PATTERN RECOGNITION

CHAPTER 1
INTRODUCTION

Dr.Negm Shawky
EMAIL: negmshawky@gmail.com
Basics

2
How you can concentrate in the lecture
Course Contents
● INTRODUCTION

● BAYESSIAN DECISION THEORY

● HIDDEN MARKOV MODEL

● SVM , KNN ALGORITHMS

● LINEAR CLASSIFIERS (perceptron algorithm)

● NONLINEAR CLASSIFIERS (neural network)

● CLUSTERING ALGORITHMS (k‐means algorithm)

● DEEP LEARNING ALGORITHMS

 Materials
• The course slides
• Reference papers
• Texts book
R.O.Duda,P.E.Hart,D.G.Stork,
Pattern Classification,2nd edition, John
Wiley&Sons,Inc.,2000.

K.Koutroumbas,S.Theodoridis,
Pattern Recognition,3td/4th edition,
AcademicPress,2006/2009.
 Materials
•Texts book
Christopher M. Bishop, Pattern Recognition
.
and Machine Learning, 2011

Richard Szeliski,Computer Vision: Algorithms

and Applications. 2nd Edition,2022

Ian Goodfellow, Yoshua Bengio, and

Aaron Courville Deep Learning, 2016
Introduction
Definition of Pattern Recognition
 First let’s understand the meaning of following topics that are
some of the most exciting technologies in today’s world.
 Artificial intelligence (AI), machine learning (ML), and deep
learning (DL) are often everyday mentioned in the news, movies.

 Pattern recognition is a type of machine learning (It is a central

part of machine learning) since it uses machine learning
algorithms to recognize patterns.
Definition of Pattern Recognition

 What is Pattern Recognition (PR)?

What is Pattern?
What is Recognition?
What is Pattern Recognition?

 Why the Pattern Recognition is required for us?

 this refer to the necessity and importance for pattern

recognition

 How to perform Pattern Recognition?

 this refer to the building blocks of a pattern recognition

system
 What is Pattern?
Pattern is a description of an object that is based on human
perception to understand of how to discriminate the objects

The description is represented by a set of measurements

describing a “physical” object
Note:
Some examples…… Pattern is everything
around in this digital
world can be observed
mathematically.
Example: a pattern is
represented using vector
feature values.

Barcode Geometry Footprint

 What is Pattern?
 Human Perception
Based on sensing (objects)in the environment and taking actions
according to what is observed, Humans perception need to highly
sophisticated skills.

For example
• Recognizing a face,
• Understanding spoken words,
• Reading hand writing

Instead of human we could like to give similar capabilities to machines

 What is Pattern? (Cont.)

 There are various kinds of patterns

● Visual patterns such as eyes, nose, mouth, face, fingerprint, etc.

● Temporal patterns such as speech, audios, videos, data streams,

etc.

● Logical patterns such as characters, strings, images, etc.

 What is Recognition?
Recognition can be defined as Identification of a pattern to be
specified into category we already know” this mean classifying an
object to a pattern class (caterogry)or identifying unknown patterns
using past experience

There are two types of recognition

Classification Categories are known and the task is to

Classification assign a proper class label (or category) for
each pattern

Clustering Categories are unknown and the task is

Clustering to learn or create categories as groups of
………………………………..similar and different patterns
 Classification vs. Clustering
Classification: An example

We already know the categories

of characters, and then classify
the handwritten characters into
category “A” and category “B”

Clustering: An example

We do not know the categories of

symbols, and then create the
categories with grouping the
symbols according to similarity
 What is Pattern Recognition?
 Pattern recognition is:
 The assignment of a physical object to one of several
prespecified categories “by Duda, Hart and Stork”

 Pattern recognition is about guessing or predicting the unknown

object and a discrete quantity is obtained such as black or white,
one or zero, sick or healthy, real or fake” by Devroye, Györfi and
Lugosi

 Pattern recognition is the process of recognizing patterns by using a

machine learning algorithm.

