Course overview

This course offers an introduction to fundamental principles and real-world applications of 2D, 3D, and deep learning-based computer vision. Major topics include image filtering, feature detection and matching, recognition and tracking, scene understanding, camera imaging geometry, stereo vision, and deep learning-based vision. Students will learn to implement interesting computer vision algorithms in a series of well designed projects. Students will also explore intriguing research questions during a final project. This course will be of particular interest to students seeking to delve into fields of image processing and computer vision. 

Prerequisites

This course requires students to have basic knowledge in linear algebra, basic probability, and Python programming. Knowledge in signal processing and machine learning is recommended but not necessary. 

Recommended Materials

Computer Vision: Algorithms and Applications (http://www.szeliski.org/Book/)

Assignments and Grading

* Students will be allowed a total of five late days per course. Each additional late day will incur a 10% penalty.

1. Five homework assignments (60% = 12%*5)

• Convolution and cross-correlation 

• Canny Edge Detector, Hough transform 

• Harris corner detector, RANSAC, HOG descriptor 

• Face detection

• Lukas-Kanade tracking

2. A final project (35%)

• Project proposal presentation (5%)

• Final project blog post (20%) 

• Final presentation (10%)

3. Class participation (5%)