Laplacian of Gaussian Filtering

size

kernel display

matrix

continuous 3D

image

baby

letters

Kernel		Result

This Demonstration shows the filtering of an image using a 2D convolution with the Laplacian of a Gaussian kernel.

This operation is useful for detecting features or edges in images.

The kernel is sampled and normalized using the Laplacian of the Gaussian function

-

1

π

4

σ

1-

2

x

+

2

y

2

σ

-

2

x

+

2

y

2

σ



.

The standard deviation

σ

is chosen to be one fifth of the width of the kernel.

Yu-Sung Chang