Function convolve3d 

pub fn convolve3d(
    volume: &ArrayBase<OwnedRepr<f64>, Dim<[usize; 3]>>,
    kernel: &ArrayBase<OwnedRepr<f64>, Dim<[usize; 3]>>,
    padding: Padding3D,
) -> Result<ArrayBase<OwnedRepr<f64>, Dim<[usize; 3]>>, NdimageError>

3D discrete convolution.

Convolves volume with kernel using the chosen padding strategy. The kernel is flipped (mirrored) along all three axes before sliding, which is the definition of mathematical convolution.

Arguments

• volume - Input 3D array.
• kernel - Convolution kernel (any odd-or-even size is accepted).
• padding - Border handling strategy.

Returns

An Array3<f64> with the same shape as volume.

Errors

Returns NdimageError::InvalidInput if the kernel has zero extent on any axis.

Example

use scirs2_ndimage::convolution3d::{convolve3d, Padding3D};
use scirs2_core::ndarray::Array3;

let vol = Array3::<f64>::ones((8, 8, 8));
let kernel = Array3::<f64>::ones((3, 3, 3));
let out = convolve3d(&vol, &kernel, Padding3D::Zero).unwrap();
assert_eq!(out.shape(), [8, 8, 8]);