Scribd
Upload
5 views42 pages

1c

The document covers key concepts in digital image processing, focusing on image representation, boundary following algorithms, and shape descriptors such as Fourier descriptors. It discusses techniques for analyzing image attributes, including color histograms and texture analysis, while also providing references for further reading. The content is structured as a lecture by Professor Gundimeda Venugopal, aimed at students in a computer vision course.

Uploaded by

freeguyfreeguy67
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
5 views42 pages

1c

The document covers key concepts in digital image processing, focusing on image representation, boundary following algorithms, and shape descriptors such as Fourier descriptors. It discusses techniques for analyzing image attributes, including color histograms and texture analysis, while also providing references for further reading. The content is structured as a lecture by Professor Gundimeda Venugopal, aimed at students in a computer vision course.

Uploaded by

freeguyfreeguy67
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/ 42

Computer Vision

(Course Code: 4047)

Module-1:Lecture-3: Image Representation and Analysis


Gundimeda Venugopal, Professor of Practice, SCOPE
Key Stages in Digital Image Processing
Image Representation and Description

Image Description deals extracting attributes that result in


some quantitative information of Interest.
Boundary Following Algorithm
Many algorithms require the points on the boundary of region be ordered Clockwise (or anticlockwise)
We assume
a) Object and Background are represented by 1 and 0 respectively
b) Image are padded with 0s to eliminate the possibility of object merging with image border
Boundary Following Algorithm
Polygonal Approximation: Minimum Parameter Polygon
MPP

b) Diagonal Mirror locations in the outer wall of bounding region denoted with orange dots
MPP Approximation Advantages
Merging: Polygonal Approximation
Splitting: Polygonal Approximation
Boundary Representation: Chain Code
Chain Code
❖Solution: Resample the boundary
by selecting a larger grid spacing

❖Then, as the boundary is


traversed, a boundary point is
assigned to each node of the large
grid, depending upon the
proximity of original boundary of
the node
❖The resampled boundary can now
be represented by 4 or 8 code
Chain Code
Chain Code
Shape Number
Fourier Descriptor
• Fourier descriptors are a way of encoding the shape of a two-dimensional object
by taking the Fourier transform of the boundary, where every point on the
boundary is mapped to a complex number.

• Fourier descriptors refer to a set of numbers derived from Fourier coefficients


that characterize the frequency content of a shape, allowing for a compact
representation of the shape by capturing its main features while ignoring small
variations caused by noise.

• The original shape can be recovered from the inverse Fourier transform. However,
if only a few terms of the inverse are used, the boundary becomes simplified,
providing a way to smooth or filter the boundary.

• Few Fourier descriptors can be used to capture the gross essence of a boundary. To define a curve that passes through
They can be used as a basis for differentiating between distinct boundary shapes. the k sampled points, we need to
consider only k/2 coefficients.
• Local frequency components provide most of the features a shape (boundary).
In practice, Fourier descriptors are
• High Frequency components account for finer detail ( corners). They are easily computed for fewer coefficients than
affected by noise and only represent detail that is of little value to recognition. the limit of k/2.

https://demonstrations.wolfram.com/FourierDescriptors/
Fourier Descriptor

At P=8, Reconstructed starts looking like a square than a circle


At P=56, Corners start to break out of the sequence
At P=61, the curves begin to straighten
Fourier Descriptors: Effect of Rotation, Translation & Scaling

Reference: https://cis.temple.edu/~lakaemper/courses/cis595_2004/papers/fourierShape.pdf
Regional Descriptors
Boundary Representation: Signature
Signature
Skeletons
Skeletons
Skeletons
Skeletons

(Not a Isolated or End point)


(Not a corner or junction point)

(This is Not part of a horizontal


/vertical line
Skeletons
Skeletons
Simple Boundary Descriptors
Simple Boundary Descriptors
Topological Descriptors
Topological Descriptors
Color Histogram
Color Histogram

A color histogram is a representation of


the distribution of colours in an image.

For digital images, a color histogram


represents the number of pixels that
have colours in each of a fixed list of
color ranges, that span the image's color
space, the set of all possible colours.

The color histogram can be built for any kind of color space, although the term is more often used for three-dimensional
spaces such as RGB or HSV. For monochromatic images, the term intensity histogram may be used instead.

For multi-spectral images, where each pixel is represented by an arbitrary number of measurements (for example,
beyond the three measurements in RGB), the color histogram is N-dimensional, with N being the number of
measurements taken. Each measurement has its own wavelength range of the light spectrum, some of which may be
outside the visible spectrum.
Global Color Histogram and Local Color Histogram
GCH is the most known color histogram used to detect similar images.
Feature extraction algorithm:
1.Discretize your color-space (images’ colors) into n color (You may use just 8*8*8 =512 color instead
of 256*256*256=16777216 color).
2.Create a bin for each color.
3.Count number of pixels for each color and store it in histogram’s bins.

Matching function:
The most common matching function for this method is Euclidean distance.
. To compare 2 images A, B.
A(R,G,B) : represents number of pixels in color = (R,G,B). (for example A(6,2,4) represents the
number of discretized pixels of color R=6,G=2 and B=4).
D: sum Euclidean distances.

Local Color Histogram:


LCH includes information about color’s distribution in different regions. It’s the same as GCH but at first we divide the image into
different block. Where each pair of the blocks (one of them in the first image and the other in the second) will be computed separately
using GCH. After that the total distance between the two images will be the sum of all GCH distances between them.

Remember : the larger the distance value, the less similar the images are.
GCH Example

Here C has the same color histogram as B but A is different from them.

Using Euclidian distance for these color histograms we found that D(A,C) = D(A,B) and D(B,C) = 0 but There’s a
problem here that B, C are not similar at all so D(B,C) shouldn’t be zero and D(A,C) should be smaller than D(A,B)
because A,C have the same pixels except for only two pixels.

GCH doesn’t include information about color spatial distribution.


What is Texture ?
Texture Characteristics
Texture Analysis
Texture Analysis
Texture Analysis: Structural Approach
Texture Analysis: Statistical Approach

Low number of athletes running 100m in less than 11sec


Texture Analysis: Statistical Approach
Texture: Statistical
Text Books and References
❖Text Books
➢ Richard Szeliski. ”Computer vision: Algorithms and Applications. Springer Nature, Second Edition”, 2022
➢ E. R. Davies, Computer Vision Principles, Algorithms, Applications, Learning, Elsevier,5th Edition, 2017

❖ References
➢ Rafael C. Gonzales, Richard E. Woods, “Digital Image Processing”, Fourth Edition, Pearson Education, 2018.
➢ Richard Szeliski ,”Computer Vision: Algorithms and Applications” , Springer 2015
➢ Intro to Digital Image Processing
➢ Digital Image Processing 2nd edition

PDF online

❖ Credits: Slides and Video content borrowed from:


➢ First Principles of Computer Vision
➢ https://robotacademy.net.au/
➢ MIT 6.S094: Computer Vision
➢ MIT Introduction to Deep Learning | 6.S191
➢ Digital Image Processing 2nd edition

You might also like