CSU44053 – Computer Vision

Module CodeCSU44053
Module Name Computer Vision
ECTS Weighting [1]5 ECTS
Semester TaughtSemester 1
Module Coordinator/s  Dr. Kenneth Dawson Howe

Presentation for CSU44053

Module Learning Outcomes

On successful completion of this module, students will be able to:

  1. Design solutions to real-world problems using computer vision;
  2. Develop working computer vision systems using C++;
  3. Critically appraise computer vision techniques;
  4. Explain, compare and contrast computer vision techniques.

Module Content

The aim of this module is to give students a firm understanding of the theory underlying the processing and interpretation of visual information and the ability to apply that understanding to ubiquitous computing and entertainment related problems. It provides them with an opportunity to apply their problem-solving skills to an area which, while it is firmly part of computer science/engineering, draws strongly from other disciplines (physics, optics, psychology).

The module is based around problems so that the technology is always presented in context and during some tutorials students work in groups to design solutions to real world problems using the techniques that they have been taught. In addition, the module has a significant practical component so that students can appreciate how difficult it can be to apply the technology.


Specific topics addressed in this module include:

  • Image digitisation and colour;
  • Camera modelling;
  • Binary image processing;
  • Region based processing including connected components analysis, watershed segmentation and mean shift segmentation;
  • Video analysis;
  • Geometric image transforms;
  • Edge based processing including edge detection, contour extraction and representation;
  • Feature processing including basic corner detection techniques and SIFT;
  • Recognition techniques including template matching, statistical pattern recognition, and the Hough transform;
  • Deep Learning in Computer Vision, an introduction;
  • Topics will change somewhat from year to year.

Teaching and Learning Methods

We are restricted (by College) to two classes per week, which would be insufficient to deliver the material and to discuss it and practice its application.  As a result the Computer Vision module is adopting a ‘Flipped Classroom’ where students are expected to study the relevant material in advance of class each week, and classes will be used to deal with questions, introduce some more advanced/applied material and work through solving real world problems.

In a typical week we have to in person classes:

  1. A Q&A discussion and advanced topics sessions: Questions about the lecture material just covered will be discussed with the whole class. Time permitting we will also look at more advanced topics or a tutorial problem. The main lecture material is provided online with slides, prerecorded lectures, and access to various textbooks and sites. Students must study the material before the relevant Q&A sessions.
  2. Tutorial sessions. In the tutorials, application problems will be presented to the class and groups of (3-4) students will try to come up with a solution (based on the techniques covered so far in the module). After around 10 minutes, different groups will present their solutions which we will discuss(in terms of issues/appropriateness). We typically deal with 2 problems in a tutorial.

Students are also asked to solve real problems in the assignments, so that the difficulties of applying the technology can be better understood.

Assessment Details

Assessment ComponentBrief Description Learning Outcomes Addressed% of TotalWeek SetWeek Due
Examination2 hour in-person examination where students must answer 2 out of 3 questions.LO1, LO3, LO450%N/AN/A
Mini-testsSmall quizzes to be completed after studying the prerecorded lecture session in an area. Students must submit a meaningful attempt at least 7 out of the 14 mini-tests in this module.LO3, LO410%Week 1-11Week 1-11
Platform
Familiarity
Assignment where students are asked to do certain tasks using Open CV to provide familiarity with the platform.L020%Week 1Week 3
Development
Assignment
Computer Vision Problem Solving Assignment, including design, implementation, evaluation and report writing. Students must submit a meaningful attempt at this assignment.LO1, LO2, LO3, LO430%Week 4Week 8/9
Problem
Solving
Assignment
Computer Vision Sample Exam Question Assignment, to give practice (and feedback) on answering exam style questions.LO1, LO3, LO410%Week 9Week 10/11

Reassessment Details

Examination (3 hours, 100%).

This exam will be an in-person examination. The supplemental mark in this module is based only on the written supplemental examination. This examination has one mandatory question which draws on the coursework and/or other material. In addition, similar to the annual written examination, students must answer 2 of the other 3 questions.

Contact Hours and Indicative Student Workload

Contact Hours (scheduled hours per student over full module), broken down by: 22 hours
Lectures (mainly Q&A and presentation of advanced topics)11 hours
Laboratory0 hours
Tutorial or seminar11 hours
Other0 hours
Independent Study (outside scheduled contact hours), broken down by:100 hours
Preparation for classes/exams and review of material50 hours
Completion of assessments (including examination, if applicable)50 hours
Total Hours122 hours

Recommended Reading List

  • A Practical Introduction to Computer Vision with OpenCV, by Kenneth Dawson-Howe, Wiley, May 2014.
  • Image Processing, Analysis and Machine Vision by Milan Sonka, Vaclav Hlavac & Roger Boyle, Thompson, Third Edition 2008.

Module Pre-requisites

Prerequisite modules: N/A

Other/alternative non-module prerequisites: Competence in C++. Competence in advanced mathematics.

Module Co-requisites

N/A

Module Website

Blackboard

Link to classes for the first two weeks (for students who may wish to switch to this module)