National University of Sciences and

Technology (NUST)
IGIS

Digital Image Processing

Khawar Khurshid 1
What are your
expectations?

2
What should be my
expectations?

3
What do you want to
learn in this class?

4
Ground rules

5
Schedule

• Office Hours

– Monday 2:00pm - 4:00pm

or By Email

• Attendance
- Minimum 75% to take the final exam.

6
Distribution - Tentative

• Quizzes
5-10%
• Assignments
5-10%
• OHTs
30%
• Project
10-20%
• Final
35-40%

7
Plagiarism

8
Plagiarism

9
ADIP - Books

10
ADIP - Software

Python

Vs

Matlab
11
Applications
Of
Image Processing

12
Contrast Enhancement

13
Image Sharpening

14
Color Enhancement

15
Area Classification

16
Geographic Information System

◼ Manipulation of Satellite Imagery

◼ Terrain Classification

◼ Weather Imaging

17
Digital Elevation Model

18
Image Segmentation

19
Industrial Automation

20
Medical Imaging

21
Face Detection

22
Face Detection

23
 Emotion Classification

◼ Implicit customer feedback

Khawar Khurshid 24
 Emotion Detection

https://azure.microsoft.com/en-us/services/cognitive-services/emotion/
25
Morphing

26
Morphing

27
Sign Language/Gesture Recognition

British Sign Language Alphabet

28
 Lip Reading

Khawar Khurshid 29
 Biometrics

 Physiological Biometrics
◼ Face, IRIS, DNA, Finger Prints

 Behavioral Biometrics
◼ Typing Rhythm, Handwriting, Gait

Khawar Khurshid 30
 Text Recognition

Khawar Khurshid 31
 Automated Mail Sorting

Khawar Khurshid 32
 Content Based Video Retrieval

Khawar Khurshid 33
 Common Operations on Images

Khawar Khurshid 34
 Common Operations on Images

Khawar Khurshid 35
 Common Operations on Images

Khawar Khurshid 36
Overview

• Basic Level
– Image Acquisition, Image Representation
– Image Digitization, Quantization
– Morphological Operations, Color Corrections
– Enhancements, De-noising, Sharpening

• Mid Level
– Segmentation, Registration
– Compression, Data Hiding
– Image Restoration, Blur Correction
– Image Transforms

• High Level
– Feature Detection, Feature Analysis
– Intro to Biomedical Imaging
– Classification
37
Outline (Subject to Modifications)

• Image Fundamentals

• Sampling, Quantization, Interpolation

• Intensity Corrections, Color Processing

• Morphological Operations

• Transforms

• Restoration

• Compression

• Segmentation

• Image Analysis

• Feature Detection and Analysis

38
 Types of Images

– Binary

– Gray Scale

– Color

– Indexed

Khawar Khurshid 39
 Pixels

Khawar Khurshid 40
 Pixels
1
2
3
4
5

Size of image: 6
7
? rows, 8
? columns 9
10
11
12
13
14
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Khawar Khurshid 41
 Pixels
1
2
3
4
5

Size of image: 6
7
14 rows, 8
15 columns 9
10
11
12
13
14
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Khawar Khurshid 42
 Pixels

Size of image: 1
2
14 rows, 3

15 columns 4
5
6
Image Color: 7

RGB 8
9
10
11
12
13
14
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Khawar Khurshid 43
 Pixels
Size of image:
14 rows,
15 columns

Image Color:
RGB

Red: 8 bits
Green: 8 bits
Blue: 8 bits

Storage size of
image=?
Khawar Khurshid
 Pixels
Size of image:
14 rows,
15 columns

Image Color:
RGB

Red: 8 bits
Green: 8 bits
Blue: 8 bits

Storage size of
image=? 14*15*8*3
Khawar Khurshid
 Image Storage Requirements

Khawar Khurshid 46
 Image Storage Requirements

Khawar Khurshid 47
 Introduction

• An image is a 2-dimensional function fxy where x and y

are the spatial coordinates, and fxy is the intensity of
the image at that point.

• A digital image is the representation of a continuous

image f(x,y) by a 2-d array of discrete samples. The
amplitude of each sample is quantized to be
represented by a finite number of bits.

• Each element of the 2D array of samples is called a

pixel (for ‘picture element’)
– Color Images have 3 values per pixel.
– Monochromatic Images have 1 value per pixel.

Khawar Khurshid 48
 Intensity Resolution

Khawar Khurshid 49
 Intensity Resolution

Khawar Khurshid 50
 Image Representation

Khawar Khurshid 51
 Image Representation

Khawar Khurshid 52
 Image Representation

Khawar Khurshid 53
 Image Representation

Khawar Khurshid 54
 RGB -> Gray -> Binary
B = im2bw(G, level);

Level = 0.1

I = imread(‘rain.jpg’) I -> 480x720x3

Level = 0.3

Level = 0.5

Khawar Khurshid
G = rgb2gray(I); 55
 R-G-B Color Components

G = I;
G(:,:,1) = 0;
G(:,:,3) = 0;

Khawar Khurshid 56
 R-G-B Color Components

RGB -> RBG

I(:,:,2) = B(:,:,3);
I(:,:,3) = G(:,:,2);
Khawar Khurshid 57
 Resolution

• How many samples and gray levels are required for a

good approximation?

• Quality of an image depends on number of pixels and

number of gray-levels.

• The more these parameters are increased, the closer

the sampled/quantized array approximates the original
image.

• But storage and processing requirements increase

rapidly as a function of N, M, and k.

Khawar Khurshid 58
 Resolution

Depends on what is in the image and what you would like

to do with it.

The picture on the right is fine for counting the number of cars,
but not for reading the number plate.

Khawar Khurshid 59
 End
Introduction

Khawar Khurshid 60