Pattern Recognition: Theory and Applications

by Dr. Kenneth K.M. Lam
Department of Electronic and Information Engineering
The Hong Kong Polytechnic University


Suject Code: EIE 522

Objective:

This course offers an up-to-date review of the state of the art in pattern recognition. In particular, it outlines the need for pattern recognition, its different algorithms, decision theoretic, syntactic, and neural network approaches including learning algorithms, and different classical image processing and character recognition techniques. The course will emphasize practical techniques for implementing useful pattern recognition systems. It will also provide a base for practice and progress in matters related to research.

 

 

 

 

Syllabus:
  1. Introduction
    • The Subproblems of Pattern Recognition
    • Patterns and Pattern Vectors
  2. Segmentation and Features
    • The concept of Texture
    • Edge-Detection Methods
    • Shape Characterization
    • Coefficient Expansions
  3. Statistical Approaches to Pattern Recognition
    • Approaches to developing StatPR classifier
    • Supervised Learning using Parametric & Nonparametric Approaches
    • Unsupervised Learning and Clustering
    • Linear Discriminant Functions
    • Principal Component Analysis and Fisher's Linear Discriminant
 
Notes:
  1. Introduction to Pattern Recognition
  2. Feature Extraction (curvature, central moments)
  3. Statistical Pattern Recognition
  4. Parametric and Non-parametric Supervised Training (Example: Page37-38)
  5. Unsupervised Training (Proof) (Supplement)
  6. Component Analysis and Discriminant
 
Announcement:

Because of the poor conditions of the classroom W210, we will move to the following classrooms for our lectures.
Y409 (19 and 26 September)
TU201 (3, 10, 17, and 31 October)

Y302 (24 October; 7, 14, 21, 28 November; 5 December)

Test 1:
Date: 26 September 2006
Time: 7:00pm - 8:00pm

Test 2:
Date: 17 October 2006
Time: 7:00pm - 8:00pm

Supplementary Test 2: If your score in Test 2 is lower than 50, you may seat for this supplementary test. However, you must attend a tutorial so that you are allowed to take this test.
Date: 2 December 2006 (Saturday)
Time: 2:00pm (Tutorial)

Venue: CD634
The tutorial is open to all students. After the tutorial, you will be given the test paper and may complete it at home. You must submit this test paper on or before 4 December 2006.
Please click here to download the Supplementary Test 2.

Please click here to download the solution (**New**).

 

Teaching schedule: Please click here to obtain the updated teaching schedule.

 
Tutorials:
  1. Feature Extraction (solution)
  2. Statistical PR
  3. Supervised Training (solution) *New
  4. Unsupervised Training (solution) *New (Note that the original solution to Q1(b) is correct!)
 
Animation:
Correlation, CMean, LDA
 
Assignment:
  1. Assignment 1 (to be sumbitted by 3 October 2006)
  2. Assignment 2 (to be submitted by 11 November 2006)
 
Laboratory:

Lab 1: Statistical Pattern Recognition
Date: 11 November 2006 (Saturday)
Time: 2:00pm - 5:00pm
Place: CF105 and CF105a
The submission deadline has been extended to 8 December 2006 (Friday). Please note that this is the absolute deadline.

Please click here to download the laboratory materials and "an introduction to MatLab"

 
Consultation Hours:

Every Monday: 8:00pm - 10:00pm
Tutors: Thomas Tse, Wing-Pong Choi, and Hei-Sheung Koo

Additional Consultation Hours:
Date: 2 November 2006 (Thursday)
Time: 8:00pm - 9:30pm
Date: 8 November 2006 (Wednesday)
Time: 8:30pm - 10:00pm
Date: 12 December 2006 (Tuesday)
Time: 7:00pm - 9:00pm, Place: CD634
Date: 14 December 2006 (Thursday)
Time: 7:00pm - 8:30pm, Place: CD634

 
Others:
  1. Greedy Snake Algorithm
  2. Circular Gabor
 
Teaching schedule: Please click here to obtain the teaching schedule.