07 - Morphology

Image Processing & Computer vision
Morphological Image Processing

2 Contents
In this topic we will consider
– What is morphology?
– Simple morphological operations
– Compound operations
– Morphological algorithms
– Applications
Image Processing and Computer Vision by Dr. Praveen Kumar @ CSE, VNIT
3 What Is Morphology?
• Morphological image processing (or morphology)
describes a range techniques for extracting image
components that are useful in the representation &
description of region shape such as boundaries,
skeletons etc.
• Some of the operations are frequently applied as
post-processing to remove imperfections introduced
during segmentation, and so typically operate on
binary images.
4 Quick Example 1
Problem here
How do we fill “missing pixels”?
5 Quick Example 2
om Gonzalez & Woods, Digital Image Processing (2002)
Image after segmentation Image after segmentation and

morphological processing
6 Preliminaries
Language of mathematical morphology is set theory
7 Preliminaries
Logic operations involving binary images
8 Preliminaries
Reflection and Translation:
ˆ = {w | w  −b,
B for b  B}
( A) z = {c | c  a + z , for a  A}
9 Structuring Elements
• Structuring elements are small sets/sub-images
used to probe an image under study
• For each SE, define its origin
• Shape and size must be adapted to geometric
properties for the objects
– For simplicity we will use rectangular structuring elements
with their origin at the middle pixel
10 Structuring Elements, Hits & Fits
B Structuring Element
Fit: All on pixels in the
structuring element cover
A on pixels in the image
C
Hit: Any on pixel in the
structuring element covers
an on pixel in the image
11 Fitting & Hitting
0 0 0 0 0 0 0 0 0 0 0 0
1 1 1
0 0 0 1 1 0 0 0 0 0 0 0
1 1 1
0 0 1 B1 1 1 1 0 C
0 0 0 0
1 1 1
0 1 1 1 1 1 1 1 0 0 0 0
Structuring
0 1 1 1 1 1 1 1 0 0 0 0 Element 1
0 0 1 1 1 1 1 1 0 0 0 0 0 1 0
0 0 1 1 1 1 1 1 1 0 0 0 1 1 1
0 0 1 1 1 1 1 A1 1 1 1 0 0 1 0
0 0 0 0 0 1 1 1 1 1 1 0 Structuring
0 0 0 0 0 0 0 0 0 0 0 0 Element 2
12 Fundamental Operations
• Fundamentally morphological image processing is
very much like spatial filtering
• The structuring element is moved across every
pixel in the original image to give a pixel in a new
processed image
• The value of this new pixel depends on the
operation performed
• There are two basic morphological operations:
erosion and dilation
13 Erosion
Erosion of image A by structuring element B represented
as sets in Z2, is defined as:
The set of all points z such that B, translated by z, is contained by A.

Informally, the structuring element B is positioned with its
origin at (x, y) and the new pixel value is determined using
the rule:

1 if B fits A
g ( x, y ) = 
0 otherwise
14 Erosion Example
Original Image
Processed Image
Structuring Element
15 Erosion Example
Original Image Processed Image With Eroded Pixels
Structuring Element
16 Erosion Example 1
Original image Erosion by 3*3 Erosion by 5*5

square structuring square structuring
element element
Watch out: In these examples a 1 refers to a black pixel!
17 What Is Erosion For?
Erosion can split apart joined objects
Erosion can strip away extrusions
Watch out: Erosion shrinks objects

18 Example 2
19 Erosion Example 3
After erosion
Original with a disc of
image radius 15
After erosion After erosion

with a disc of with a disc of
radius 11 radius 45
20 Dilation
Dilation of image A by structuring element B is given by:
The set of all displacements z, such that and A

overlap by at least one element.
Informally: The structuring element B (if symmetric with
origin at center, else reflected) is positioned with its
origin at (x, y) and the new pixel value is determined
using the rule: g ( x, y ) = 

1 if B hits A
0 otherwise
21 Dilation Example
Original Image
Processed Image
Structuring Element
22 Dilation Example
Original Image Processed Image With Dilated Pixels
Structuring Element
23 Dilation Example 1
Original image Dilation by 3*3 Dilation by 5*5

square structuring square structuring
element element
Watch out: In these examples a 1 refers to a black pixel!
Original image After dilation
Structuring element
26 Quiz:
What is the result of this dilation?
0 0 0 0 0 1 0
0 1 1 0 ⊕ 0 1 1 =
0 0 0 0 0 0 0
27 Quiz:
What is the result of this dilation?
0 0 0 0 0 1 0
0 1 1 0
0 1 1 0 ⊕ 0 1 1 = 0 1 1 1
0 0 0 0 0 0 0 0 0 0 0
28 What Is Dilation For?
Dilation can repair breaks
Dilation can repair intrusions
Watch out: Dilation enlarges objects

