Table of Contents

Vision

Color Quantization

References

RGB cone

YUV polygon

HSV rectangle

Lab

Image segmentation

Floodfill

References

Watershed

References

Filters

References

Anti-noise

(smoothing)

References

Median Filter

Vector Median Filter

Kuwahara filter

Peer Group Filtering

References

Anisotropic Filtering

Gradient

Prewitt

Objective

Basic kernels for 1st order gradient.

Quick Def

Horizontal kernel (transpose for vertical) :

-101
-101
-101

Roberts

Objective
Quick Def

First kernel :

10
0-1

Second kernel :

01
-10

Intensity : I = sqrt(I1^2 + I2^2)

Direction : theta = arctan(I2 / I1) + pi/4

Sobel

Objective

Most popular 1st order kernels

Quick Def

Horizontal kernel (transpose for vertical) :

-101
-202
-101

Intensity : I = sqrt(Iv^2 + Ih^2)

Direction : theta = arctan(Iv / Ih)

Full Def

The filter can be seen as the convolution of two filters.

Derivative :

-1
0
1

Smoothing :

121

Scharr

Objective

May give more accurate results than Sobel

Quick Def

Horizontal kernel (transpose for vertical) :

-303
-10010
-303

Laplace

Objective

2nd order kernel

Quick Def

Kernel :

010
1-41
010

Morphological

References

Dilation

Dilatation

Erosion

Erosion

Opening

Ouverture

Closing

Fermeture

Edge detection

Canny detector

Objective

Best theoritical edge detector.

Quick Def

Noise filtering, gradient with horizontal vertical and diagonal kernels, thresholding with hysteresis taking into account the direction of the gradient.

Canny-Deriche