Enum Interpolation 

#[non_exhaustive]
pub enum Interpolation {
    Nearest,
    Bilinear,
}

How convert() resamples when it resizes.

Every backend honours it: the CPU backend through its resizer, the OpenGL backend through its texture filtering and NV shader choice, and G2D, which has a fixed hardware filter, by declining resizing converts under Interpolation::Nearest so another backend runs them. Converts that do not resize read each source pixel exactly in either mode.

Override at runtime with EDGEFIRST_INTERPOLATION=nearest|bilinear, which takes precedence over ImageProcessorConfig::interpolation and ImageProcessor::set_interpolation.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

Nearest

Nearest-neighbour: each output pixel copies one source pixel. The fastest mode, at the cost of aliasing that lowers detection accuracy on resized inputs.

Bilinear

Bilinear with half-pixel centres (default): two taps per axis at every scale on both the CPU and GL backends, matching OpenCV INTER_LINEAR, the preprocessing most detectors and segmenters are trained with. It does not widen its kernel when downscaling, so it smooths less than PIL’s bilinear below 1×.

Trait Implementations

impl Clone for Interpolation

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for Interpolation

impl Debug for Interpolation

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for Interpolation

fn default() -> Self

Returns the “default value” for a type.

impl Eq for Interpolation

impl PartialEq for Interpolation

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for Interpolation