SCV0954 - ARTIFICIAL INTELLIGENCE FOR SIGNAL ANALYSIS

insegnamento

ID:

SCV0954

Durata (ore):

48

CFU:

SSD:

INFORMATICA

Informatica

Anno:

2026

Periodo di attività

Secondo Semestre (22/02/2027 - 28/05/2027)

Obiettivi Formativi

The course aims to provide the basic knowledge of analyzing and processing a variety of signals, including multimedia data (audio and images) and physiological signals, such as electrocardiogram (ECG), skin conductance and electroencephalogram (EEG). At the end of the course, the student will be able to:

1) know the basics of the transition from analog to digital signals, sampling, quantization, and coding.
2) Manage and process digital signals using linear time-invariant systems and frequency analysis.
3) Apply the signal processing techniques in the case of audio signals and images (sampling, quantization, and filtering).
4) understand how to process physiological signals: feature extraction and classification tasks through classical machine learning techniques (like for example decision trees, K-Nearest Neighbors and Support Vector Machines).
5) understand and apply digital signal processing to research topics like affective computing and brain-computer interfaces.

Prerequisiti

Basic concepts of Mathematical Analysis and linear algebra.

Metodi didattici

The course consists of 48 hours of lectures.

Verifica Apprendimento

The examination consists of an oral discussion of a research article chosen from a pool of journal articles provided by the teacher. The paper will include several topics related to the course, and the discussion aims to assess what has been learned during the lessons. The student should be able to present the main findings and methodology of the paper, discussing and explaining the related topics studied during the course. The final mark (out of 31) depends on the completeness and correctness of the exposition (70%), the clarity of the exposition (20%), and the adequateness of answers (10%).

Contenuti

Definition of one-dimensional signals, two-dimensional signals, and N-dimensional signals. Analog signal, digital signal, media, variance, energy and power, noise (6 h, teaching goal 1).

Signals in the transformed domain: Fourier Transform for periodic, continuous, and discrete signals. Convolution theorem (8 h, teaching goal 2).

Analog to digital conversion: sampling theorem, quantization, anti-aliasing filtering, Shannon theorem (10 h, teaching goals 1-2).

Introduction to Linear Time invariant Systems (LTI): definitions, input / output equation, convolution, filtering (8 h, teaching goals 1- 2)

Audio signals and images: sampling and quantization, filtering (8 h, teaching goals 1-3).

Physiological signals (ECG, skin conductance, EEG) are applied in human machine interaction and intelligent system applications. Learning from physiological data: From Affective Computing to Brain Computer Interfaces. How to manage input and output data in emotion recognition tasks (8h, teaching goals 4-5).