CS4770 Pattern Recognition 3-0-2-4
Professor: P. J. Narayanan
Monsoon 2006
Room 305
Timing: TueFr 8:30-10:00
Course outline:
Pattern Recognition and Pattern Classification deal with the tools
that discover structure in the data about the real world. They are
used to automatically classify a physical objects based on abstract
multidimensional patterns. Pattern recognition developed primarily
with image processing and machine vision, but is an area completely
independent of it. PR techniques find applications in text
categorization, speech recognition, data minining, DNA sequence
identification, etc.
In this course, we will learn the fundamental techniques and tools
used in pattern recognition such as Bayes classifier, Linear
discriminant functions, Neural networks, Hidden-Markov models,
clustering, etc.
Prerequisites:
Mathematical fundamentals. (Matrix and Linear Algebra, Probability
and Statistics, Calculus)
Programming. (For project, assignments)
Desire to work hard.
Teaching Assistant:
Uday Kumar Visesh
Course page
Assignments:
Lecture material:
Tentative Grading Plan:
| Two Tests: | 30-40% |
| Final Exam: | 25-30% |
| Project: | 20-30% |
| 5-6 Assignments: | 10-15% |
Others: | 5-10% |
Approximate Syllabus outline:
- Introduction, Feature Extraction, classification.
- Mathematical preliminaries
- Bayesian Classifier
- Linear Discriminant Functions
- Feature Selection
- Neural Networks
- Decision Trees (other non-metric methods)
- Hidden Markov Models
- Clustering
Textbook:
Pattern Classifiction. Richard Duda, Peter Hart, and David Stork.
John Wiley, 2001. Indian Edition available.