Convolution

2D and 3D Convolution with linear kernels More...

Detailed Description

2D and 3D Convolution with linear kernels

Collaboration diagram for Convolution:

Functions
template<class T >
RES_T	gaussianFilter (Image< T > &imIn, int radius, Image< T > &imOut)
	gaussianFilter() - 3D Gaussian filter
template<class T >
RES_T	horizConvolve (const Image< T > &imIn, const vector< double > &kernel, Image< T > &imOut)
	horizConvolve() - 2D Horizontal convolution
template<class T >
RES_T	vertConvolve (const Image< T > &imIn, const vector< double > &kernel, Image< T > &imOut)
	vertConvolve() - 2D Vertical convolution
template<class T >
RES_T	convolve (const Image< T > &imIn, const vector< double > &kernel, Image< T > &imOut)
	convolve() - 2D Convolution

Function Documentation

gaussianFilter()

template<class T >

RES_T gaussianFilter	(	Image< T > &	imIn,
		int	radius,
		Image< T > &	imOut
	)

gaussianFilter() - 3D Gaussian filter

Gaussian Filter : convolution de l'input image against a Gaussian Kernel of size 2 * radius + 1

The kernel is defined, in each direction, by :

Parameters

[in]	imIn	: input image
[in]	radius	: radius of the gaussian kernel
[out]	imOut	: output image

Example

 
import smilPython as sp
 
imIn = sp.Image("https://smil.cmm.minesparis.psl.eu/images/lena.png")
imOut = sp.Image(imIn)
imIn.show()
imOut.show()
 
sp.gaussianFilter(imIn, 3, imOut)
 

horizConvolve()

template<class T >

RES_T horizConvolve	(	const Image< T > &	imIn,
		const vector< double > &	kernel,
		Image< T > &	imOut
	)

horizConvolve() - 2D Horizontal convolution

2D horizontal convolution using a 1D kernel

Parameters

[in]	imIn	: input image
[in]	kernel	: an 1D kernel (as a vector)
[out]	imOut	: output image

Example:

import smilPython as sp
 
im1 = sp.Image("https://smil.cmm.minesparis.psl.eu/images/lena.png")
im2 = sp.Image(im1)
kern = [ 0.0545, 0.2442, 0.4026, 0.2442, 0.0545 ]
sp.horizConvolve(im1, kern, im2)
 
im1.show()
im2.show()

vertConvolve()

template<class T >

RES_T vertConvolve	(	const Image< T > &	imIn,
		const vector< double > &	kernel,
		Image< T > &	imOut
	)

vertConvolve() - 2D Vertical convolution

2D vertical convolution using a 1D kernel

Parameters

[in]	imIn	: input image
[in]	kernel	: an 1D kernel (vector)
[out]	imOut	: output image

See also

horizConvolve()

convolve()

template<class T >

RES_T convolve	(	const Image< T > &	imIn,
		const vector< double > &	kernel,
		Image< T > &	imOut
	)

convolve() - 2D Convolution

2D convolution by & 1D kernel. Vertical convolution followed by an horizontal convolution using the same 1D kernel.

Parameters

[in]	imIn	: input image
[in]	kernel	: the 1D kernel (vector)
[out]	imOut	: output image

See also

horizConvolve()