is an option that specifies the method to be used for resampling images or arrays.


  • In all of the interpolations, the window is normalized so that its values sum to 1.
  • With the setting Resampling->Automatic, the method of resampling is selected automatically.
  • Specific settings for Resampling are typically used to achieve different tradeoffs with respect to prefiltering of data, order of interpolation, and complexity of computation.
  • Nearest neighbor resamplings are fast, and except for "Nearest" do not introduce any new values:
  • "Nearest"nearest neighbor, use average for tie
    "NearestLeft"nearest neighbor, use left for tie
    "NearestRight"nearest neighbor, use right for tie
  • Spline interpolations are relatively fast, based on polynomial interpolation of order with continuous derivatives:
  • "Constant"piecewise constant interpolation
    "Linear"piecewise linear interpolation
    "Quadratic" spline interpolation of order 2
    "Cubic"spline interpolation of order 3
    "Quartic"spline interpolation of order 4
    "Quintic"spline interpolation of order 5
    {"Spline",n}spline interpolation of order up to
  • Gaussian and B-splines of higher orders are practically isotropic resamplings. They are fast approximations that blur the data rather than interpolations:
  • "Gaussian"Gaussian weighted resampling using and
    {"Gaussian",r,σ}Gaussian with a specific radius and sigma
    {"BSpline",n}B-spline approximation of order up to
  • Classic polynomial interpolations up to order :
  • "Dodgson"Dodgson polynomial interpolation
    {"Keys",a}Keys polynomial interpolation (default )
    "CatmullRom"CatmullRom (Meijering) cubic polynomial interpolation
    "German"German polynomial interpolation
    {"Hermite",n}^(th)-order Hermite interpolation
    {"Schaum",n}^(th)-order Schaum (Lagrange) polynomial interpolation
    {"Meijering",n}odd ^(th)-order Meijering polynomial interpolation
  • Optimal sampling of maximal order and minimal support (o-MOMS) gives the best resampling for a given order, and may give only continuous or even discontinuous filter kernel:
  • {"OMOMS",n}o-MOMS of order up to
  • Windowed sinc interpolations give ideal resamplings regularized by windows of the form or . The following possible window specifications can be given:
  • {"Bartlett",r}Bartlett (default )
    {"Blackman",r}Blackman (default )
    {"Connes",r,α}Connes (default , )
    {"Cosine",r,α}cosine (default , )
    {"Hamming",r}Hamming (default )
    {"Hann",r,α}Hann (default , )
    {"Kaiser",r,α}Kaiser (default , )
    {"Lanczos",r}Lanczos (default )
    {"Parzen",r}Parzen (default )
    {"Welch",r,α}Welch (default , )


Basic Examples  (2)

Downsample an image using Gaussian interpolation:

Upsample an image using a higher-order interpolation:

Wolfram Research (2010), Resampling, Wolfram Language function, (updated 2014).


Wolfram Research (2010), Resampling, Wolfram Language function, (updated 2014).


Wolfram Language. 2010. "Resampling." Wolfram Language & System Documentation Center. Wolfram Research. Last Modified 2014.


Wolfram Language. (2010). Resampling. Wolfram Language & System Documentation Center. Retrieved from


@misc{reference.wolfram_2024_resampling, author="Wolfram Research", title="{Resampling}", year="2014", howpublished="\url{}", note=[Accessed: 24-July-2024 ]}


@online{reference.wolfram_2024_resampling, organization={Wolfram Research}, title={Resampling}, year={2014}, url={}, note=[Accessed: 24-July-2024 ]}