Computer Vision

Code:

M.EIC029

Acronym:

VC

Keywords
Classification	Keyword
OFICIAL	Interaction and Multimedia

Instance: 2024/2025 - 2S

Active?	Yes
Responsible unit:	Department of Informatics Engineering
Course/CS Responsible:	Master in Informatics and Computing Engineering

Cycles of Study/Courses

Acronym	No. of Students	Study Plan	Curricular Years	Credits UCN	Credits ECTS	Contact hours	Total Time
M.EIC	49	Syllabus	1	-	6	39	162

Teaching language

Suitable for English-speaking students

Objectives

Computer vision is a subfield of computer science that focuses on extracting "useful information" from images and videos. The goal of computer vision is to "discover from images what is present in the world, where things are located, what actions are taking place" (Marr, 1982). Examples of "useful information" include, for example, recovering the 3D geometry of objects in an image, detection and identification of human faces and gestures, and tracking moving people or vehicles in a video sequence. Computer vision algorithms have found a wide range of applications in the industrial, military and medical fields, as well as in the ever-growing entertainment field.

This course is an introduction to basic concepts and methods in computer vision. It is mainly suited for MIEIC students who are interested in following research in this area. The covered topics include: image formation, basic image processing and analysis methods, as well as more advanced methods like 3D scene reconstruction, motion analysis, machine learning / deep learning, and object recognition. 

Learning outcomes and competences

Upon completion of this course, students will:

• understand and be able to explain the basic concepts of computer vision and the fundamental algorithms for the manipulation of images and video sequences;

• have knowledge of existing methods for visual data analysis and be able to apply them in practical situations;

• acquire skills to use a library, like OpenCV, that implements some of the analyzed algorithms and to implement novel algorithms described in the literature;

• be able to analyze and understand selected scientific papers in image processing and analysis, and computer vision.

Working method

Presencial

Pre-requirements (prior knowledge) and co-requirements (common knowledge)

Approval in the Programming, Algorithms and Data Stuctures, and Algebra courses (or equivalent) is advisable.

Program

Introduction to Computer Vision

Image acquisition

• intensity images (2D) and distance/position images (3D)
• geometric and radiometric model of a camera

Processing and analysis of intensity images

• filtering
• feature extraction
• segmentation

Motion

• motion estimation

Recognition

• feature selection
• description using local invariant features
• matching of features

Deep Learning

• fundamental concepts of Machine learning
• neural networks

Convolutional neural networks (CNNs)

• concepts and architectures
• detection and segmentation (R-CNN and YOLO models)
• visualization of CNNs

Image generation

• GAN models
• diffusion models

Attention-based models

• concept of attention
• transformers
• vision-language models

Case studies

Mandatory literature

Richard Szeliski; Computer vision, 2011. ISBN: 978-1848829350
Gonzalez , Rafael C.; Digital image processing. ISBN: 0-13-335672-8

Complementary Bibliography

David A. Forsyth, Jean Ponce; Computer vision. ISBN: 0-13-085198-1
Goodfellow , Ian; Deep learning. ISBN: 978-0-262-03561-3

Teaching methods and learning activities

• Lectures: Presentation and discussion of the course topics, and resolution of exercises.

• Practical assignments: Development of projects where the studied computer vision methods must be applied.

Software

OpenCV

Evaluation Type

Distributed evaluation without final exam

Assessment Components

Designation	Weight (%)
Trabalho laboratorial	50,00
Teste	50,00
Total:	100,00

Amount of time allocated to each course unit

Designation	Time (hours)
Estudo autónomo	70,00
Frequência das aulas	42,00
Trabalho laboratorial	50,00
Total:	162,00

Eligibility for exams

Do not exceed the absence limit and obtain a minimum of 40% in the distributed evaluation classification.

Calculation formula of final grade

Distributed evaluation, including project (PR) and minitest (MT)

PR - A group project will be developed during the semester that is developed both during classes and at home. Each group will deliver a small report (in a scientific paper format) and do an oral presentation of the project.

MT - A written minitest will be done, covering the theoretical concepts presented during the course lectures.

PR and MT are specified on a 0 to 20 scale.

Final Classification = PR * 0.5 + MT * 0.5

Oral examination: whenever needed, according to a decision of the teaching team, students may be submitted to an oral exam. In this situation the final classification will be given by the average of the classification calculated with the previous formula and the classification of the oral exam.

Observation: A minimum of 40% on the MT evaluation component is required to be approved in the course.

Second season exam: the second season exam replaces only the grade corresponding to the MT component.

Examinations or Special Assignments

See PR, in Evaluation components.

Special assessment (TE, DA, ...)

Students with a special status will be assessed in the same way as ordinary students. They have to do all the assignments and deliver them on the scheduled dates.

Classification improvement

Students can only improve the mark of the distributed evaluation component in the following year. Students can improve the mark of the written minitests at the corresponding seasons (according to the rules).