Struct Convolve 

pub struct Convolve { /* private fields */ }

Apply a convolution kernel.

Implementations

impl Convolve

pub const fn new(kernel: Kernel) -> Convolve

Convolve with kernel.

pub const fn kernel(&self) -> &Kernel

The kernel this op applies.

Trait Implementations

impl Clone for Convolve

fn clone(&self) -> Convolve

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Convolve

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

Formats the value using the given formatter.

impl Op for Convolve

fn rescaled(&self) -> Option<Arc<dyn Op>>

Never rescaled. The kernel is measured in pixels, so the same 3x3 blur over a source decoded eight times smaller is eight times the blur relative to image content. The output would be the right size and the wrong picture, which is exactly the failure the conservative default exists to prevent.

fn name(&self) -> &'static str

A short, stable name for this op, used in diagnostics.

fn output_descriptor( &self, inputs: &[ImageDescriptor], ) -> Result<ImageDescriptor, PixelsError>

Compute the output shape from the input shapes.

fn input_regions( &self, output: Region, inputs: &[ImageDescriptor], ) -> Result<Vec<Region>, PixelsError>

The input regions needed to produce output.

fn access_pattern(&self) -> AccessPattern

How this op reads its inputs; drives tile negotiation (ADR-0003).

fn compute( &self, inputs: &[Tile<'_>], output: &mut TileMut<'_>, ) -> Result<(), PixelsError>

Fill output from inputs — the kernel entry point.

fn arity(&self) -> usize

The number of inputs this op consumes.

impl PartialEq for Convolve

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for Convolve