Struct ImageArray

pub struct ImageArray<T: Gradient + Average + VectorLen + Norm> { /* private fields */ }

An array representing an image, used with the solvers. The From trait is implemented for the types GrayImage and RgbImage in the image crate.

Implementations

impl ImageArray<Array3<f64>>

pub fn into_luma(&self) -> GrayImage

Assumes Array3 axis 2 is colors, will flatten axis 2 if bigger than 1.

pub fn into_rgb(&self) -> RgbImage

Assumes Array3 axis 2 is colors, will use 3 first elements of axis 2 if bigger than 3, or cycle through the elements if smaller than 3.

impl ImageArray<Array3<f64>>

pub fn denoise( &self, lambda: f64, tau: f64, sigma: f64, gamma: f64, max_iter: u32, convergence_threshold: f64, ) -> Result<Self, ShapeError>

Image denoising algorithm for 2 dimentional shapes with 1 dimention of information (pixels) as an arbitrarily sized vector. Assumes axes 0 and 1 and the x and y coordinates of the image, and axis 2 is the pixel vector coordinate of the image.

inputs

lambda is the target value of the dual objective function, i.e. how close you want the output to be to the input: approaching 0, the output should be completely smooth (flat), approaching “infinifty”, the output should be the same as the original input.

tau and sigma affect how fast the algorithm converges, according to Chambolle, A. and Pock, T. (2011) these should be chosen such that tau * lambda * L2 norm^2 <= 1 where L2 norm^2 <= 8.

gamma updates the algorithm’s internal variables, for the accelerated algorithm of Chambolle, A. and Pock, T. (2011) the chosen value is 0.35 * lambda.

max_iter and convergence_threshold bound the runtime of the algorithm, i.e. it runs until convergence_threshold < norm(current - previous) / norm(previous) or max_iter is hit.

Trait Implementations

impl<T: Clone + Gradient + Average + VectorLen + Norm> Clone for ImageArray<T>

fn clone(&self) -> ImageArray<T>

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl<T: Debug + Gradient + Average + VectorLen + Norm> Debug for ImageArray<T>

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

Formats the value using the given formatter.

impl<T: Gradient + Average + VectorLen + Norm> Deref for ImageArray<T>

type Target = T

The resulting type after dereferencing.

fn deref(&self) -> &Self::Target

Dereferences the value.

impl<T: Copy + Into<f64>, D: Dimension + RemoveAxis> From<&ArrayBase<OwnedRepr<T>, D>> for ImageArray<Array<f64, D>>

fn from(value: &Array<T, D>) -> Self

Converts to this type from the input type.

impl From<&ImageBuffer<Luma<u8>, Vec<u8>>> for ImageArray<Array3<f64>>

fn from(value: &GrayImage) -> Self

Converts to this type from the input type.

impl From<&ImageBuffer<Rgb<u8>, Vec<u8>>> for ImageArray<Array3<f64>>

fn from(value: &RgbImage) -> Self

Converts to this type from the input type.