U-Net 2D - Decoding Layer

This component can be used for the decoding layers of a U-Net for 2D data (the corresponding encoding layers can be created using the U-Net 2D - Encoding Layer component). The following layers are used:

* Transposed Convolution 2D Layer
* Concatenate Layer
* Convolution 2D Layer
* Dropout Layer
* Convolution 2D Layer

The following inputs should be passed from the U-Net 2D - Encoding Layer Component as flow variables:
* Filter size for the Convolution layers
* Activation Function for the Colnvolution layers
* Kernel Size for the Convolution layers
* Dropout Rate for the Dropout layer
* Random seed for the Dropout layer and for the Kernel Initializer of the Convolution layers
* Kernel size (and Strides) for the Transposed Convolution layer (correspondes to the Pool Size from the Max Pooling layer)
* Strides for the Convolution layers
* Padding for the Convolution layers
* Kernel Initializer for the Convolution layers

The required extensions:
- KNIME Deep Learning - Keras Integration

Input ports

1. Type: Keras Deep Learning Network
   Port 1
   Convoluted Keras Layers
2. Type: Keras Deep Learning Network
   Port 2
   Convoluted and Max Pooling Keras Layers

Output ports

1. Type: Keras Deep Learning Network
   Port 1
   Encoded and Decoded U-Net