|
Instructor:
Prof. Sinisa Todorovic sinisa at oregonstate edu 2107 Kelley Engineering Center Classes: Tue 2-4pm, Thu 2-3pm BEXL 207 Office hours: Wed 3:00-3:30pm, or by appointment
|
 |
| LECTURE NOTES: |
| Date | Topics and Recommended Literature | Slides |
| 01/10 Tue 2-4pm |
Introduction: Administrative stuff; What is computer vision; Overview of fundamental problems and popular applications; Big picture -- Image formation; From image features to image understanding; |  |
| 01/12 Thu 2-3pm |
Image filtering; Feature extraction; Matlab warm-up; | |
| 01/17 Tue 2-4pm |
Interest points (RS pp. 181- 199); Keypoint descriptors; | |
| 01/19 Thu 2-3pm |
Matlab practice session |
|
| 01/24 Tue 2-4pm |
Point-based image matching (RS pp. 225-235); The Assignment problem; Hungarian algorithm; Homework 1 (due 02/03) | |
| 01/26 Thu 2-3pm |
Point-based image matching (RS pp. 225-235); The Assignment problem; Linear Program; | |
| 01/31 Tue 2-4pm |
Imaging process; Geometric primitives (RS pp. 29-32); 2D and 3D transformations (RS pp. 33-41); | |
| 02/02 Thu 2-4pm |
Projections (RS pp. 42-47; Forsyth & Ponce pp. 20-37; Hartely & Zisserman Ch. 3); Camera calibration (RS pp. 45-50; Forsyth & Ponce pp. 23-29); Camera calibration toolbox for MATLAB; | |
| 02/07 Tue 2-3pm |
Epipolar geometry (Hartely &
Zisserman Ch. 9); Forsyth & Ponce pp. 260-2);
Preparation for Exam 1 |
|
| 02/09 Thu 2-3pm |
Exam 1 | |
| 02/14 Tue 2-4pm |
MATLAB functions for multiple view geometry; Essential and fundamental matrices (RS pp. 347-354; RANSAC (Forsyth & Ponce pp. 346-351; Hartely & Zisserman Ch. 4.7-4.8); Homework 2 (due 02/27); | |
| 02/16 Thu 2-3pm |
Levenberg-Marquardt
algorithm; Triangulation (RS pp. 345-7); |
|
| 02/21 Tue 2-4pm |
2D Homography (Hartely & Zisserman Ch. 4); Homography estimation; | |
| 02/23 Thu 2-3pm |
Perceptual grouping and Gestalt laws (Forsyth & Ponce pp. 304-309); Color (RS pp. 71-76); Image classification; Image classification using deep learning; | |
| 02/28 Tue 2-4pm |
Deep Learning; Convolutional Neural Network; Autoencoder; | |
| 03/02 Thu 2-3pm |
Error backpropagation in neural networks; Homework 3 (due 03/21) | |
| 03/07 Tue 2-4pm |
Edges (RS pp. 210-226); Shape descriptors; Shape Matching; Dynamic time warping (RS pp. 485-487); | |
| 03/09 Thu 2-3pm |
Image segmentation (RS pp. 237-252); Meanshift; Homework 4 (due 03/23) | |
| 03/14 Tue 2-4pm |
Preparation for Exam 2; |
|
| 03/16 Thu 2-3:20pm |
Exam 2 |