This course provides an introduction to computer vision and computational photography. The course will cover basic principles of image processing, image recognition using both classical methods and deep learning, and multiple view geometry for visual navigation. It will explore the topics of image formation, image features, image stitching, image and video segmentation, motion estimation, tracking, and object and scene recognition.
The course is organized around several projects. Through these projects you will learn the theory and practical skills required to obtain a computer vision engineering job.
Links to the lectures can be found on ELMS .
| Date | Topic | Assigned Reading |
|---|---|---|
| 01/25 | Introduction to Computer Vision | David Jacobs' Preassessment Notes |
| 01/27 | Linear Algebra / Least Squares (Normal Equations) |
|
| 2/1 | Ridge Regression and Regularization | |
| 2/3 | Singular Value Decomposition (SVD) Principal Component Analysis (PCA) |
PCA and SVD with numpy tutorial |
| 2/8 | Filtering | |
| 2/10 | Canny Edge Detection | Szeliski textbook, Section 3.2 |
| 2/15 | Image Pyramids and Frequency Domain | Szeliski textbook, Sections 3.4 and 3.5 |
| 2/17 | Frequency Domain Continued, Intro to Corner Detection | Szeliski textbook, Section 7.1 |
| 2/22 | Corner Detection | |
| 2/24 | Feature Descriptors and SIFT | |
| 3/1 | Image Classification Bag of Words |
Szeliski textbook, Section 6.2.1 |
| 3/3 | Image Classification Support Vector Machines (SVM) |
Szeliski textbook, Section 5.1.4 |
| 3/8 | Neural Networks | |
| 3/10 | Neural Networks | |
| 3/15 | Spring Break | |
| 3/17 | Spring Break | |
| 3/22 | Convolutional Neural Networks | |
| 3/24 | Convolutional Neural Networks | |
| 3/29 | 2D Transformations Projective Coordinates |
Szeliski textbook, Section 3.6 Cyrill Stachniss Lecture |
| 3/31 | 2D Transformations Projective Coordinates |
|
| 4/5 | Homographies RANSAC |
|
| 4/7 | Geometric Camera Models | David Jacobs' 3D Geometry Notes |
| 4/12 | Geometric Camera Models | |
| 4/14 | Two-view Geometry | |
| 4/19 | Stereo and Structured Illumination | |
| 4/21 | Structure from Motion | |
| 4/26 | Segmentation | |
| 4/28 | Optical Flow | |
| 5/3 | Digital Photography | |
| 5/5 | Computational Photography | |
| 5/10 | Recent Trends in Computer Vision | |
| 5/19 | Final Project Due |
Instructor: Christopher Metzler (metzler at umd.edu)
Office: Online
Office Hours: Fridays 11:00-noon or by appointment
| Name | Office hours | |
|---|---|---|
| Kanishka Ganguly | kganguly at terpmail.umd.edu | Tuesdays and Thursday 5:00-6:00 |
| Jiaye Wu | jiayewuw at gmail.com | Mondays 3:15-5:15 |
| Shantam Bajpai | sbajpai at terpmail.umd.edu | Thursdays 2:30-4:30 |
All office hours take place online. Zoom links can be found on ELMS.
Click the name of an assignment below to see its specifications. Homework should be turned in via ELMS. Assignments are due before class time.
| Homework 1: Vision Startups | Feb. 3, 2021 |
| Homework 2: Least Squares, Ridge Regression, and PCA | Feb. 17, 2021 |
| Project 1: Filtering and Edge Detection | March 3, 2021 |
| Project 2: Bag of Words Classification | March 24, 2021 |
| Project 3: Classification with Neural Networks | April 14, 2021 |
| Project 4: Homographies and RANSAC | April 28, 2021 |
| Homework 3: 3D Geometry | May 5, 2021 (optional) |
| Final Project: Visual Odometry | May 19, 2021 |