Computer Vision
Lecture notes
Suggested schedule for lecture recordings and study:
21 January: Lecture 1 (slides 1 - 20)
Overview, and goals of computer vision
26 January: Lecture 2 (slides 21 - 38)
Pixels, sensors, and image coding (both wet and dry)
28 January: Lecture 3 (slides 39 - 55)
Neural operations on images
- week of 25 Jan 2021: Exercises 1 - 3.
2 February: Lecture 4 (slides 56 - 62)
Mathematical operations on images
4 February: Lecture 5 (slides 63 - 70)
Edge detection and its challenges
- week of 1 Feb 2021: Exercises 4 - 6.
9 February: Lecture 6 (slides 71 - 78)
Isotropic, anisotropic, and nonlinear operators
11 February: Lecture 7 (slides 79 - 88)
Multi-scale analysis and wavelets for visual coding
- week of 8 Feb 2021: Exercises 7 - 10.
16 February: Lecture 8 (slides 89 - 98)
Active contours, boundary descriptors, and SIFT
18 February: Lecture 9 (slides 99 - 112)
Functional streams, texture, and colour processing
- week of 15 Feb 2021: Exercises 11 - 15.
23 February: Lecture 10 (slides 113 - 125)
Stereo vision, motion, and optical flow
25 February: Lecture 11 (slides 126 - 135)
Surfaces and reflectance maps
- week of 22 Feb 2021: Exercises 16 - 19.
2 March: Lecture 12 (slides 136 - 149)
Shape representations, codon grammars, and vision as modelling
4 March: Lecture 13 (slides 150 - 160)
Bayesian inference and statistical classifiers
- week of 1 Mar 2021: Exercises 20 - 24.
9 March: Lecture 14 (slides 161 - 172)
Discriminant functions and convolutional neural networks
11 March: Lecture 15 (slides 173 - 192)
Face detection and recognition (2D appearance-based)
- week of 8 Mar 2021: Exercises 25 - 29.
16 March: Lecture 16 (slides 193 - 212)
3D face recognition, affect, and Facial Action Coding System
First Q&A session, covering Lectures 1 - 8 and Exercises 1 - 10
- Friday 19 February, 4:00 - 5:00pm, by Zoom call (per invitation)
Second Q&A session, covering Lectures 9 - 15 and Exercises 11 - 29
- Friday 12 March, 4:00 - 5:00pm, by Zoom call (per invitation)
Some other reference resources
For interest: reference paper on RGB-D cameras and 3D reconstruction: (18 MB pdf).
You may enjoy this collection of
dynamic, colour, and cognitive illusions.
Here is a background paper about
face recognition, and here is a paper about the breakthrough in
face recognition using the "deep learning" approach of
FaceNet.