29 What Is Dilation For?
Combined with erosion: eliminate irrelevant details
structuring element B = 13x13 pixels of gray level 1

30 Compound Operations
More interesting morphological operations can be
performed by performing combinations of erosions
and dilations
The most widely used of these compound operations
are:
– Opening
– Closing
31 Opening
The opening of image f by structuring element s,
denoted f ○ s is simply an erosion followed by a
dilation
f ○ s = (f s) s
Note: a disc shaped structuring element is used

32 Opening Example
Original Image
Processed Image
Structuring Element
33 Opening Example
Original Image Processed Image
Structuring Element
34 Opening Examples
35 Closing
The closing of image f by structuring element s,
denoted f • s is simply a dilation followed by an erosion
f • s = (f s)s
Original shape After dilation After erosion

(closing)
Note a disc shaped structuring element is used

36 Closing Example
Original Image Processed Image
Structuring Element
37 Closing Example
38 Morphological Processing Example
39 Morphological Algorithms
Using the simple technique we have looked at so far
we can begin to consider some more interesting
morphological algorithms such as:
– Boundary extraction
– Region filling
– Extraction of connected components
– Hit or Miss Transformation
– Convex Hull
– Thinning/thickening
40 Boundary Extraction
Extracting the boundary (or outline) of an object is
often extremely useful
The boundary can be given simply as
β(A) = A – (AB)
41 Boundary Extraction Example
A simple image and the result of performing boundary
extraction using a square 3*3 structuring element
Original Image Extracted Boundary

42 Connected Components
Binary image with multiple 'objects'

Separate 'objects' must be labeled
individually
6 Connected Components
43
Connectivity
Two points in an image are 'connected' if a path
can be found for which the value of the image
function is the same all along the path.
P1 connected to P2
P3 connected to P4
P1
P1 not connected to P3 or P4
P2
P4
P3
44 Finding Connected Components
Pick any pixel in the image and assign it a label

Assign same label to any neighbor pixel with the
same value of the image function
Continue labeling neighbors until no neighbors
can be assigned this label
Choose another label and another pixel not
already labeled and continue
If no more unlabeled image points, stop.
45 Connected Components Labelling
46 Recursive algorithm for CCL
Pseudo code:
1. Scan the image to find an unlabeled unity valued pixel and

assign it a new label L.
2. Recursively assign a label L to all its unity valued neighbors.
3. Stop if there are no more unlabeled unity valued pixels.
4. Go to step 1.
47 Sequential algorithm for CCL
First Pass
• Intialize CC=0 and scan pixels across rows. So when we visit pixel p, pixel l
and t have been visited and labeled.
t
• If p=0: no action; l p
• If p=1: examine l and t.
– both l and t = 0; CC++ and assign CC to p.
– only one of l and t is 1; assign its label to p.
– both l and t are 1:
* same label => assign it to p;
* different label => assign lowest no. to p and
Mark labels in neighbours as equivalent (i.e
they are the same) and store in an equivalence
table Image Processing and Computer Vision by Dr. Praveen Kumar @ CSE, VNIT
First Pass
t
• If p=0: no action; l p
First Pass
12
• If p=0: no action;
13
50 The Re-Labeling Process
Second Pass: simply replace the label with equivalent
label stored in the table
12
13
– But resolving equivalences is necessary between the two

passes. Different methods:
• Union-Find disjoint sets and operations
• Obtain transitive closure using Floyd-Warshall algorithm
51 Why Equivalence Resolution?
14
89
56
25
37
7  10
52 Using morphology
X k = ( X k −1  B)  A
53 Application: Blood cell counting
• Thresholded/binary Red
blood cell image
• Many blood cells are
separate objects
• Many touch – bad!
• Salt and pepper noise from
thresholding
• What kind of analysis can be
done ?
54 Results of analysis
• 63 separate objects detected
after CCL
• Compute features like area, BB,
centroid etc.
• Single cells have area about 50
• Noise spots
• Gobs of cells
• What should be done to get
better results?
55 Region Filling
Given a pixel inside a boundary, region filling attempts
to fill that boundary with object pixels (1s)
Given a point inside

