Download - PATTERN RECOGNITION - YunTech
1
PATTERN RECOGNITION (2014 Spring YunTech)
Instructor: Hsuan-Ting Chang, Ph.D.
Class hours: Tuesday. F,G,H;
Email: [email protected]
URL: http://teacher.yuntech.edu.tw/htchang/
Office: EN307
Phone: ext. 4263
Office hours: Thursday 10~12 AM
2014/2/18 YunTech EE Pattern Recognition
2
Text Book
Pattern Classification
R.O. Duda, P.E. Hart and D.G. Stork
Wiley-Interscience Publication, 2001
2nd Edition
2014/2/18 YunTech EE Pattern Recognition
3
Reference Books
Pattern Recognition: A Statistical
Approach
P.A. Devijver & J. Kittler
Prentice-Hall, 1982
Pattern Recognition: Statistical,
Structural and Neural Approaches
R. Schalkoff
John Wiley & Sons, 1992
2014/2/18 YunTech EE Pattern Recognition
4
Reference Books
Pattern Recognition Using Neural
Networks
C.G. Looney
Oxford University Press, 1997
Pattern Recognition Principles
J.T. Tou and R.C. Gonzalez
Addison-Wesley, 1981
2014/2/18 YunTech EE Pattern Recognition
5
Reference Books
Pattern Recognition and Image Analysis
E. Gose, R. Johnsonbaugh and S. Jost
Prentice-Hall, 1996
Numerical Recipes in C The Art of Scientific Computing (2nd)
W.H. Press, S.A. Teukolsky, W.T. Vetterling and B.P. Flannery
Cambridge, 1999
2014/2/18 YunTech EE Pattern Recognition
6
Important Academic References (Journals)
Pattern Recognition (PR)
Pattern Recognition Letters (PRL)
IEEE Trans. Pattern Analysis and Machine
Intelligence (IEEE PAMI)
IEEE Trans. System, Man and Cybernetics
(IEEE SMC) (Part A, Part B)
2014/2/18 YunTech EE Pattern Recognition
7
Important Academic References (Journals)
IEEE Trans. on Image Processing (IEEE IP)
IEEE Trans. on Circuits and Systems for Video Technology (IEEE CSVT)
IEEE Trans. on Information Theory (IEEE IT)
IEEE Trans. on Neural Networks (IEEE NN)
2014/2/18 YunTech EE Pattern Recognition
8
Important Academic References (Journals)
Computer Vision and Image Understanding (CVIU)
Graphical Modeling (GM) (Graphical Modeling and Image Processing)
Image and Vision Computing (IVC)
International Journal of Computer Vision (IJCV)
Machine Vision and Applications (MVA)
2014/2/18 YunTech EE Pattern Recognition
9
Important Academic References (Journals)
Neural Networks (NN)
影像與識別 (IPPR) 中華民國影像處理與圖形識別會刊
2014/2/18 YunTech EE Pattern Recognition
10
Important Academic References (Conferences)
International Conference on Pattern Recognition
(ICPR)
International Conference on Image Processing
(ICIP)
IEEE International Conference on Computer
Vision (ICCV)
IEEE Computer Society Conference on Computer
Vision and Pattern Recognition
(CVPR)
2014/2/18 YunTech EE Pattern Recognition
11
Important Academic References (Conferences)
International Joint Conference on Neural
Networks
(IJCNN)
IPPR Conference on Computer Vision,
Graphics and Image Processing
(Domestic)
International Computer Symposium
(ICS or NCS, Domestic)
2014/2/18 YunTech EE Pattern Recognition
12
Tentative Grading
Midterm Exam (30%)
Homework, paper presentation, & class
activities (40%)
Term project (30%)
(Implementation and Presentation)
2014/2/18 YunTech EE Pattern Recognition
Paper Presentations
Read a magazine article from IEEE, ACM,
OSA, SPIE, or other related academic
societies.
Read at least 1~2 strongly related papers
Write a report/summary (5-page Word file)
about the paper
13 2014/2/18 YunTech EE Pattern Recognition
14
Term Project
Goal
Identify the six classes of abnormal regions
about early gastric cancer (EGC) in magnified
NBI images in the stomach
Implementation of the method you use in
Matlab or C++ codes
More details will be described later
2014/2/18 YunTech EE Pattern Recognition
15
Term Project
Report Content
Abstract
Survey of related work
Proposed methods
Experimental results
Discussions and Conclusions
References
Hand in
Paper sheet and electronic file
Example
Example_of_report.doc
2014/2/18 YunTech EE Pattern Recognition
16
Term Project
Report
Electronic files
Experimental results
Power point file
The same format as lecture
Test images and results must be included
For presentation
Paper
Print of report
2014/2/18 YunTech EE Pattern Recognition
17
Course Outline (I) – In textbook
(Ch 1) Introduction
(Ch 2) Bayesian Decision Theory
(Ch 3) Maximum-Likelihood and Bayesian Parameter Estimation
(Ch 4) Nonparametric Techniques
(Ch 5) Linear Discriminant Functions
2014/2/18 YunTech EE Pattern Recognition
18
Course Outline (II) – Special topics
Support vector machine (SVM)
Principle component analysis (PCA)
Independent component analysis (ICA)
Back-propagation artificial neural network
Optical pattern recognition – architectures and algorithms
2014/2/18 YunTech EE Pattern Recognition
19
Website for downloading PDF files
http://teacher.yuntech.edu.tw/htchang/PR
1022.htm
2014/2/18 YunTech EE Pattern Recognition
20
Definition
Pattern recognition is the study of how
machines can
- observe the environment,
- learn to distinguish patterns of interest,
- make decisions about the categories of
the patterns.
