gives a resized version of image that is width pixels wide.


gives a resized version of image with a maximum pixel width or height given by size.


gives a resized version of image that has exactly the specified pixel width and height.


gives a video in which every frame is resized.


gives a resized version of a 3D image with the specified dimensions.

Details and Options

  • Image resizing changes the dimensions of an image. This is needed when enlarging an image for viewing purposes or creating thumbnails when previewing a collection of images. It is also used to conform collections of images into a common size when batch processing, training, etc.
  • ImageResize[image,w] and ImageResize[image,{s}] preserve the original aspect ratio of image.
  • Specifications for width, height and depth can be any of the following:
  • dd pixels
    {d}maximum d pixels
    Scaled[s]a multiple s of the original size
    Allpreserve the original dimension
    Automaticdetermine value from aspect ratio
    Tiny,Small,Medium,Largepredefined absolute sizes
  • ImageResize[image,w] is equivalent to ImageResize[image,{w,Automatic}].
  • ImageResize[image,{size}] is equivalent to ImageResize[image,{{size},{size}}].
  • To obtain an image with an aspect ratio r, use ImageResize[image,w{1,r}].
  • The following options can be given:
  • Padding "Reversed"padding method
    Resampling Automaticresampling method


open allclose all

Basic Examples  (1)

Resize an image to be 100 pixels wide:

Scope  (5)

Data  (3)

Resize a grayscale image:

Resize frames of a video:

Resize a 3D image:

Parameters  (2)

Specify the width of the resulting image:

Resize by a factor of :

Use Automatic to explicitly preserve the aspect ratio of the image:

Resize an image to have the larger dimension no greater than 200 pixels:

Specify both width and height dimensions:

Resize in one dimension only:

Specify a named size:

Resize a 3D image:

Resize given the new width:

Resize given the new height:

Options  (4)

Padding  (3)

By default, "Reversed" padding is used:

Use constant padding:

When resampling method is set to "Nearest", all padding methods return the same result:

With constant padding, the padding value is converted to the color space of the target image:

Resampling  (1)

By default, "Lanczos" resampling method is used when resizing an image:

Use method "Nearest":

Applications  (3)

Create an image thumbnail:

Conform and classify a list of day and night images.

Here is an example set containing labeled images of day and night:

Calculate the average dimension of the example images:

Create the classifier:

Try this classifier on a few test images:

Get a list of images from the web and then conform them:

Conform dimensions of the list of images:

Conform while imposing an aspect ratio that averages over the images:

Properties & Relations  (4)

Resampling method "Nearest" is used when increasing the size of an image with only a few pixels:

Use "Linear" resampling:

Use ImageResize instead of Size argument in ImageTransformation:

Thumbnail is equivalent to ImageResize using size specification Small and "Fixed" padding:

Importing an image with given ImageSize is equivalent to resizing after Import:

Possible Issues  (2)

Downsizing with no interpolation will give poor results for thin objects:

Use first-order interpolation:

Reducing the size of an image with high-frequency content may result in a moire pattern:

Blur the image to reduce the moire effect:

Interactive Examples  (1)

Create a pan and zoom application while fixing the displayed image size:

Wolfram Research (2008), ImageResize, Wolfram Language function, (updated 2021).


Wolfram Research (2008), ImageResize, Wolfram Language function, (updated 2021).


Wolfram Language. 2008. "ImageResize." Wolfram Language & System Documentation Center. Wolfram Research. Last Modified 2021.


Wolfram Language. (2008). ImageResize. Wolfram Language & System Documentation Center. Retrieved from


@misc{reference.wolfram_2024_imageresize, author="Wolfram Research", title="{ImageResize}", year="2021", howpublished="\url{}", note=[Accessed: 15-June-2024 ]}


@online{reference.wolfram_2024_imageresize, organization={Wolfram Research}, title={ImageResize}, year={2021}, url={}, note=[Accessed: 15-June-2024 ]}