COS 429 - Computer Vision

Fall 2019

Assignments

Assignments

There will be four assignments, each consisting of a written section and a programming section. Each assignment will count for 14% of the final grade.

These are to be completed individually or with a partner. If working with a partner, please submit one writeup together. You may work with a partner on every assignment but note that you may work with the same partner on at most two assignments.

Assignment 1: Image processing and feature detection (due Oct. 3, 11:59 PM)
Assignment 2: Face Detection and Model Fitting (due Oct. 17, 11:59 PM)
Assignment 3: Tracking (due Nov. 21, 11:59 PM)
Assignment 4: Deep Learning (due Dec. 5, 11:59 PM)

To help you get started, there is also

Assignment 0: Setup (due Sep. 19)

as well as some great notebooks with demos courtesy of Prof. David Fouhey (at U Michigan)

Midterm

There will be a midterm worth 20% of the final grade. The midterm will be during class time on Thu, Oct 24. No alternate times will be offered, except in cases of medical emergenies.

Final project

In addition, there will be a final project for the course, to be completed in groups of 1-3. This will be worth 24% of the final grade. No late submissions are accepted for any part of the final project.

Final Project (milestone due Dec. 13, poster session Jan 13 3-6pm in the Friend Convocation Room, report due Dean's Date Jan. 14)

Collaboration Policy

You may complete each assignment individually or with a partner. You may discuss assignments with others, but everything handed in must be the work of you and/or your partner. All code must be written by you and/or your partner -- you may not use code from other classmates, the internet, or any other source.

For the final project, you are expected to collaborate within your group, and to look at other projects and publications for inspiration. However, you must acknowledge any ideas that are not your own, and you may not directly re-use others' code.

Late Policy

Assignments 1 through 4 are due at 11:59 PM on the due date. You are given 4 free late days that you can use any time during the semester, with at most 2 late days per assignment. Once you use up your 4 free late days, you will incur a penalty of 20% for each additional late day.

Note that no assignment will be accepted more than 2 days late.

No late presentations or reports will be permitted for the final project.

Notes on Grading

Since several of the assignments for this course do not involve enormous amounts of coding, a significant part of the grade will depend not just on implementing the algorithms, but also on analyzing how well they perform on a variety of inputs. This will be based on the writeups - they should not only describe the algorithm and any implementation notes, but also include the results you get on a variety of test cases and an explanation of why you might be getting those results.

Submitting Your Work

We will use Gradescope to handle submission of code and writeups. The code for this class is 97KR4W. When creating your account, make sure you use your Princeton email, which is @princeton.edu. Any email that does not follow this format will be removed from the roster.
You can resubmit files as needed up until the submission deadline.
To submit homework as a group of 2, have one of the members submit the work. Then on the page where you view your submission, you can add group members by clicking "Add Group Member" in the top right corner.

Last update 2-Dec-2019 09:56:55