33239 - APPLIED VIDEO SEQUENCES ANALYSIS

This is a non-sworn translation intended to provide students with information about the course

Information of the subject

Code - Course title: 33239 - APPLIED VIDEO SEQUENCES ANALYSIS

Degree: 725 - Máster Erasmus Mundus en Tratamiento de Imágenes y Visión Artificial

767 - Máster Erasmus Mundus en Tratamiento de Imágenes y Visión Artificial

Faculty: 350 - Escuela Politécnica Superior

Academic year: 2022/23

1. Course details

1.1. Content area

1.2. Course nature

Compulsory

1.3. Course level

Máster (EQF/MECU 7)

1.4. Year of study

1.5. Semester

Second semester

1.6. ECTS Credit allotment

6.0

1.7. Language of instruction

THEORY: Group 619: English

1.8. Prerequisites

Secure Verification Code: Date: 18/01/2023

Signed by: This teaching guide will not be signed by CSV until the closing of the proceedings

1/5
URL Verification: Page: 1/5
 1.9. Recommendations

The following skills are necessary for this class:

• Math: Linear algebra, vector calculus, and probability.
• Image processing: image representation, pixels, histograms, image features, gradients/edges, spatial filtering.
• Programming: A good working knowledge for C programming. C++ programming is a plus although it is not formally
required by this course.

Related subjects that may be helpful for this course:

• Machine Learning: theory and applications [Aprendizaje Automático: teoría y aplicaciones]
• Applied Bayesian Methods [Métodos bayesianos aplicados]
• People detection and Biometrics [Biometría]

1.10. Minimum attendance requirement

In order to be assessed on each individual part during the ordinary exam period, attendance is mandatory.
Students are requested to assist to a minimum of 90% of lectures/labs for each individual part. The absence to lecture/labs will
be only permitted under exceptional circumstances which should be justified (e.g. medical reasons). If students fail to meet the
90% assistance criterion, a fail grade will be given and the students will be evaluated according to the extraordinary exam
period.

1.11. Subject coordinator

Juan Carlos San Miguel Avedillo

https://autoservicio.uam.es/paginas-blancas/
1.12. Competences and learning outcomes

1.12.1. Competences

1.12.2. Learning outcomes

By the end of the course, the student must be able to:

1. Be familiar with both the theoretical and practical aspects of video processing systems
2. Have described the foundation of methods for foreground object detection
3. Have implemented common methods for foreground object detection
4. Be familiar with Bayesian and probabilistic formulation of target tracking in video
5. Grasp the principles of state-of-the-art prediction-based target tracking
6. Be able to evaluate video processing algorithms
7. Have developed the practical skills necessary to build video processing tools.
8. Be able to properly report practical work in written form.

1.12.3. Course objectives

The main objective of video analysis for video-surveillance is the extraction of useful semantic information: detection of people
presence and trajectory, detection of irregular events, etc. The general objective of this course is to present automatic video-
surveillance systems (with single or multiple cameras), to describe the practical application of generic video analysis techniques
in this domain, and to present specific video-surveillance applications: people detection and tracking, abandoned/stolen object
detection, etc.

1.13. Course contents

The contents of the course are as follows:

1. Introduction to video processing systems

2. Foreground/Objects detection and segmentation
• Background subtraction: parametric models

Secure Verification Code: Date: 18/01/2023

Signed by: This teaching guide will not be signed by CSV until the closing of the proceedings

2/5
URL Verification: Page: 2/5
 • Background subtraction: non-parametric models
• Shadow detection
• Specific object detectors
3. Video object tracking
• Target modeling andTemplate matching
• Mean-shift tracking
• Kalman and Particle Filters
• Multi-target tracking

1.14. Course bibliography

1. S.A. Velastin, P. Remagnino, "Intelligent Distributed Video Surveillance Systems", IET, 2006
2. O. Javed, M. Shah, "Automated Multi-camera Surveillance: Algorithms and Practice", Springer 2008
3. T. Bouwmans, F. Porikli, B. Hörferlin, A. Vacavant. "Background modeling and Foreground Detection for video
surveillance: Traditional and Recent Approaches, Benchmarking and Evaluation". CRC Press, Taylor and Francis
Group. 2014
4. E. Maggio, A. Cavallaro, "Video Tracking: Theory and Practice", Wiley, 2011.
5. Selected papers in relevant journals and conferences for topics covered in the course

2. Teaching-and-learning methodologies and student workload

2.1. Contact hours

2.2. List of training activities

The course involves lectures, lab assignments (including implementation and documentation), and exams of the lectures'
content. The list of course activites is as follows:

Activity Hours % Hours %

Presential Lecture sessions 24 16%

Practical programming sessions 28 18.6% 56 37.33%

Tests and exams 4 2.6%

Non-presential Weekly study of lectures 24 16%

Practical work (programming and reporting) 60 40% 94 62.66%

Preparation of tests and exams 10 6.6%

TOTAL WORKLOAD: 25 hours x 6 ECTS 150 100%

3. Evaluation procedures and weight of components in the final grade

3.1. Regular assessment

In the ordinary exam period, we will consider the following scheme for overall evaluation:
Grade = 40%TE + 60%PR

TE is the grade obtained from theoretical lectures. TE grade is obtained as follows:

TE = 50%TE1 + 50%TE2
where TE1 and TE2 are two exams, each one corresponding to aproximately half of the lectures' content.

PR is the grade obtained from lab assignments. PR grade is obtained as follows:

Secure Verification Code: Date: 18/01/2023

Signed by: This teaching guide will not be signed by CSV until the closing of the proceedings

3/5
URL Verification: Page: 3/5
 PR = 25%PR1 + 25%PR2 + 25%PR3 + 25%PR4
where PR1...PR4 correspond to four lab assignments distributed along the course.
The grading range is from 0.0 to 10.0. The maximum grade to be obtained is 10.0 and each of the parts (labs and exam) will be
also graded with the same grading range. In order to pass the course, it is necessary to have a pass grade (equal or greater
than 5.0) in the overall evaluation, as well as a pass grade (equal or greater than 4.0) in the two individual parts (TE and PR).
The grades of the individual parts are kept for the extraordinary exam period.

3.1.1. List of evaluation activities

Type Evaluation activity Percent (overall grade)

Lectures

Exam TE1 20%

Exam TE2 20%

Labs

Assignment PR1 15%

Assignment PR2 15%

Assignment PR3 15%

Assignment PR4 15%

3.2. Resit

In case of a fail grade in the ordinary period, in the extraordinary exam period, the grade will be determined by:
• 40 % - Theoretical exam considering all lectures' content
• 60 % - Project on a proposed research topic in video analysis (programming and documentation).

In order to pass the course, it is necessary to have a pass grade (equal or greater than 5.0) in the overall evaluation, as well as
a pass grade (equal or greater than 4.0) in the two individual parts.

3.2.1. List of evaluation activities

Type Evaluation activity Percent (overall grade)

Lectures Unique exam 40%

Project Unique practical assignment 60%

4. Proposed workplan

The proposed workplan for the course considers 14 weeks. If the teaching period contains less than 14 weeks for a given year,
the workplan will be adjusted accordingly. The workplan is as follows:

Week Lectures Labs Lab deadlines

1 Introduction to video processing systems Introduction to OpenCV

2 Foreground/Objects detection and segmentation PR1 Foreground segmentation

3 Foreground/Objects detection and segmentation PR1 Foreground segmentation

4 Foreground/Objects detection and segmentation PR1 Foreground segmentation

5 Foreground/Objects detection and segmentation PR2 Blob analysis Deadline PR1

6 Foreground/Objects detection and segmentation PR2 Blob analysis

Secure Verification Code: Date: 18/01/2023

Signed by: This teaching guide will not be signed by CSV until the closing of the proceedings

4/5
URL Verification: Page: 4/5
 7 Exam TE1 PR2 Blob analysis

8 Video object tracking PR3 Target tracking Deadline PR2

9 Video object tracking PR3 Target tracking

10 Video object tracking PR3 Target tracking

11 Video object tracking PR4 Target tracking Deadline PR3

12 Video object tracking PR4 Target tracking

13 Video object tracking PR4 Target tracking

14 Exam TE2 - Deadline PR4

Secure Verification Code: Date: 18/01/2023

Signed by: This teaching guide will not be signed by CSV until the closing of the proceedings

5/5
URL Verification: Page: 5/5