0% found this document useful (0 votes)
43 views36 pages

Image Feature Extraction and Segmentation

This document discusses image segmentation and feature extraction. It covers: 1. The purpose of image features is to describe objects in a meaningful way to assist with object classification. Feature extraction involves extracting and generating features. 2. Characteristics of good features include robustness, shift/rotation/size invariance, discrimination, reliability, and noise resistance. Features are stored as a feature vector for classification. 3. Image segmentation algorithms covered include discontinuity-based (point/line/edge detection) and similarity-based (thresholding, region growing, split and merge) methods. Edge detection finds boundaries between regions with distinct intensities.

Uploaded by

Monark Mehta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
43 views36 pages

Image Feature Extraction and Segmentation

This document discusses image segmentation and feature extraction. It covers: 1. The purpose of image features is to describe objects in a meaningful way to assist with object classification. Feature extraction involves extracting and generating features. 2. Characteristics of good features include robustness, shift/rotation/size invariance, discrimination, reliability, and noise resistance. Features are stored as a feature vector for classification. 3. Image segmentation algorithms covered include discontinuity-based (point/line/edge detection) and similarity-based (thresholding, region growing, split and merge) methods. Edge detection finds boundaries between regions with distinct intensities.

Uploaded by

Monark Mehta
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 36

CSE4019

Image Segmentation and Feature Extraction


Before reading this presentation
file, go through our
class recorded content

Presentation by:
S.Rajkumar
Associate Professor
School of Computer Science and Engineering
email: [email protected]
Image Feature
• The purpose of an image feature is to describe
the object in a meaningful manner.
• This helps in the discrimination of these
objects
• Feature extraction is the process of extraction
and generation of features to assist the task of
object classification.
• The quality of features influence the task of
object classification.
Image Feature
• The extended set of features is stored as a
vector called the feature vector.
• A feature vector or pattern vector is a vector
that contains ‘n’ measured values
• The classifier takes the feature vector as input
and performs the classification.
Characteristics of a good feature
• Robustness – The property of a features
invariance to operations such as translation,
rotation, scaling and properties such as
illumination, noise and artefacts are called as
robustness.
• Shift invariance – The ability of the feature to
remain constant when shift operations are
performed
Characteristics of a good feature
• Rotation invariance - This is the ability of the
feature to remain constant when rotated.
• Size invariance – This is the ability of the
feature to remain constant when size is
changed.
• Occlusion invariance – When all or some parts
of the object are hidden, the property of the
feature do not change.
Characteristics of a good feature
• Discrimination – The properties should
distinguish one object from the other.
• Reliability – The similar objects should have
similar values.
• Resistance to noise – A good feature should be
immune to noise
• Not all features exhibit these properties.
• Suitable parameters that distinguish the object
uniquely should be identified and extracted
Feature Extraction
• The segmentation process segments the
foreground object using a boundary.
• This boundary should be encoded into a form
that can be processed by the computer.
• It involves coding of the boundary region.
• This stage is called as object representation.
Feature Extraction
• Once the boundary is identified , the features
that characterize the object should be
described.
• This stage is called object description.
Feature Extraction
• It is a dimensionality reduction
• It represents the interesting parts of an image
in a more compact representation
• Mostly preferred in the case of large images
• Features – The process of extracting useful
information from an image
• The features should carry unique attributes of
the image
Example – Features paper
(I)Image segmentation and purpose
(II)Algorithms
1. Discontinuity based
(a) Point detection
(b) Line detection
(c) Edge detection
2. Similarity based
(a) Thresholding
(b) Region growing
(c) Split and Merge
(III)Image segmentation analysis
Image Segmentation
• It is the process of sub dividing the image in to its
constituent parts
• ROI
• Locate the Objects

---------RajkumarS----------
Point detection

• A point is said to be detected at the location


on which the mask is centered if |R|>=T
Line detection
Line
Detection
• In line detection, four types of masks are used
to get the responses R1, R2, R3, and R4 for
four directions vertical, horizontal, 450 and -
450
• These masks are applied to the image.
• R1 is the response for moving the mask from
left to right of its image
• R2 is the response for moving the mask from
top to bottom of its image
Line
Detection
By moving the mask over the
entire image, it detect all
those points which lie on the
horizontal line
• If we move this mask on the
entire image it will help to detect
all the point in the image which
are lying on the line which is
inclined at an angle of 45
degrees.
• Help to detect all the point
which lie on the vertical line
• Detect the points which
are lying on the line which
are inclined at an angle of -
45 degrees
Line Detection

Mask Image Matrix


= -1*z1-1*z2-
1*z3:2*z4+2*z5+2*z6-1*z7-1*z8-
1*z9
• For line detection apply all the 4 masks on the
image.
• If the above condition is satisfied, the
corresponding point is more associated with
the line with the corresponding mask i.
Line Detection
• R(x, y) = max(|R1(x, y)|, |R2(x, y)|, |R3(x, y)|,
|R4(x, y)|)
Edges
Edges
• Boundary between 2 regions in an image having distinct
intensity / grey levels
• It is the region but there is a variation either from the low
intensity to a high intensity or vice versa
• Edge provide the outline of the object
• Edges correspond to the discontinuities in depth and light
variations
• Set of connected pixels that lies boundary between 2 regions
that differ in grey value
• The pixels on an edge are called as edge points
• A reasonable definition of the edge requires the ability to
measure grey level transititions in meaningful manner.
Edge detection
• Consider an image strip I = 60 60 60 100 100 100
First order derivative:
0 0 40 0 0
Second order derivative: 0 40 -40 0
High value 40 indicates the presence of an edge in first
order derivative
Sign change in second order derivative represents the
edges
Thresholding
Histogram based Thresholding
Global Thresholding
Algorithm
• Steps of an iterative algorithm for choosing a
thresholding value:
1. Choose an initial threshold T=T0 where T0 is the
mean of the two peaks or the mean of pixel
value.
2. Calculate the mean value of the pixel below the
threshold(µ1) and the mean value of the pixels
above this threshold (µ2)
3. Compute a new threshold as T = µ1+µ2/2
4. Repeat 2,3 until no change in T
Thresholding
Problems in Global thresholding and why we go for Adaptive Thresholding?
Thresholding
Problems in Global thresholding and why we go for Adaptive Thresholding?
References
1. R.C. Gonzalez & R.E. Woods, Digital Image Processing, Pearson Education,
Third Edition

You might also like