< 

bimagicLab

> 

bioimage informatics lab

 


home \ people \ research \ publications \ teaching \ contact


42-703/18-799: Wavelets and Multiresolution Techniques


The instructor for this course is Jelena Kovačević.

 

Basic course information


All course material will be posted at the course site. Lecture topics can change without notice depending on the students enrolled and their backgrounds.

 

A prerequisite for this course is 18-396 Signals and Systems or equivalent.

 

Course goals


The goal of this course is to expose you to multiresolution signal processing methods and their use.  Upon successful completion of this course, you will be able to:

Explain the importance and use of Hilbert spaces and signal representations in building more sophisticated signal processing tools, such as wavelets.

Think in basic time-frequency terms.

Describe how Fourier theory fits in a bigger picture of signal representations.

Use basic multirate building blocks, such as a two-channel filter bank.

Characterize the discrete wavelet transform and its variations.

Construct a time-frequency decomposition to fit the signal you are given.

Explain how these tools are used in various applications.

Through an independent project, you will learn to recognize techniques covered in the course, evaluate their usefulness in the real world and compare them to other techniques available. You will also develop your own algorithm to solve a practical problem.

 

My stress in this course is on the process of learning. If you strive to understand and apply the concepts you learned in class, you will be successful in it. Asking questions and doing is the best way to learn. There are no stupid questions. You are not in class to impress me but to learn and develop one step closer to being an independent researcher. Asking a lot and early is the way to go. Do not wait for five minutes before homework due time/midterm/project presentation to ask a question because I will not have sufficient time to go into details with you.

 

I will continuously assess how the course is going by using feedback from you. I will give you an anonymous 5-minute questionnaire at the end of each class where I will ask you to rate and give comments on a variety of topics to do with the course: from my preparation and delivery, to the level of difficulty of homeworks and anything else you wish to add. I will use that to answer your concerns and to improve your learning experience. I will not see these forms, someone else will type them for me.

 

Course materials


Notes will be distributed during the class, based on the upcoming textbook by the instructor.

For the more mathematically oriented, I also recommend A Wavelet Tour of Signal Processing, by Stephane Mallat.

Grading policies


Class participation

Active class participation is very important. It will count as much as homework towards the final grade. After each class, I will note your level of participation. Active participation means getting to class prepared, reading the assigned text, doing your homework and getting involved in discussions. You will be expected to read the material indicated on the course site prior to coming to class (except for the first day of class). I praise effort, not necessarily right answers. I do not expect you to have mastered the material before coming to class (this is what we do together), but I do expect you to put in a serious effort to try and master it.

Homeworks

Homeworks will be given and solutions posted at this course site. Students are required to turn in their homeworks on time, by the beginning of the class, on the day the homework is due. Homeworks will count toward the final grade.

Midterm

Midterm will be given in class.

Final

There will be no final exam.

Project

You will be expected to do an independent project.

Grading

Grading is absolute, not on a curve. This means I will grade you based solely on your work and will not compare you to the others in the class. This is done so that you can obtain a grade based on your independent performance and not in competition with others. This also means that everyone can get an A (everyone can get an R as well, but I am hoping you will strive for better).

The final grade will be calculated as follows: 10%: class participation, 10%: homework, 30%: midterm, 50%: project.

 

Policies on cheating


Please make sure you read this section. Often, students caught cheating claim they did not know they were doing anything wrong. That is why I took the time to spell out what is and what is not acceptable. By enrolling in this course, you acknowledge that you have read and understood these rules and will abide by them.

 

I encourage collaboration in this course. That is why I encourage you to do your projects in pairs. I also encourage you to discuss problems in class as well as homework problems. Ultimately, I assume you are taking this course because you want to learn the material and come one step closer to being an independent researcher. Therefore, I will also assume that you will do your best to come to the solutions mostly on your own.

 

Following standard practice in scientific publishing, I will ask that every homework you hand in have an Acknowledgment section. This section should detail your collaboration on that specific homework. Make sure you let me know if you gave help and to whom, received help, discussed the solution or got to the solution together. I will not penalize you for collaborating. I will penalize you though, if you collaborate and do not let me know.

 

Similarly, collaborating with someone else than your partner on your project is fine in the following areas: help with code, help with literature, brainstorming ideas towards a possible solution. Again, as for homework, in the Acknowledgement section, make sure you put down all the people who were involved in the work. This is standard practice in scientific journals.

 

Most of you are coming to class to learn and will not descend to cheating. For a tiny fraction to whom this does not apply: As a rule of thumb, if you are handing in something you do not understand, you probably cheated. If you are allowing someone to copy a solution from you, you cheated. Any collaboration during midterm is strictly prohibited and constitutes cheating.

 

If you are unsure if something is or is not allowed, check with me. If you are at the point where you are considering cheating because you just cannot do it on your own, contact me. We will figure out a way for you to succeed in this course.

 

If I suspect that someone has cheated in class-related work, I will look into it on an individual basis. Penalty will also be based on individual circumstances as well as on your response to the situation. If you inform me on your own that you cheated or are unsure if you did, I will certainly be more lenient than if I discover it myself and have to confront you.

 

Please read the University Policy on cheating.