CS4670/5670 - Computer Vision

Picture credit: xkcd

About this course

Humans are extremely good at perceiving the world from visual input alone. This comes so easily to us that we underestimate how difficult perception it is, and how hard it is for machines, as the webcomic above illustrates.
Computer vision is a subfield of AI focussed on getting machines to see as humans do, and has been around for almost half a century. This course will cover the basics of computer vision: the underlying mechanics of images, the core problems that the field focuses on, and the array of tools and techniques that have been developed. The emphasis will be on covering the fundamentals which underly both computer vision research and applications. A tentative list of topics is below:
  • Geometry / Physics of image formation
  • Properties of images and basic image processing
  • 3D reconstruction
  • Grouping (of image pixels into objects)
  • Machine learning in computer vision: basics, hand-designed feature vectors, convolutional networks
  • Detecting and localizing objects
A detailed but tentative list of learning outcomes can be found below. This course is intended for undergraduate students and MEng. students. Knowledge of basic probability and linear algebra will be useful. A primer on the aspects of linear algebra that will be useful is available here.

Quick info

Instructor: Bharath Hariharan
Lecture time: MWF 1:25pm - 2:15pm
Lecture venue: Ives 305
  • Stephanie Chang
  • Lindsay Fei
  • Apoorv Khandelwal
  • Jiwon Kim
  • Frank Li
  • Mason Liu
  • Alisha Mithal
  • Riley Niu
  • Kane Tian
  • Albert Tsao
  • Ziyang Wu
  • Xiaokai (Steven) Ye
Instructor Office Hours: M / F 2:30 - 3:30 pm in Gates 311
All Office Hours:

Plan for going online

Lectures: We will go with primarily asynchronous lectures, i.e., recorded videos. For people with not very good internet connections, I will attempt to post a detailed ``script'' as a pdf. The remote lectures will be linked from the calendar and will also be added to this channel.

Office Hours: We will have remote office hours. Instead of the usual 1 hour office hour on two different days, I will instead have three 1-hour office hours spread over different days and different times of the day:
Mondays , Fridays 2:30 - 3:30 pm EST (Zoom)
Wednesdays, 9 - 10 am EST. (Zoom)

(If you can't attend office hours on zoom, we will all still be looking at Piazza)

(Ungraded) Quizzes: One challenge with remote and recorded lectures is that it can be hard to stay engaged. So every week we will have a short ungraded quiz on cms you can attempt to make sure you are following.

Learning outcomes Calendar and lecture notes Resources Grading policy