here, can we fill the
whole circle?
56 Why Region filling?
Helpful in the
later stages of
processing
skeleton before skeleton after
closing closing
57 Area fill and closing
Areafill Operation: extract connected components of dark regions and check
their area . If the area is less than the threshold value they will be treated as
holes. Then the dark holes can be removed by changing their intensity into
light/white color.
Another criteria can be compactness (defined as perimeter^2 /area)
58 Region Filling (Floodfill Operation)
The key equation for region filling is
X k = ( X k −1  B)  Ac k = 1,2,3.....
Where X0 is simply the starting point inside the
boundary, B is a simple structuring element and Ac is
the complement of A
This equation is applied repeatedly until Xk is equal to
Xk-1
Finally the result is unioned with the original boundary
59 Region Filling Step By Step
0
1
X0=p
X k = ( X k −1  B)  Ac
60 Region Filling Example
Original Image One Region All Regions

Filled Filled
63 The Hit-or-Miss Transformation
• Useful as a very basic tool for shape detection
• To find the location of a shape B1 in an image A
– Erosion of A Ө B1 gives all places where B1 fits in A
• But B1 fits in any sufficiently large shape
• So also require the boundary (B2) around the
shape, to be empty
– Erosion of Ac Ө B2 gives all places where B2 fits in empty places of A

B = ( AOX )  Ac O(W − X )
A＊  B = ( X ,W − X )

B = ( AOX )  Ac O(W − X )
A＊  B = ( X ,W − X )
66 Corner Detection
• Method can also be used to look for particular patterns of
foreground and background pixels
• In such case background is not required and hit or miss transform
reduces to simple erosion
• Using a Group of Structuring Elements, containing 0s, 1s
and sometimes don’t cares(x!)
Ex: Find all the

right angle convex
corners of a region
in a given image
as shown below
67 Corner Detection
Structuring Elements representing four corners
Contains 0s, 1s and don’t care’s
x x x x x x
x x x x x x
Apply each Structuring Element

Use OR operation to combine the four results
68 Convex Hull
• A set A is said to be convex if the straight
line segment joining any two points in A
lies entirely within A.
• The convex hull H of an arbitrary set S is
the smallest convex set containing S.
• Application in shape analysis etc.
69 Convex Hull
Let B i , i = 1, 2, 3, 4, represent the four structuring elements.

The procedure consists of implementing the equation:
X ki = ( X k −1O
* Bi )  A
i = 1, 2, 3, 4 and k = 1, 2, 3,...
with X 0i = A.
When the procedure converges, or X ki = X ki −1 , let D i = X ki ,
the convex hull of A is
4
C ( A) =  D i
i =1
70 Convex Hull
71 Skeletonization
• Is a concise representation of a shape - the set of
all points that are equally distant from two closest
points of the object boundary
72 Skeletonization
Results are
sensitive to small
perturbations in
the boundary
73 Thining
1. Used to remove selected foreground pixels
from binary images
2. After edge detection, lines are often thicker
than one pixel.
3. Thinning can be used to thin those line to one
pixel width.
4. It is particularly useful for skeletonization
74 Thining
The thinning of a set A by a structuring element B,
defined: A  B = A − ( AO * B)
= A  ( AO
* B )c
A more useful expression for thinning A symmetrically
is based on a sequence of structuring elements:
 B = B1 , B 2 , B 3 ,..., B n 
where B i is a rotated version of B i -1
The thinning of A by a sequence of structuring element {B}
A  {B} = ((...(( A  B1 )  B 2 )...)  B n )
75 Thining
76 Grayscale Morphology
• Instead of binary images, assume we have gray
values f(x,y), where x,y are integer pixels
• Let b(x,y) be the gray level structuring element
b(x,y) 250
200
150
100
• We will only focus on flat (constant 50
value) structuring elements 0

