Instructor: Jean Gao

Email: gao@uta.edu

Office: 538 Engineering Research Building, phone: 817-272-3628.

Office Hours: Mondays, Wednesdays, 12:00 - 1:00pm or by appointment.

Course Information: Tuesdays and Thursdays, 12:25 - 1:45pm.

Lecture Room: ERB 129

Course Description:

Pattern recognition - the act of taking raw data and making decisions based on the categories of the pattern - has applied to such diverse areas as character recognition, data mining, medical diagnosis, image processing, computer vision, bioinformatics, speech recognition, fraud detection, and stock market prediction. This course will provide underlyingof principles and various approaches of pattern recognition and decision making processes. The topics include diverse classifier designs, evaluation of classifiability, learning algorithms, and feature extraction and modeling. The goal of this course is to introduce students to the fundamental models of decision making in order to prepare them for applying the associated concepts to information processing.

Course Schedule.

Textbook:

## References:

K. Fukunaga, Introduction to Statistical Pattern Recognition, Academic Press, second edition, 1990. (0-12-269851-7). (available in the library)

Sergios Theodoridis and Konstantinos Koutroumbas, Pattern Recognition, Academic Press, 2006. (Third edition)

Evaluation:

Homework (25%): There will be about 5~6 HWs. Some of the homework assignments may include small computer projects. You are free to choose your most comfortable programming language. Homework assignments are collected before class starts.Exams (50%): There will be two midterms. Each will cover approximately 50% of the course materials.

Presentation (15%): Students will give one class presentation on selected topics focusing on the diverse applications of pattern

Class Attendance (10%): Attendance is expected for your best learning.

recognition.

Cooperative efforts at understanding the material and the assignments are encouraged. However, you are required to present your work that you have completed individually.

Prerequisites:

A basic background in probability theory and linear algebra, or consent of instructor.

Tentative Major Topics to Be Covered:

1. Introduction

A. Problems in decision making processes

B. Mathematical formulation2. Pattern recognition and learning machines

A. Review of probability theory and linear algebra

B. Hypothesis test (Bayesian test)

C. Parametric classifier design (LDA, QDA)

D. Non-parametric classifier design (KNN, Parzen window)

E. Estimation of classifiability

F. Classifier evaluation

G. Learning algorithms3. Data analysis

A. Feature extraction for signal representation (PCA, ICA)

B. Feature extraction for classification (FSS)

C. Clustering

D. Modeling and validity tests