bicubische

Bicubische interpolation, also known as bicubic interpolation, is a resampling method used to resize digital images and other raster data. It estimates new pixel values by fitting cubic polynomials to the four-by-four neighborhood around the target position, resulting in smoother images than bilinear or nearest-neighbor methods.

In practice, when scaling an image, each target pixel corresponds to a location in the source image. The algorithm gathers a 4x4 grid of surrounding source pixels and applies a cubic convolution kernel along both the horizontal and vertical directions to compute the interpolated value. Common kernel families include Keys, Mitchell-Netravali, and Catmull-Rom, with a parameter that controls sharpness influence.

Advantages of bicubic interpolation include smoother gradations and better preservation of details compared with simpler methods, producing more natural-looking results and less blockiness. It is widely adopted in image editing software and graphics libraries due to its balance between quality and computational efficiency.

Disadvantages include higher computational cost than nearest-neighbor or bilinear methods and the potential to blur very fine edges or introduce mild ringing artifacts near strong contrast transitions in some images or implementations.

Applications and implementation: Bicubic interpolation is standard in many image processing toolkits and software packages. In programming libraries, it is often exposed as a dedicated resizing method (for example, labeled as bicubic or as an INTER_CUBIC option). It is also used in video scaling and texture resampling in graphics pipelines.

Limitations: Like all interpolation methods, bicubic cannot recreate details that were not present in the original image. The results depend on the input image characteristics and the chosen kernel variant.