| Category | Difficulty |
|---|---|
| HW | 2 |
| Exam | 6 |
Pattern Recognition Theory is offered in the fall semester. It is an graduate level course. You need some understanding of probability theory and linear algebra course to follow along. The following topics are covered in the course- Decision theory, parameter estimation, density estimation, non-parametric techniques, supervised learning, linear discriminant functions, clustering, unsupervised learning, artificial neural networks, feature extraction, support vector machines, and pattern recognition applications (e.g., face recognition, fingerprint recognition, automatic target recognition, etc.). It is not as hard as other computer vision related courses at CMU.
Lecture was once a week so you have one class of four hours a week. The class was not fast paced so that the basic concepts can be covered well. Having a good background in probability theory and linear algebra really helps. Towards the end of the semester more topics were covered every lecture so it may feel rushed.
- You have four homeworks. There is a coding assignment as well as a written part. It's not too hard and you can use OH if you have any doubts about the homeworks.
- There was one midterm exam which was harder than the topics covered in class.
- There is also a group project which goes on the entire semester.
- Attend and understand lectures. You will have access to the lecture recordings in case you cannot sit for four hours.
- It is an easy course and there won't be too much work involved except for the project.
- The project is a major part of the course so make sure you do it well.
- Overall the midterm is quite different from the homeworks so you need to prepare properly for it.