|
|
|||||||||||||||||
Course Catalog 2011-2012
SGN-2506 Introduction to Pattern Recognition, 4 cr |
Additional information
Lectures and exercises in English.
Person responsible
Jussi Tohka, Ulla Ruotsalainen
Lessons
| Study type | P1 | P2 | P3 | P4 | Summer | Implementations | Lecture times and places |
|
|
|
|
|
|
|
|
|
| Study type | Hours | Time span | Implementations | Lecture times and places |
|
|
|
| SGN-2506 2011-02 |
|
Requirements
Final examination and active participation in exercises.
Completion parts must belong to the same implementation
Principles and baselines related to teaching and learning
-
Learning outcomes
After completing the course, the student will know the basic structure of pattern recognition systems and the statistical bases of the classification theory (the Bayes classifier). He will distinguish supervised learning methods from the unsupervised ones. He will be able to apply supervised learning methods (model-based maximum likelihood, k-nearest neighbours) to the classifier design. The student will be able to apply k-means clustering algorithm.
Content
| Content | Core content | Complementary knowledge | Specialist knowledge |
| 1. | The basic structure of pattern recognition systems. Supervised and unsupervised learning. | The design cycle of pattern recognition systems. | |
| 2. | Basics of multivariate probability and statistics, class conditional density function, Bayesian decision theory, Bayes classifier | The Bayes minimum risk classifier | |
| 3. | Parametric (model-based maximum likelihood) and nonparametric techniques (Parzen windows, k-nearest neighbours) for the estimation of density functions and the design of pattern classifiers based on training data. | ||
| 4. | Linear classifier, Perceptron algorithm | Minimum squared error method | |
| 5. | Testing of pattern recognition systems. | ||
| 6. | Algorithms for unsupervised classification. K-means clustering. | EM-algorithm |
Evaluation criteria for the course
In order to pass the course the student has to pass the exam and make at least 30% of the exercises. There will be bonus from extra exercises. To pass the exam at least half of the maximum points of the exam has to be reached. Lecture notes and exercises are enough to good grade in exam.
Assessment scale:
Numerical evaluation scale (1-5) will be used on the course
Partial passing:
Study material
| Type | Name | Author | ISBN | URL | Edition, availability, ... | Examination material | Language |
| Book | "Pattern Classification" | Duda RO, Hart PE, Stork DG | 2nd edition, Wiley, 2001 | English | |||
| Summary of lectures | "Introduction to Pattern Recognition" | Jussi Tohka | English |
Additional information about prerequisites
Basics of signal processing and probability
Prerequisite relations (Requires logging in to POP)
Correspondence of content
| Course | Corresponds course | Description |
|
|
|
|
|
|
|
More precise information per implementation
| Implementation | Description | Methods of instruction | Implementation |
| Teaching: 32 hours lectures (non-mandatory), 16 hours exercises (30% mandatory). No particular exercise session is mandatory although in some of them it might be stated so in POP. The contents and requirements are the same as in the previous implementation of SGN-2506 in Fall 2011. Detailed information for the summer implementation is available on the course web page: http://www.cs.tut.fi/kurssit/SGN-2506/ |