2014/2/18 YunTech EE Pattern Recognition
21
What is a pattern?
Watanabe defines a pattern “as opposite of
chaos; it is an entity, vaguely defined, that
could be given a name”.
Examples:
- fingerprint,
- handwritten cursive word,
- a human face,
- a speech signal...
2014/2/18 YunTech EE Pattern Recognition
22
Types of recognition
Supervised classification: input pattern is
identified as a member of a pre-defined
class.
Unsupervised classification: input pattern
is assigned to a hitherto unknown class.
2014/2/18 YunTech EE Pattern Recognition
23
Applications (1) Problem
Domain
Application Input Pattern
Classes
Bioinformatics Sequence Analysis DNA/Protein
Sequence
Known types
of genes
Data Mining Search for
meaningful
patterns
Points in
multi-dimensional
space
Compact and
well-separated
clusters
Document
Classification
Internet Search Text document Semantic
categories
Document Image
Analysis
Reading machine
for the blind
Document image Alphanumeric
characters,
words
Industrial
Automation
PCB inspection Intensity or range
image
Defective/non-
defective
2014/2/18 YunTech EE Pattern Recognition
24
Applications (2)
Problem
Domain
Application Input Pattern
Classes
Multimedia
database
retrieval
Internet Search Video clip Video genres
Biometrics
recognition
Personal
identification
Face, iris,
fingerprint
Authorized users
for access control
Remote sensing Forecasting crop
yield
Multi-spectral
image
Land use
categories
Speech
recognition
Telephone
directory inquiry
Speech
waveform
Spoken words
2014/2/18 YunTech EE Pattern Recognition
25
Components of a PR System
Data acquisition and pre-processing
Data representation
Decision making
2014/2/18 YunTech EE Pattern Recognition
26
Pattern Recognition Methods
Template matching
Statistical approach
Syntactic approach
Neural networks
2014/2/18 YunTech EE Pattern Recognition
27
Template Matching
A template (typically a 2D shape) or a
prototype of the pattern to be recognized is
available.
Compute the similarity between the
template and the pattern to be matched.
Take into account pose (rotation,
translation) and scale changes.
2014/2/18 YunTech EE Pattern Recognition
28
Issues of concern
Choice of template
Computational complexity
Rigidity assumption (use deformable
template models)
2014/2/18 YunTech EE Pattern Recognition
29
Statistical Approach
E ach pattern is represented in terms of d
features, and is viewed as a point in a d-
dimensional space
The goal is to choose those features that
allow pattern vectors belonging to different
categories to occupy compact and disjoint
regions in a d-dimensional feature space.
2014/2/18 YunTech EE Pattern Recognition
30
Syntactic Approach
Use hierarchical structures to represent
complex patterns.
The simplest unit is called: primitives
Complex pattern is represented in terms of
the interrelationships (grammars) between
the primitives.
Grammatical rules can be learned by
training.
2014/2/18 YunTech EE Pattern Recognition
31
Issues of concern
Can be used in situations where the
patterns have a definite structure such as
EKG waveforms, shape analysis of contours.
However, it’s usually difficult to segment
noisy patterns and infer grammar from the
training set.
May yield a combinatorial explosions of
possibilities to be investigated.
2014/2/18 YunTech EE Pattern Recognition
32
Neural Networks
Massively parallel computing systems
consisting of an extremely large number of
simple processors with many interconnections.
Can learn complex non-linear input-output
relationships.
Feed-forward networks such as multilayer
perceptron and Radial Basis Function network
are useful for pattern classification.
2014/2/18 YunTech EE Pattern Recognition
33
Pattern Recognition Models
Approach Representation Recognition
Function
Typical
Criterion
Template
Matching
Samples, pixels,
curves
Correlation,
distance
measure
Classification
error
Statistical Features Discriminant
function
Classification
error
Syntactic or
structural
Primitives Rules, grammar Acceptance
error
Neural Nets Samples, pixels,
features
Network
function
Mean square
error
2014/2/18 YunTech EE Pattern Recognition
34
Reference
A.K. Jain, R.P.W. Duin and J. Mao,
“Statistical Pattern Recognition: A Review”,
IEEE Transactions on Pattern Analysis and
Machine Intelligence (PAMI), Vol. 22, No. 1,
pp. 4-37, Jan. 2000.
2014/2/18 YunTech EE Pattern Recognition
35
Homework assignment
Mathematical functions in Appendix A of
the text book
Notation
Linear algebra
Lagrange optimization
Probability theory
2014/2/18 YunTech EE Pattern Recognition