What is Nearest Neighbor Interpolation?

Nearest-neighbor resampling maps each destination coordinate to the closest discrete source sample and copies that value unchanged. It performs no weighted average, so the output contains only values already present in the input. This makes the method computationally simple and semantically safe for categorical rasters, while diagonal edges and photographs develop stair steps or block-shaped regions. It is used in image transforms, texture sampling, and preprocessing when preserving exact sample identities matters more than smoothness.

Choose nearest neighbor scaling for pixel art, masks, or label maps where intermediate values would be invalid. For photographs, its hard transitions usually look rougher than blended interpolation methods.