在最终找到如何正确读取数据集之后，我遇到了这个问题： “ ValueError：检查目标时出错：预期activation_1（Softmax）有3个维度，但得到的数组有形状（32,416,608,3）” 我使用了功能API，因为我采用了以下的FCNN架构： https://github.com/divamgupta/image-segmentation-keras/blob/master/Models/FCN32.py

稍微修改并根据我的任务进行调整（IMAGE_ORDERING =“channels_last”（TensorFlow后端））。     谁能帮帮我吗？     提前大力谢谢。     下面的体系结构适用于FCNN，我尝试将其用于分段。     这是架构（在调用model.summary（）之后）     [1]：https://i.stack.imgur.com/2Ou5z.png
    [2]：https://i.stack.imgur.com/zOFAz.png
    [3]：具体错误是：https://i.stack.imgur.com/DVo2k.png
    [4]：“导入数据集”功能：https://i.stack.imgur.com/UY2FE.png
    [5]：“Fit_Generator方法调用”：https://i.stack.imgur.com/VskLj.png

IMAGE_ORDERING = 'channels_last'
def getFCN32(nb_classes,input_height=416, input_width=608):

    img_input = Input(shape=(input_height,input_width,3))

    #Block 1
    x = Convolution2D(64, (3, 3), activation='relu', padding='same', name='block1_conv1', data_format=IMAGE_ORDERING)(img_input) 
    x = BatchNormalization()(x)
    x = Convolution2D(64, (3, 3), activation='relu', padding='same', name='block1_conv2', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool', data_format=IMAGE_ORDERING)(x)
    f1 = x
    # Block 2
    x = Convolution2D(128, (3, 3), activation='relu', padding='same', name='block2_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(128, (3, 3), activation='relu', padding='same', name='block2_conv2', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool', data_format=IMAGE_ORDERING )(x)
    f2 = x

    # Block 3
    x = Convolution2D(256, (3, 3), activation='relu', padding='same', name='block3_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(256, (3, 3), activation='relu', padding='same', name='block3_conv2', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(256, (3, 3), activation='relu', padding='same', name='block3_conv3', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool', data_format=IMAGE_ORDERING )(x)
    f3 = x

    # Block 4
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block4_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block4_conv2',data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block4_conv3',data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool', data_format=IMAGE_ORDERING)(x)
    f4 = x

    # Block 5
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block5_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block5_conv2',data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block5_conv3', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool', data_format=IMAGE_ORDERING)(x)
    f5 = x

    x = (Convolution2D(4096,(7,7) , activation='relu' , padding='same', data_format=IMAGE_ORDERING))(x)
    x = Dropout(0.5)(x)
    x = (Convolution2D(4096,(1,1) , activation='relu' , padding='same',data_format=IMAGE_ORDERING))(x)
    x = Dropout(0.5)(x)

    #First parameter = number of classes+1 (de la background)
    x = (Convolution2D(20,(1,1) ,kernel_initializer='he_normal' ,data_format=IMAGE_ORDERING))(x)
    x = Convolution2DTranspose(20,kernel_size=(64,64), strides=(32,32),use_bias=False,data_format=IMAGE_ORDERING)(x)
    o_shape = Model(img_input,x).output_shape

    outputHeight = o_shape[1]
    print('Output Height is:', outputHeight)
    outputWidth = o_shape[2]
    print('Output Width is:', outputWidth)
    #https://keras.io/layers/core/#reshape
    x = (Reshape((20,outputHeight*outputWidth)))(x)
    #https://keras.io/layers/core/#permute
    x = (Permute((2, 1)))(x)
    print("Output shape before softmax is", o_shape)
    x = (Activation('softmax'))(x)
    print("Output shape after softmax is", o_shape)
    model = Model(inputs = img_input,outputs = x)
    model.outputWidth = outputWidth
    model.outputHeight = outputHeight
    model.compile(optimizer = 'adam', loss = 'categorical_crossentropy', metrics =['accuracy'])

    return model

Answer 1

FCNN架构示例中的原始代码使用输入维度(416, 608)。而在您的代码中，输入维度为(192, 192)（忽略通道维度）。现在，如果你仔细注意，这个特定的层

x = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool', data_format=IMAGE_ORDERING)(x)

生成维度(6, 6)的输出（您可以在model.summary()中进行验证。）

下一个 convoltuion图层

o = (Convolution2D(4096,(7,7) , activation='relu' , padding='same', data_format=IMAGE_ORDERING))(o)

使用大小为(7, 7)的卷积滤镜，但您的输入已缩小到小于该值的卷积（即(6, 6)）。首先尝试修复它。

另外，如果您查看model.summary()输出，您会注意到它不包含 block5_pool 图层后定义的图层。其中有一个transposed convolution图层（基本上会对您的输入进行上采样）。您可能需要查看并尝试解决此问题。

