ECMM426 | 2023 | University of Exeter

MODULE TITLE	Computer Vision	CREDIT VALUE	15
MODULE CODE	ECMM426	MODULE CONVENER	Dr Sareh Rowlands (Coordinator)

DURATION: TERM	1	2	3
DURATION: WEEKS	0	15	0

Number of Students Taking Module (anticipated)	75

DESCRIPTION - summary of the module content

How do we recognise objects and people? How can we catch a ball? How do we navigate our way from our desk to the coffee machine, without bumping into each other? These seemingly simple tasks have represented a challenge for AI scientists for decades. Recent developments in computer vision have seen significant improvement in important applications (face detection in cameras, body tracking, and autonomous cars).

This module expects prior knowledge in linear algebra and probability theory (e.g., ECM1416), programming with Python (ECMM1400), object-oriented programming (ECM1414 and ECM1410). Prior knowledge in machine learning is also desirable (such as ECM3420 or ECMM422).

AIMS - intentions of the module

This module will provide you with the fundamentals of computer vision, covering the essential challenges and key algorithms for solving a variety of vision problems. The course will provide both theoretical grounding in the relevant theories and a blend of classical and state-of-the-art approaches to computer vision problems. The course will focus on practical applications of computer vision and cover a broad range of problems, from low-level image processing to object recognition, tracking and 3D vision.

INTENDED LEARNING OUTCOMES (ILOs) (see assessment section below for how ILOs will be assessed)

On successful completion of this module you should be able to:

Module Specific Skills and Knowledge

1. Explain key computer vision problems and their mathematical formulation.
2. Design and implement vision algorithms in a high-level language.

Discipline Specific Skills and Knowledge

3. Analyse and propose solutions for computer vision problems.
4. Select appropriate statistical representations, features and algorithms to suit problem specificities.

Personal and Key Transferable / Employment Skills and Knowledge

5. Understand and appreciate the limitations of the state-of-the-art.
6. Critically read and report on research papers.

SYLLABUS PLAN - summary of the structure and academic content of the module

The course will cover the following topics:

Image formation: geometry, light, and cameras

Image processing: convolution, linear filters, Fourier transforms, image gradients, geometric transformations

Feature extraction & matching: corners, edges, blobs, and lines; feature descriptors (SIFT), feature matching and tracking

Object detection and recognition: K-NN, bag-of-words, scanning windows & Viola-Jones

Image segmentation: active contours, Markov random fields, graph cuts

Dense image correspondences: dense motion estimation, optical flow, stereo

Shape reconstruction: 2D and 3D shape modelling and fitting, active appearance models, 3D morphable models

3D vision: 3D pose estimation, calibration, structure from motion, SLAM, shape from shading, motion capture

Deep learning for vision: neural networks, convolutional neural networks, object detection, semantic and instance segmentation, recurrent neural networks

LEARNING AND TEACHING

LEARNING ACTIVITIES AND TEACHING METHODS (given in hours of study time)

Scheduled Learning & Teaching Activities	33	Guided Independent Study	117	Placement / Study Abroad	0

DETAILS OF LEARNING ACTIVITIES AND TEACHING METHODS

Category	Hours of study time	Description
Scheduled Learning & Teaching activities	22	Lectures
Scheduled Learning & Teaching activities	11	Workshops/tutorials
Guided independent study	48	Coursework preparation
Guided independent study	69	Wider reading and self study

ASSESSMENT

FORMATIVE ASSESSMENT - for feedback and development purposes; does not count towards module grade

SUMMATIVE ASSESSMENT (% of credit)

Coursework	60	Written Exams	40	Practical Exams	0

DETAILS OF SUMMATIVE ASSESSMENT

Form of Assessment	% of Credit	Size of Assessment (e.g. duration/length)	ILOs Assessed	Feedback Method
Coursework: workshop code	60	48 hours, code submission	All	Written feedback and model answers
Quiz	40	2 hours	1,3,4,5,6	Written feedback via ELE

DETAILS OF RE-ASSESSMENT (where required by referral or deferral)

Original Form of Assessment	Form of Re-assessment	ILOs Re-assessed
Coursework: workshop code	Coursework: workshop code	All
Quiz	Quiz (2 hours)	1, 3, 4, 5, 6

RE-ASSESSMENT NOTES

Reassessment will be by coursework/quiz in the failed or deferred element only. For referred candidates, the module mark will be capped at 50%. For deferred candidates, the module mark will be uncapped.

RESOURCES

INDICATIVE LEARNING RESOURCES - The following list is offered as an indication of the type & level of
information that you are expected to consult. Further guidance will be provided by the Module Convener

Basic reading:

ELE: vle.exeter.ac.uk

Web based and Electronic Resources:

CVOnline: an online compendium of computer vision techniques: http://homepages.inf.ed.ac.uk/rbf/CVonline/

Website of R. Szelinski’s Computer Vision book (including a free electronic version of the book): http://szeliski.org/Book

I. Goodfellow, Y Bengio & A. Courville's Deep Learning (free chapters): https://deeplearningbook.org

Reading list for this module:

Type	Author	Title	Edition	Publisher	Year
Set	Bishop, John C.	Pattern recognition and machine learning		Springer	2006
Set	Forsyth, David & Jean Ponce	Computer vision: a modern approach	2nd	Pearson	2011
Set	Szeliski, Richard	Computer vision: algorithms and applications	2nd	Springer	2021

CREDIT VALUE	15	ECTS VALUE	7.5

PRE-REQUISITE MODULES	None
CO-REQUISITE MODULES	None

NQF LEVEL (FHEQ)	7	AVAILABLE AS DISTANCE LEARNING	No
ORIGIN DATE	Tuesday 10th July 2018	LAST REVISION DATE	Friday 9th December 2022

KEY WORDS SEARCH	Computer vision, object recognition and detection, semantic and instance segmentation, tracking, pattern recognition, deep learning.

Computer Vision - 2023 entry