Bicubic interpolation

<p>In <a href="/facts/Mathematics/pxTouaz4">mathematics</a>, bicubic interpolation is an extension of <a href="/facts/Cubic_interpolation/S29nwRUF">cubic spline interpolation</a> (a method of applying cubic interpolation to a data set) for <a href="/facts/Interpolation/8PyetX4f">interpolating</a> data points on a <a href="/facts/Two-dimensional/f4b397W8">two-dimensional</a> <a href="/facts/Regular_grid/8vDdL4UB">regular grid</a>. The interpolated surface (meaning the kernel shape, not the image) is <a href="/facts/Smooth_function/mffL4ch2">smoother</a> than corresponding surfaces obtained by <a href="/facts/Bilinear_interpolation/yoke6SzF">bilinear interpolation</a> or <a href="/facts/Nearest-neighbor_interpolation/AGHTPfEu">nearest-neighbor interpolation</a>. Bicubic interpolation can be accomplished using either <a href="/facts/Lagrange_polynomial/MnpMmhtC">Lagrange polynomials</a>, <a href="/facts/Cubic_spline/S29nwRUF">cubic splines</a>, or cubic convolution algorithm.
</p><p>In <a href="/facts/Image_processing/zenalNCf">image processing</a>, bicubic interpolation is often chosen over bilinear or nearest-neighbor interpolation in <a href="/facts/Resampling_(bitmap)/qNoeujzm">image resampling</a>, when speed is not an issue. In contrast to bilinear interpolation, which only takes 4 <a href="/facts/Pixel/UzwBvXBV">pixels</a> (2×2) into account, bicubic interpolation considers 16 pixels (4×4). Images resampled with bicubic interpolation can have different interpolation <a href="/facts/Spatial_anti-aliasing/od0SgvLy">artifacts</a>, depending on the b and c values chosen.
</p>

Bicubic interpolation open-in-new

Bicubic interpolation