Input Image

At first we need an image . In the given example we see a image as a matrix . Basically, images are represented as a 2D matrix .

image

Kernel

What is a kernel ?

image

We use this kernel for this example given below :

image

At first flip/rotate the kernel into 180 degrees .

image

Then apply zero padding into the image . Then multiply and sum all the things to find a value .

An example to find the value

image

Put it into the output image are given below . That’s we calculate our convoution operation into image processing .

Output image

Using the operation we found the output image .

image

References :

Definition in Bangla Convolution 2D Convolution FAQ