CS 4476

Computer Vision

(Spring 2025)

Course Description

This is an introductory course to computer vision which strikes the balance of classic vision topics such as light and color, image filtering, feature detection, transforms, stereo vision etc., and modern vision topics such as deep learning based visual recognition, generative models, transformers and diffusion models. The focus of the course is to develop intuitions and mathematics of the methods in the lectures, and then to learn about the difference between theory and practice in the projects.

Lectures

Schedule is tentative and subject to change.

Links to recent/next lecture slides PDF and lecture recordings are shared through Canvas.

Textbook is not required; however, for those seeking an accessible textbook to accompany the course, you can check out the resources here. (Szeliski 2nd ed)

We use Ed Dicussion for more direct online communication.

Assignments, Exams and Grading

Four Assignments with 15 pts each, total 60 pts.
- Assignments are released through Canvas, and submitted through GradeScope.
- You have four "late days" for the whole course without penalty. You will lose 10% each day for late projects, and late submissions will not be accepted after 1 week.
- There are extra credit points for several projects and the total score for those can be up to 110% of 15pts.
- We'll likely provide two more optional assignments to help your learning, but those will not be graded.
Two in-class Exams with 20 pts each, total 40 pts.
Extra credit for participation can be claimed, up to 3pts.

Final Grades: A: 90%+, B: 80%+, C: 70%+, D: 60%+, F<60%.

Teaching Assistants

TA Office Hours:

2hrs/weekday, see table below.
Extra office hours are possible if the time doesn't work for you, contact your head TAs as needed.

Location:

Tentatively KLAUS 2nd Floor Atrium.

TA Contacts:

Useful Links:

https://catalog.gatech.edu/policies/student-absence-regulations/

Acknowledgement:

Course materials are built on top of contents from great colleagues including but not limited to: Judy Hoffman, James Hays, Svetlana Lazbnik, Fei-Fei Li, Kristen Grauman, Devi Parikh, Derek Hoiem, Bill Freeman, Alexei Effros and many more.

Google Sites

Report abuse