convolution and correlation

2D convolution
 A filter is a technique with which certain
frequency components can be chosen or
rejected
 Convolution based filters use spatial mask
 Spatial mask are also known as kernels,
templates and windows
 The spatial mask is convolved with given image
to achieve the required smoothing or sharpening
effect.
Spatial Filtering
A spatial filter consists of (a) a neighborhood, and (b) a
predefined operation

Linear spatial filtering of an image of size MxN with a filter

of size mxn is given by the expression

a b
g ( x, y ) =   w(s, t ) f ( x + s, y + t )
s =− a t =− b

m = 2a + 1; n = 2b + 1
Spatial Convolution
The convolution of a filter w( x, y ) of size m  n
with an image f ( x, y ), denoted as w( x, y ) f ( x, y )

a b
w( x, y ) f ( x, y ) =   w(s, t ) f ( x − s, y − t )
s =− a t =− b
Spatial Correlation
The correlation of a filter w( x, y ) of size m  n
with an image f ( x, y ), denoted as w( x, y ) f ( x, y )

a b
w( x, y ) f ( x, y ) =   w(s, t ) f ( x + s, y + t )
s =− a t =− b
2D convolution (kernel size: 3x3,
stride=1)

205 203
Shifting (stride =1)
 X={1 2 3 4} m=4
 Y={1 2 3 0} n=4
Zero padding