Introduction to pattern recognition

■ Outline:
1. What is pattern recognition?
2. How to perform pattern recognition?
3. Some examples
4. Components of a pattern recognition system
5. How to evaluate a pattern recognition system?
 Introduction

■ Outline:
1. What is pattern recognition?
2. How to perform pattern recognition?
3. Some examples
4. Components of a pattern recognition system
5. How to evaluate a pattern recognition system?
Examples of patterns
 What is pattern recognition?

▪ One of the most common applications of machine learning

▪ Important component of AI systems

▪ Aiming to give human perception capabilities to machines

▪ Process of recognizing regularities in data by using machine
learning algorithms, based on statistical information,
historical data, or the machine’s memory.

▪ During recognition given objects are assigned to prescribed

classes
Applications of pattern recognition

People detection
Applications of pattern recognition

Face recognition

Paul
Alex
Bob
Applications of pattern recognition

License plate recognition

 Applications of pattern recognition

■ OCR (Optical Character Recognition): handwritten, printed text,…

■ Biometrics: face, finger prints, speech,…

■ Medical diagnosis: X-ray, MRI analysis,…

■ Biology: fruit, leaf,…

■ Education

■ Smart transportation: traffic light, traffic sign,…

 Applications of pattern recognition

■ Security surveillance
■ Entertainment, sport

■ Agriculture

■ Commercial
■ Military technology: target recognition, satellite image analysis,…

and more…
 Introduction

■ Outline:
1. What is pattern recognition?
2. How to perform pattern recognition?
3. Some examples
4. Components of a pattern recognition system
5. How to evaluate a pattern recognition system?
Grid-by-grid comparison

From: Talal Alsubaie

Grid-by-grid comparison

From: Talal Alsubaie

Grid-by-grid comparison

From: Talal Alsubaie

Grid-by-grid comparison

From: Talal Alsubaie

 The statistical way

■ Problems with grid-by-grid comparison: too costly

■ Solution: artificial intelligence/machine learning

■ Two phases: learning and classification

 Learning: learn the rule from data (supervised and unsupervised)

 Classification: decide a pattern to a known category/class
 Variability challenges

■ Intra-class
 Ex: different typefaces of
letter “T”
 Variability challenges

Inter-class

 Ex: lisianthus (hoa cát tường) and carnation (hoa cẩm chướng)
 Introduction

■ Outline:
1. What is pattern recognition?
2. How to perform pattern recognition?
3. Some examples
4. Components of a pattern recognition system
5. How to evaluate a pattern recognition system?
 Example 1

Sea bass

Fish

Salmon
 Problem analysis

 Set up a camera

 Take some sample images

 Note physical differences (features):

 Length

 Lightness

 Width

 Number and shape of fins

 Position of the mouth, etc…

 Classify fish
Implementation

To isolate fishes from one another

and from the background

To reduce the data by measuring

certain features

To discriminate salmon and sea

bass (by statistical, by machine
learning)
 Feature selection

■ Sea bass is generally longer than salmon  Length becomes

a feature.

Length is a poor feature

 Feature selection

■ Lightness feature: better than length feature but not adequate

■ Move the decision boundary toward smaller values of lightness to
reduce the number of sea bass that are classified as salmon

Lightness is better
 Feature selection

■ Multiple features
■ Lightness + width Fish x = [x1, x2]

Lightness width

Combination is better
 How many features

■ More is not always better

■ There could be redundancy

■ Over complication can occur

 Classification model selection

 There is a tradeoff between complexity of the decision rules and

their performances to unknown patterns.

 Generalization: The ability of the classifier to produce correct

results on novel patterns.

 Simplify the decision boundary!

Example 2
 System building

 Data:

 Images with people (positive)

 Images without people (negative)

 System building

 Modeling:
Human
 HOG features

 SVM classification model

Non-human
 System building

 Detecting:

 Sliding window
 Detecting a

person/non-person

within a sliding
window
(HOG+SVM)
 System building

 Detecting:

 Sliding window
 Detecting a

person/non-person

within a sliding
window
(HOG+SVM)
 System building

 Image pyramic:

 To change the window size

https://www.youtube.com/watch?v=sDByl84n5mY
 Introduction

■ Outline:
1. What is pattern recognition?
2. How to perform pattern recognition?
3. Some examples
4. Components of a pattern recognition system
5. How to evaluate a pattern recognition system?
 Pattern recognition system

Testing data
 The design cycle

Performance evaluation
Computational complexity evaluation
 Teachable machine

■ https://teachablemachine.withgoogle.com/
■ A web-based tool that:
 makes creating machine learning models (for sites, apps,…) fast,
easy, and accessible to everyone – no coding requires
 teaches a model to recognize images (from files/webcam), sounds
(MP3, *.wav), poses (from files/webcam)
 How to use teachable machine

https://youtu.be/DFBbSTvtpy4 https://youtu.be/n-zeeRLBgd0?t=50

https://youtu.be/CO67EQ0ZWgA
 ■ Fruit recognition:
https://www.youtube.com/watch?v=cBOSGQo1A10
■ Grading fruit:
https://www.youtube.com/watch?v=a8UZkCQqzNQ
■ Lung cancer prediction:
https://www.youtube.com/watch?v=JfZIJzWWhI0