Course syllabus

Jump to today

Introduction
This course introduces computational methods and techniques for computer vision, extended towards real-world problems such as vision-guided robotics based on 3D scene description. The students will interact with 3D vision equipments based in the city campus Intelligent Vision Systems Lab (www.ivs.auckland.ac.nz). A particular feature of the course work is the emphasis on complete system design. This year the course project will look at the integration of 3D stereo-cameras into a 3D map using a simplified version of SLAM technology. Tentative non-compulsory field demos using Unmanned Aerial Vehicles (quadcopter) will be planned at Leigh Marine Science campus either towards the end of the semester or during Easter break.

Classes will take place on Monday (2-3pm) and Tuesday (2 hours within the 1-4pm block). Practical experiment sessions (2 across the semester) will take place over three hours on Tuesday 2 April and Tuesday 14 May (1-4pm).

Assignment resources: here

Lecture notes will be available in the relevant canvas section

Timetable (Semester 1, 2019)

Day

Time

Where

Lecture

Mon  4 March

2-3pm

303-G11

1) Introduction

Tues 5 March

1-3pm

303-B11

2-3) 2-3D geometry, Image features, camera calibration

Mon 11 March 

2-3pm

303-B11

4) Image features, camera calibration, Camera calibration exercises, Trigonometry

Tues 12 Mar

1-3pm

303-B11

5-6) Bits and pieces about calibration: rotation/translation exercises, Features, Tsai calibration: summary, Tsai's paper, Stereo web camera, Camera distortion, 773 lecture notes on calibration (zip)

Mon 18 Mar

2-3pm

303-B11

7) Camera calibration wrap-up - Literature review and experimental protocol design

Tues 19 Mar

1-3pm

303-B11

8-9) Experimental procedure, literature review discussion, Presentation discussion, preparation to the assignment(A2).

Mon Mar 25

2-3pm

303-B11

10)Camera calibration

Tues Mar 26

1-4pm

303-B11

11-12) Camera calibration (exercises/wrap-up), A2 preparation

Mon 1 April

2-3pm

303-B11

13) Introduction to 3D Vision

Tues 2 April

1-3pm

303-B11

14-15) In-class assignment (A2)

Mon 8 April

2-3pm

303-B11

16) Stereo Geometry 

Tues 9 April

1-3pm

303-B11

17-18) Stereo Matching+Stereo Rectification

Mid-semester break: April 13 - April 27

Mon 29 April

2-3pm

303-B11

19) Auto-Calibration + Introduction to Multi-Image 3D Reconstruction

Tues April 30

1-3pm

303-B11

20-21) Feature Detection + Visual Odometry

Mon 6 May

2-3pm

303-B11

22) Solving Optimization Problems + Bundle Adjustment

Tues 7 May

1-3pm

303-B11

23-24) Point Clouds 1 + 2

Mon 13 May

2-3pm

303-B11

25) Preparation to assignment A3

Tues 14 May

1-4pm

303-B11

26-28) In-class assignment (A3)

Mon 20 May

2-3pm

303-B11

29) Machine Learning in Computer Vision

Tues 21 May

1-3pm

303-B11

30-31) ANN + Deep Learning with CNN

Mon 27 May

2-3pm

303-B11

32) Case Study: Stereo Matching by CNN

Tues 28 May

1-3pm

303-B11

Open class, lecturers available for help

Mon 3 June

2-3pm

303-B11

No class

Tues 4 June

1-4pm

303-B11

Final demo, individual assessment

Top

Lecture notes and additional materials

Image Processing review:

Image Processing And Analysis: A Primer

G Gimel'farb, P. Delmas; Image Processing and Analysis: A Primer, World Scientific Europe, ISBN 978-1-78634-581-3, 2018.

Recommended reading:

  • R. Hartley and A. Zisserman: Multiple View Geometry in Computer Vision. 2nd edition. Cambridge University Press: Cambridge, UK, 2006.
    • Part 0: Projective geometry, 2D/3D transformations.
    • Part I: Camera geometry and single view geometry.
    • Part II: Two-view geometry: epipolar geometry, fundamental matrix, 3D reconstruction.
    • C. M. Bishop: Pattern Recognition and Machine Learning. Springer Science+Business Media, LLC: New York, NY, USA, 2006.
      • Chapter 8: Graphical models: inference, factor graphs, sum-product and max-sum algorithms, dynamic programming, loopy belief propagation.

    Course summary:

    Date Details Due