ResizeLayer

ResizeLayer[{d}]

represents a layer performing one-dimensional resizing of a two-dimensional array.

ResizeLayer[{d1,,dn}]

represents a layer performing n-dimensional resizing of a (n+1)-dimensional array.

Details and Options

  • ResizeLayer is typically used inside NetChain, NetGraph, etc. to resize one or several dimensions of input arrays by resampling.
  • By default or with the setting InterleavingFalse, ResizeLayer treats the first dimension as a channel dimension, which remains unchanged. With the setting InterleavingTrue, the channel dimension is taken to be the last dimension of the input and output arrays.
  • In ResizeLayer[{d1,,dn}], any of the di can be one of the following:
  • nscale dimension to be size n
    Scaled[r]scale dimension to be a ratio r of the original size
    Allleave the dimension unchanged
  • The following optional parameters can be included:
  • InterleavingFalsewhether to assume channels are interleaved
    ResamplingAutomaticresampling method
  • Possible explicit settings for the Resampling method include:
  • "Linear"piecewise linear interpolation
    "Nearest"nearest neighbor
  • When the Resampling method is "Nearest", only the special cases of ResizeLayer[{Scaled[m],Scaled[n],}], where m, n, ... are positive integers, are currently implemented.
  • ResizeLayer exposes the following ports:
  • "Input"an array
    "Output"an array of the same rank as the input
  • ResizeLayer[][input] explicitly computes the output from applying the layer.
  • ResizeLayer[][{input1,input2,}] explicitly computes outputs for each of the inputi.
  • When given a NumericArray as input, the output will be a NumericArray.
  • Options[ResizeLayer] gives the list of default options to construct the layer. Options[ResizeLayer[]] gives the list of default options to evaluate the layer on some data.
  • Information[ResizeLayer[]] gives a report about the layer.
  • Information[ResizeLayer[],prop] gives the value of the property prop of ResizeLayer[]. Possible properties are the same as for NetGraph.

Examples

open allclose all

Basic Examples  (2)

Create a ResizeLayer that resizes a three-dimensional array to be of size c×8×8:

Create a ResizeLayer that resizes a matrix to be of size c×2:

Apply the layer to an input matrix (where c is 1):

Scope  (2)

Create a ResizeLayer that resizes a three-dimensional array to be a specific size:

Apply the layer to an input:

Plot the result:

Construct a ResizeLayer that preserves the height and multiplies the width of an input image by 1.5:

Apply the layer to an image:

Thread the layer across a batch of images:

Options  (2)

Interleaving  (1)

Create a ResizeLayer with InterleavingFalse and one input channel:

Create a ResizeLayer with InterleavingTrue and one input channel:

Resampling  (1)

Upsample an array by a factor of 3 using linear interpolation:

Upsample an array with a factor of 3 using nearest neighbor interpolation:

Wolfram Research (2017), ResizeLayer, Wolfram Language function, https://reference.wolfram.com/language/ref/ResizeLayer.html (updated 2021).

Text

Wolfram Research (2017), ResizeLayer, Wolfram Language function, https://reference.wolfram.com/language/ref/ResizeLayer.html (updated 2021).

CMS

Wolfram Language. 2017. "ResizeLayer." Wolfram Language & System Documentation Center. Wolfram Research. Last Modified 2021. https://reference.wolfram.com/language/ref/ResizeLayer.html.

APA

Wolfram Language. (2017). ResizeLayer. Wolfram Language & System Documentation Center. Retrieved from https://reference.wolfram.com/language/ref/ResizeLayer.html

BibTeX

@misc{reference.wolfram_2022_resizelayer, author="Wolfram Research", title="{ResizeLayer}", year="2021", howpublished="\url{https://reference.wolfram.com/language/ref/ResizeLayer.html}", note=[Accessed: 15-August-2022 ]}

BibLaTeX

@online{reference.wolfram_2022_resizelayer, organization={Wolfram Research}, title={ResizeLayer}, year={2021}, url={https://reference.wolfram.com/language/ref/ResizeLayer.html}, note=[Accessed: 15-August-2022 ]}