1

我有一个连体神经网络，我想对提取的图像特征应用线性变换，以使用 PCA 或自动编码器降低维度。如何在展平层之后实现它？

这是我的代码：

    input_a = Input(shape=(input_shape))
    input_b = Input(shape=(input_shape)) 

    # Convolutional Neural NetworK
    seq = Sequential()
    seq.add(Conv2D(32, (5,5), activation='relu',padding='same',input_shape=input_shape,
      kernel_initializer=initializers.RandomNormal(mean=0.0 ,stddev=0.1, seed=None),bias_initializer= initializers.Zeros()))
    seq.add(MaxPooling2D(pool_size=(2,2) ,strides=(2,2)))
    seq.add(Conv2D(64, (5,5), activation='relu',padding='same',
           kernel_initializer=initializers.RandomNormal(mean=0.0 ,stddev=0.1, seed=None),bias_initializer= initializers.Zeros()))
    seq.add(MaxPooling2D(pool_size=(2,2),strides=(2,2)))    
    seq.add(Flatten())
    

    processed_a = seq(input_a)
    processed_b = seq(input_b)
    #here i want to preform linear transformation

    L2_distance = Lambda(euclidean_distance, output_shape=eucl_dist_output_shape,name='L2')([processed_a, processed_b])    
    a = Lambda(function,output_shape=eucl_dist_output_shape,name='out1')(L2_distance)
   

    model = Model([input_a, input_b],a)
4

1 回答 1

0

要在连体神经网络的末端添加线性变换层，或者更好地说它的编码器，您可以执行以下两个步骤：

  1. 您需要构建一个自定义层。您可以使用以下一个并删除偏差项：https ://keras.io/guides/making_new_layers_and_models_via_subclassing/#the-layer-class-the-combination-of-state-weights-and-some-computation
  2. 您需要在seq.add(Flatten())之后添加该层
于 2021-01-25T11:04:36.810 回答