Course Name Pattern Recognition
English Name Pattern recognition
Pre-Course probability statistics,better if you are familiar with information theory, optimization
Course Code Hours 40 Credit
Teacher Wang Yunhong Course Character Time spring
Applied Discipline Pattern recognition and Intelligent systems, Control theory and Control engineering,Computer science and technology
Teaching Material 《Pattern Recognition》Second Edition,Zhaoqi Bian,Xuegong Zhang et al. Tsinghua University Press,2000.
Reference Material 《Pattern recognition》Second Edition,Richard O.Duda et al.
Course Objective

By taking and finishing this course, students are supposed to master the elementary concepts, basic principles, fundamental analysis methods and algorithms of pattern recognition and acquire the ability of designing and implementing the simple classification algorithms in pattern recognition, which can lay a foundation for their future study and research in this field.

Content Outlines and Time Allocation:
Chapter 1 Introduction (2 Hours)
Introduce fundamental conception of pattern ,pattern recognition and pattern recognition system;Some basic problems about pattern recognition.

Chapter 2 Bayesian Decision Theory (4 Hours)
some staple decision rules(Minimum error rate,Minimum risk,Limited class of minimum error rate conditions makes another kind of error rate ,Min max decision、Sequential classification method etc.);Statistical decision under the condition of gaussian distribution;Classifier error rate

Chapter 3 Estimation of Probability Density Function (4 Hours)
Basic concept of parametric estimation(maximum likelihood estimation,Bayesian estimation);Supervised parametric estimation of gaussian distribution;Unsupervised parameter estimation;Nonparametric estimation of overall distribution(basic methods,Parzen Window Method,KN- nearest neighbor estimation);Estimation of classifier error rate.

Chapter 4 Linear Discriminant Function; (3 Hours)
basic concept of linear discriminant function;Qeneralized linear discriminant function;Fisher linear discriminant;Perception criterion function; The minimum misclassification sampling rule;Minimum mean square error criterion;Linear discriminant criterion function for random minimum error rate; Basic concept of multi class problem.

Chapter 5 Non-linear Discriminant Function (3 Hours)
The concept of piecewise linear discriminant function;Representation of piecewise linear discriminant functions (using union of concave functions); Piecewise linear classifier using the sample design of the intersection;Brief introduction to two discriminant function.

Chapter 6 Neighbor Method (3 Hours)
Nearest neighbor method (decision rule, error rate analysis);K-nearest neighbor method; The improved method of nearest neighbor method (fast algorithm, clipping method, compression method); The nearest neighbor method for the refusal strategy;Best distance nearest neighbor method.

Chapter 7 Empirical Risk Minimization and Structural Risk Minimization (3 Hours)
Concept of average risk minimization and empirical risk minimization;Estimation of linear boundary weight vector; Properties of growth function;Estimation of deviation of empirical optimal decision rules;Structural risk minimization.

Chapter 8 Feature Selection and Extraction (8 Hours)
Basic concepts of feature selection and extraction;Separability criterion(distance in classes and between classes, criterion based on probability distribution, criterion based on entropy function);Feature extraction method (Euclidean distance, probability distance, divergence criterion function, discriminant entropy minimization);Feature selection method (optimal search, sub optimal search, simulated annealing algorithm, tabu search algorithm,Genetic algorithm); Feature extraction based on K-L expansion.

Chapter 9 Non-supervised Learning Method (3 Hours)
Separation method of single peak subset;Indirect method of category separation;Hierarchical clustering method.

Chapter 10 Fuzzy Pattern Recognition Method (3 Hours)
Basic concepts of fuzzy sets;The concept of fuzzy feature and fuzzy classification;The characteristics of fuzzy evaluation Fuzzy clustering method;Fuzzy k- nearest neighbor classifier.

Chapter 11 Statistical Learning Theory and Support Vector Machine(4 Hours)
Brief introduction to statistical learning theory;The basic principle and Realization of the support vector machine algorithm.

Main reference books

《Pattern Recognition》Second Edition,Zhaoqi Bian,Xuegong Zhang et al. Tsinghua University Press,2000.
《Pattern recognition》Second Edition,Richard O.Duda et al.Mechanical Industry Press,2003.

This course mainly introduces the basic theory of image processing, digital image transformation, image enhancement, digital image restoration, the basic concept of digital image coding, digital image analysis, the basic method and the realization of the important algorithms. Through learning of this course, students should master the basic concept of the digital image processing, the basic method and the realization of the important algorithms of the transformation of digital image, digital image enhancement, digital image restoration, image coding and digital image analysis and so on, which can lay a foundation for computer graphics, multimedia processing, pattern recognition and other engineering application and subsequent courses .

Chapter 1 Overview of Image Processing (2 Hours, requirement to understand)
1.1 Concept of Image Processing and Main Contents of Digital Image Processing(1 hour)
1.2 Image Processing and Image Processing System(1 Hour)

Chapter 2 Digital Transformation in Image Processing(2 Hours,requirement to master)
2.1 Point Operation and Histogram (1 Hour)
2.2 Algebraic and Geometric Operations(1 Hour)

Chapter 3 Orthogonal Transformation in Image Processing(4 Hours,requirement to master)
3.1 The Basic Concepts and Properties of Fourier Transformation (1 Hour)
3.2 Fast Discrete Fourier Transform Algorithm (1 Hour)
3.3 Discrete Cosine Transform (2 Hours)

Chapter 4 Image Enhancement and Recovery (4 Hours, requirement to master)
4.1 The Histogram Modification of Image and Image Smoothing(1 Hour)
4.2 Contrast Enhancement (Gray Scale Transformation Method,Local Statistical Method )(1 Hour)
4.3 Inverse Filtering Restoration and Wiener Filtering Restoration (1 Hour)
4.4 Median Filtering and Restoration of Motion-blurred Image (1 Hour)

Chapter 5 Image Coding(4 Hours,requirement to master)
5.1 Image Coding Standards and International Standards(1 Hour)
5.2 Statistical Coding (1 Hour)
5.3 Predictive Coding (1 Hour)
5.4 Transform Coding (1 Hour)

Chapter 6 Image Analysis (2 Hours,requirement to master)
6.1 Image Segmentation(1 Hour)
6.2 Classification and Estimation(1 Hour)