maxUnpooling2dLayer

Max unpooling layer

Description

A 2-D max unpooling layer unpools the output of a 2-D max pooling layer.

Creation

Syntax

layer = maxUnpooling2dLayer

layer = maxUnpooling2dLayer('Name',name)

Description

layer = maxUnpooling2dLayer creates a max unpooling layer.

layer = maxUnpooling2dLayer('Name',name) sets the Name property. To create a network containing a max unpooling layer you must specify a layer name.

Properties

Name — Layer name

'' (default) | character vector | string scalar

Layer name, specified as a character vector or a string scalar. For Layer array input, the trainnet and dlnetwork functions automatically assign names to unnamed layers.

The MaxUnpooling2DLayer object stores this property as a character vector.

Data Types: char | string

NumInputs — Number of inputs

3 (default)

Number of inputs of the layer.

There are three inputs to this layer:

• 'in' — Input feature map to unpool.

• 'indices' — Indices of the maximum value in each pooled region. This is output by the max pooling layer.

• 'size' — Output size of unpooled feature map. This is output by the max pooling layer.

Use the input names when connecting or disconnecting the max unpooling layer to other layers using connectLayers or disconnectLayers, respectively.

Data Types: double

InputNames — Input names

{'in','indices','size'} (default)

Input names of the layer.

There are three inputs to this layer:

• 'in' — Input feature map to unpool.

• 'indices' — Indices of the maximum value in each pooled region. This is output by the max pooling layer.

• 'size' — Output size of unpooled feature map. This is output by the max pooling layer.

Use the input names when connecting or disconnecting the max unpooling layer to other layers using connectLayers or disconnectLayers, respectively.

Data Types: cell

NumOutputs — Number of outputs

Read-only: 1 (default)

This property is read-only.

Number of outputs from the layer, stored as 1. This layer has a single output only.

Data Types: double

OutputNames — Output names

Read-only: {'out'} (default)

This property is read-only.

Output names, stored as {'out'}. This layer has a single output only.

Data Types: cell

Examples

Create Max Unpooling Layer

Create a max unpooling layer that unpools the output of a max pooling layer.

layer = maxUnpooling2dLayer

layer = 
  MaxUnpooling2DLayer with properties:

          Name: ''
     NumInputs: 3
    InputNames: {'in'  'indices'  'size'}

Unpool Max Pooling Layer

Create a max pooling layer, and set the 'HasUnpoolingOutputs' property as true. This property gives the max pooling layer two additional outputs,'indices' and 'size', which enables unpooling the layer. Also create a max unpooling layer.

layers = [
    maxPooling2dLayer(2,'Stride',2,'Name','mpool','HasUnpoolingOutputs',true)
    maxUnpooling2dLayer('Name','unpool')]

layers = 
  2×1 Layer array with layers:

     1   'mpool'    2-D Max Pooling     2×2 max pooling with stride [2  2] and padding [0  0  0  0]
     2   'unpool'   2-D Max Unpooling   2-D Max Unpooling

Add the layers to a dlnetwork object.

net = dlnetwork;
net = addLayers(net,layers)

net = 
  dlnetwork with properties:

         Layers: [2×1 nnet.cnn.layer.Layer]
    Connections: [1×2 table]
     Learnables: [0×3 table]
          State: [0×3 table]
     InputNames: {'mpool'  'unpool/indices'  'unpool/size'}
    OutputNames: {'mpool/indices'  'mpool/size'  'unpool'}
    Initialized: 0

  View summary with summary.

Unpool the output of the max pooling layer, by connecting the max pooling layer outputs to the max unpooling layer inputs.

net = connectLayers(net,'mpool/indices','unpool/indices');
net = connectLayers(net,'mpool/size','unpool/size');

Extended Capabilities

C/C++ Code Generation

Generate C and C++ code using MATLAB® Coder™.

Usage notes and limitations:

If equal max values exists along the off-diagonal in a kernel window, implementation differences for the maxPooling2dLayer might cause minor numerical mismatch between MATLAB^® and the generated code. This issue also causes mismatch in the indices of the maximum value in each pooled region.

For example, consider the following input to maxPooling2dLayer. The output from MATLAB is shown.

For the same input, the output from the generated code is shown.

GPU Code Generation

Generate CUDA® code for NVIDIA® GPUs using GPU Coder™.

Usage notes and limitations:

If equal max values exists along the off-diagonal in a kernel window, implementation differences for the maxPooling2dLayer might cause minor numerical mismatch between MATLAB and the generated code. This issue also causes mismatch in the indices of the maximum value in each pooled region.

For example, consider the following input to maxPooling2dLayer. The output from MATLAB is shown.

For the same input, the output from the generated code is shown.

Version History

Introduced in R2017b