 Pattern recognition can be defined as the classification of data based

on statistical information extracted from patterns.
 What is Pattern Recognition? (Cont.)
 Three main elements are required to be studied for creating any
application by the machine based on pattern recognition

● Perceive view
● Process view
● Prediction view

 Perceive: Observe the environment (i.e. interact with the real

world)

 Process: Learn to distinguish patterns of interest from their

background‐ learn to create the model for discriminating
the object

Prediction: make right and reasonable decisions about the

categories of the patterns
 Why the pattern recognition is required for us?
 The goal of pattern recognition is based on the classification of
objects into a number of categories or classes.

 the objects may be images or signal waveforms or any type of

measurements are needed to be classified.

 The real power of human thinking is based on recognizing patterns

thus, instead of human thinking a better machine can achieve the
tasks of pattern recognition according to the applications required
in our life that lead to more humanlike.

 the science of Pattern recognition based on better machine

almost is used in designing all automated and intelligent
systems!
Why the pattern recognition is required for us? (Cont.)
For humans, pattern recognition is natural & easy

● recognize a face
● understand spoken words
● read handwritten characters
● identify items by feel (decide whether an apple is ripe by its
smell ) ……

For computers, pattern recognition is never easy

in spit of, pattern recognition is important, useful, attractive, but has

challenging

face the Challenges obtain an application of intelligent system

Applications of Pattern Recognition
1) Character Recognition

Input:
Output:
images with characters
the identified
(normally contaminated
character strings
with noise)
(Earham encourage)

Useful in scenarios such as automatic license plate

recognition (ALPR), optical character recognition (OCR),
etc.
Applications of PR (Cont.)
2) Speech Recognition

Input: Output:
acoustic signal (e.g. contents of the
sound waves) speech

Useful in scenarios such as speech‐to‐text (STT), voice

command & control etc.
Applications of PR (Cont.)
3) Fingerprint Recognition

Input:
Output:
fingerprints of some
the person’s identity
person

Useful in scenarios such as

computerized access control,
criminal pursuit, etc.
Applications of PR (Cont.)
4) Signature Verification

Input:
Output:
signature of some person
the signatory’s
(sequence of dots)
identity

Useful in scenarios such as

digital signature verification,
credit card anti‐fraud, etc.
 Applications of PR (Cont.)
5) Face Detection

Input: Output:
images with several people locations of the
peoples’ faces in the
1 image

1
4
5

3 2

Useful in scenarios such as

digital camera capturing,
video surveillance, etc.
 Applications of PR (Cont.)
6) Text Categorization

Input: Output: category of

document, web pages, etc. the text, such as
political, economic,
military, sports, etc.

Useful in scenarios such as

information retrieval,
document organization, etc.
Commercial Applications in Pattern Recognition

English hand writing recognition

 Commercial Applications in Pattern Recognition
Biometric recognition
Biometric recognition refers to the advanced method of recognizing biological and
behavioral characteristics of an individual.
Types of Biometrics:
Biological Characteristics: These are related to the shape of the body.
Examples include:
Fingerprint: The unique patterns on our fingertips.
Face Recognition: Analyzing facial features.
Iris Recognition: Examining the iris of the eye.
Voice: Distinctive vocal characteristics.
Behavioral Characteristics: These relate to patterns of behavior:
Mouse Movement: How we move the mouse.
Signature: Our individual handwriting.
Commercial Applications in Pattern Recognition
Autonomous navigation
Commercial Applications in Pattern Recognition

Cancer detection and grading using microscopic tissue data

Commercial Applications in Pattern Recognition

Land cover classification using satellite data

 Pattern Recognition Applications
Problem Domain Application Input Pattern Pattern Classes
Document Internet search Text document Semantic categories
classification
Document Junk mail filtering Email Junk/non‐junk
classification
Natural language Information Sentences Parts of speech
processing extraction

Medical Computer aided Microscopic image Cancerous/healthy cell

diagnosis

Military Automatic target Optical or infrared Target type

recognition image

Industrial Printed circuit board Intensity or range Defective/non‐defective

automation inspection image product
 Pattern Recognition Applications

Problem Domain Application Input Pattern Pattern Classes

Industrial Fruit sorting Images taken on a Grade of quality
automation conveyor belt

Bioinformatics Sequence analysis DNA sequence Known types of genes

 END

P
P
P
P R
R
R
R

57