注意：在我的所有维度中，我忽略了频道维度。

编辑以下详细解答

首先，这是我的keras.json文件。它使用 Tensorflow 后端，image_ordering设置为 channel_last 。

{
    "floatx": "float32",
    "epsilon": 1e-07,
    "backend": "tensorflow",
    "image_data_format": "channels_last"
}

接下来，我复制粘贴我的确切模型代码。请特别注意以下代码中的内联注释。

from keras.models import *
from keras.layers import *

IMAGE_ORDERING = 'channels_last' # In consistency with the json file

def getFCN32(nb_classes = 20, input_height = 416, input_width = 608):

    img_input = Input(shape=(input_height,input_width, 3)) # Expected input will have channel in the last dimension

    #Block 1
    x = Convolution2D(64, (3, 3), activation='relu', padding='same', name='block1_conv1', data_format=IMAGE_ORDERING)(img_input) 
    x = BatchNormalization()(x)
    x = Convolution2D(64, (3, 3), activation='relu', padding='same', name='block1_conv2', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block1_pool', data_format=IMAGE_ORDERING)(x)
    f1 = x
    # Block 2
    x = Convolution2D(128, (3, 3), activation='relu', padding='same', name='block2_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(128, (3, 3), activation='relu', padding='same', name='block2_conv2', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block2_pool', data_format=IMAGE_ORDERING )(x)
    f2 = x

    # Block 3
    x = Convolution2D(256, (3, 3), activation='relu', padding='same', name='block3_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(256, (3, 3), activation='relu', padding='same', name='block3_conv2', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(256, (3, 3), activation='relu', padding='same', name='block3_conv3', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block3_pool', data_format=IMAGE_ORDERING )(x)
    f3 = x

    # Block 4
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block4_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block4_conv2',data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block4_conv3',data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block4_pool', data_format=IMAGE_ORDERING)(x)
    f4 = x

    # Block 5
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block5_conv1', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block5_conv2',data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = Convolution2D(512, (3, 3), activation='relu', padding='same', name='block5_conv3', data_format=IMAGE_ORDERING)(x)
    x = BatchNormalization()(x)
    x = MaxPooling2D((2, 2), strides=(2, 2), name='block5_pool', data_format=IMAGE_ORDERING)(x)
    f5 = x

    x = (Convolution2D(4096,(7,7) , activation='relu' , padding='same', data_format=IMAGE_ORDERING))(x)
    x = Dropout(0.5)(x)
    x = (Convolution2D(4096,(1,1) , activation='relu' , padding='same',data_format=IMAGE_ORDERING))(x)
    x = Dropout(0.5)(x)

    x = (Convolution2D(20,(1,1) ,kernel_initializer='he_normal' ,data_format=IMAGE_ORDERING))(x)
    x = Convolution2DTranspose(20,kernel_size=(64,64), strides=(32,32),use_bias=False,data_format=IMAGE_ORDERING)(x)
    o_shape = Model(img_input, x).output_shape

    # NOTE: Since this is channel last dimension ordering, the height and width dimensions are along [1] and [2], not [2] and [3]
    outputHeight = o_shape[1]
    outputWidth = o_shape[2]

    x = (Reshape((outputHeight*outputWidth, 20)))(x) # Channel should be along the last dimenion of reshape
    # No need of permute layer anymore

    print("Output shape before softmax is", o_shape)
    x = (Activation('softmax'))(x)
    print("Output shape after softmax is", o_shape)
    model = Model(inputs = img_input,outputs = x)
    model.compile(optimizer = 'adam', loss = 'categorical_crossentropy', metrics =['accuracy'])

    return model

model = getFCN32(20)
print(model.summary())

接下来，我将提供我的model.summary()外观的摘要。如果你看看最后几层，它是这样的：

这意味着，conv2d_transpose图层在应用 softmax 之前会生成尺寸为(448, 640, 20)的输出，并且将其展平。所以输出的维度是(286720, 20)。同样，您的target_generator（在您的情况下为mask_generator）也应生成类似维度的目标。同样，您的input_generator也应该生成大小为[batch size, input_height,input_width, 3]的输入批次，如您函数的img_input中所述。

希望这可以帮助您找到问题的根源并找出合适的解决方案。请查看代码中的细微变化（以及内嵌评论）以及如何创建输入和目标批次。

Answer 2

您可能会在color_mode='grayscale'面罩电话中丢失flow_from_directory()。 RGB是color_mode的默认值。

flow_args = dict(
    batch_size=batch_size,
    target_size=target_size,
    class_mode=None,
    seed=seed)

image_generator = image_datagen.flow_from_directory(
    image_dir, subset='training', **flow_args)

mask_generator = mask_datagen.flow_from_directory(
    mask_dir, subset='training', color_mode='grayscale', **flow_args)

Keras（FIT_GENERATOR） - 检查目标时出错：预期activation_1具有3个维度，但得到了具有形状的数组（32,416,608,3）

2 个答案: