Applied Signal Processing with artificial intelligence (ELE640)

We surround ourselves with smart phones, watches and sensors. With the help of such equipment we have conversations, listen to music, watch films, receive information about the world around us as well as monitor our surroundings and ourselves. We need advanced signal and image processing to be able to interpret and give meaning to data from such sensors, including medical equipment. This course builds on subjects such as Signal Processing, Image Processing and Computer Vision as well as Machine Learning. We learn basic theory, some new techniques and establish "building blocks", which we use in specific applications. We learn to extract features from data we can feed into machine learning programs.


Course description for study year 2025-2026. Please note that changes may occur.

Facts

Course code

ELE640

Version

1

Credits (ECTS)

10

Semester tution start

Spring

Number of semesters

1

Exam semester

Spring

Language of instruction

Norwegian

Offered by
Time table

View course schedule

Content

The students will learn about:

Classical machine learning models such as:

  • Support Vector Machines (SVM).
  • Random Forest (RF).
  • XGBoost.

Time and frequency analysis using:

  • Short Time Fourier Transform (STFT).
  • The Gabor Transform.
  • The Wavelet Transform.

Texture analysis using:

  • Co-occurrence matrices.
  • Local binary pattern (LBP).

Techniques and methods for signal and image compression using:

  • The Discrete Cosine Transform (JPEG).
  • Wavelets (JPEG2000).

Multirate signal processing using polyphase decomposition.

Generative deep learning models such as:

  • Autoencoders.
  • General Adversarial Networks (GAN).

We will use Python actively to apply the aforementioned machine- and deep learning methods as well as signal- and image processing methods to:

  • Build biomedical decision systems for disease classification.
  • Generate features from datasets consisting of magnetic resonance imaging (MRI) of the brain and electroencephalography (EEG) of the brain for classification of different types of dementia in the elderly.
  • Compress photographs.
  • Enable signals from different sensors to work together.
  • Noise reduction.

Learning outcome

Knowledge:

The student will learn advanced signal- and image processing techniques as well as machine- and deep learning techniques, building modules mentioned in ELE500 Signal Processing, ELE510 Image Processing and Machine Vision, and ELE520 Machine Learning.

The student will gain knowledge of several essential classical machine learning models (SVM, RF, XGBoost), signal and image processing tools within the topics of time and frequency analysis (STFT, the Gabor Transform, and wavelets), texture analysis (co-occurrence matrices and LBP), techniques and methods for signal and image compression (JPEG and JPEG2000), multirate signal processing (polyphase decomposition), and generative deep learning methods such as autoencoders and GANs.

Skills:

The student will apply the aforementioned methods in practical applications using programming tools in Python. The student will gain insight into how machine learning and deep learning models, as well as signal and image processing techniques, can be used in concrete applications such as disease classification, image compression, resampling of data from multiple sensors, and noise reduction.

General competence:

After completing this course, the student will have a theoretical understanding of important classical machine learning and deep learning models, as well as advanced signal and image processing concepts. The student will have a basic understanding of how these methods can be applied to relevant data using programming tools in Python.

Required prerequisite knowledge

None

Recommended prerequisites

ELE500 Signal Processing, ELE510 Image Processing and Computer Vision
It would be an advantage if students have experience with or are concurrently studying signal processing, image processing, and machine learning equivalent to ELE500, ELE510, and ELE520.

Exam

Form of assessment Weight Duration Marks Aid
Written exam 1/1 4 Hours Letter grades No printed or written materials are allowed. Approved basic calculator allowed

Coursework requirements

Assessement, Attendance at lectures

Programming assignments using Python:

6 out of 6 assignments must be approved by the course teacher within the specified deadlines.

The course has mandatory attendance.

Course teacher(s)

Course coordinator:

Ketil Oppedal

Head of Department:

Tom Ryen

Method of work

The course is divided into six topics, with each topic lasting approximately two weeks. Students will be provided with resources to familiarize themselves with the topics (videos and theoretical material). Furthermore, students will have 4-6 hours of work in supervised groups per week, where the goal is to apply the studied methods to relevant data using programming tools in Python. All groups will present their results in various ways throughout the semester at the end of each two-week period.

Open for

Admission to Single Courses at the Faculty of Science and Technology
Cybernetics and Applied AI - Master of Science Degree Programme

Course assessment

There must be an early dialogue between the course supervisor, the student union representative and the students. The purpose is feedback from the students for changes and adjustments in the course for the current semester.In addition, a digital course evaluation must be carried out at least every three years. Its purpose is to gather the students experiences with the course.

Literature

Search for literature in Leganto