0 50 100 150 200 250 300 350 400
Distance along profile
• Definition of dilation with flat structuring element
(f  b)(x, y) = maxf (x − s, y − t)
(s,t)b
• Reflect b(s,t), slide it past f(x,y); at each position, take the maximum
of f(x,y) within the window of b(s,t)
b(s,t)
• Definition of dilation with flat structuring element
(f  b)(x, y) = maxf (x − s, y − t)
(s,t)b
• Reflect b(s,t), slide it past f(x,y); at each position, take the maximum
of f(x,y) within the window of b(s,t)
b(s,t)
79 Gray Scale Erosion
• Definition of erosion (with flat SE)
(f  b)(x, y) = (s,tmin
)b
f (x + s, y + t)
• Slide b(s,t) past f(x,y); at each position, take the minimum of f(x,y)
within that window
b(s,t)
80 Gray Scale Erosion
• Definition of erosion (with flat SE)
(f  b)(x, y) = minf (x + s, y + t)

(s,t )b
• Slide b(s,t) past f(x,y); at each position, take the minimum of f(x,y)
within that window
b(s,t)
81 Example
82 Opening and Closing
• Similar to binary case
f b = ( f θb) b
– Opening is erosion followed by
dilation
– Closing is dilation followed by f •b = ( f b)θb
erosion
• Geometric interpretation of opening
– Push the SE up from below against the
underside of f
– Take the highest values achieved at
every point
Opening removes small,

bright details
83 Opening and Closing
• Geometric interpretation of closing
– Push the SE down from above against the topside of f
– Take the lowest values achieved at every point
Closing removes
small, dark details
84 Example
• Segment the coins from the image
I = imread('pennies.png');
I = rgb2gray(I);
imshow(I,[]);
% Pick a disk that will just fit inside the pennies
SE = strel('disk', 6, 0);
I2 = imopen(I,SE);
figure, imshow(I2,[]);
% Threshold to get the pennies
B = Iopen > 150;
figure, imshow(B);
L = bwlabel(B);
props = regionprops(L);
figure(1);
for i=1:length(props)
rectangle('Position', props(i).BoundingBox, 'EdgeColor', 'r’);
end
85 Top Hat Transform
• Open the image with a structuring element, subtract it
from the original h = f − ( f b)
• Leaves only details smaller than the structuring element
• Matlab example
I = imread('rice.png');
% Pick structuring element larger than a rice grain
background = imopen(I,strel('disk',15));
subplot(1,2,1), imshow(I);
subplot(1,2,2), imshow(background);
I2 = I-background;
figure, imshow(I2, []);
86 Quiz
• Describe a morphological algorithm to count the number
of objects with holes and the number of objects without
holes in below image
87 Summary
• The purpose of morphological processing is
primarily to remove imperfections added during
segmentation
• The basic operations are erosion and dilation
• Using the basic operations we can perform
opening and closing
• More advanced morphological operation can then
be implemented using combinations of all of these
• Analogous things could be done with grayscale

07 - Morphology

Uploaded by

Copyright:

Available Formats

07 - Morphology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

07 - Morphology

Uploaded by

Copyright:

Available Formats

Image Processing & Computer vision

Morphological Image Processing

How do we fill “missing pixels”?

Image after segmentation Image after segmentation and

The set of all points z such that B, translated by z, is contained by A.

Original image Erosion by 3*3 Erosion by 5*5

Watch out: In these examples a 1 refers to a black pixel!

Erosion can strip away extrusions

Watch out: Erosion shrinks objects

After erosion After erosion

The set of all displacements z, such that and A

Original image Dilation by 3*3 Dilation by 5*5

Watch out: In these examples a 1 refers to a black pixel!

What is the result of this dilation?

Dilation can repair intrusions

Watch out: Dilation enlarges objects

structuring element B = 13x13 pixels of gray level 1

Note: a disc shaped structuring element is used

Original shape After dilation After erosion

Note a disc shaped structuring element is used

Original Image Extracted Boundary

Binary image with multiple 'objects'

Pick any pixel in the image and assign it a label

1. Scan the image to find an unlabeled unity valued pixel and

2. Recursively assign a label L to all its unity valued neighbors.

3. Stop if there are no more unlabeled unity valued pixels.

– But resolving equivalences is necessary between the two

Given a point inside

Original Image One Region All Regions

Ex: Find all the

Apply each Structuring Element

Let B i , i = 1, 2, 3, 4, represent the four structuring elements.

• We will only focus on flat (constant 50

value) structuring elements 0

(f  b)(x, y) = minf (x + s, y + t)

Opening removes small,

You might also like

Original image Erosion by 33 Erosion by 55

Original image Dilation by 33 Dilation by 55