Propostas atribuídas ano letico 2025/2026

DEI - FCTUC
Gerado a 2025-08-31 19:11:39 (Europe/Lisbon).
Voltar

Titulo Estágio

Automatic Mood Tracking in Audio Music

Área Tecnológica

Multimédia, Engenharia de Software, Processamento de Áudio

Local do Estágio

DEI

Enquadramento

Audio music mood-based analysis is a research area of increasing interest in Music Information Retrieval (MIR). It has a wide range of applications in fields such us automatic music classification, automatic playlist generation and similarity analysis. 

In fact, recent studies identify music mood/emotion as an important criterion used by people in music retrieval and organization. Moreover, music psychology and education recognize the emotion component of music as the one most strongly associated with music expressivity. 

The analysis of audio music in terms of mood/emotion is challenging in its very nature: mood is a subjective notion, techniques are still in an embryonic stage and a uniform evaluation framework is yet to be agreed upon. Nevertheless, the Music Information Retrieval Evaluation eXchange (MIREX, a small “competition” that takes place every year), has in 2007 (and for the first time) a track on audio music mood classification, which will certainly give a strong impulse towards the improvement of techniques and creation of evaluation standards. 

Objetivo

The main objective of this thesis is to implement a software tool able to the track the mood 

variations throughout an audio musical clip, e.g., contentment, depression, exuberance and 

anxiety. This involves the study and derivation of mood-like features, mood-based similarity 

metrics (namely self-similarity metrics) and ranking. 

References: 

- Juslin P.N., Karlsson J., Lindström E., Friberg A. and Schoonderwaldt E. (2006) “Play It Again 

With Feeling: Computer Feedback in Musical Communication of Emotions”, Journal of 

Experimental Psychology: Applied, Vol. 12, No.2, pp. 79-95.  

- Yang Y.-H., Lin Y.-C., Su Y.-F. and Chen H. H. (2008). “A Regression Approach to Music 

Emotion Recognition”. IEENo. 2, pp. 448-457. 

Plano de Trabalhos - Semestre 1

1.  Study and summarization of the state-of-the art. 

2.  Requirements analysis. 

3.  Review of mood-based feature extraction. 

4. Review of (self-)similarity metrics. 

5. Design of the graphical user interface.  

6. Documentation (in each stage) and intermediate report. 


Plano de Trabalhos - Semestre 2

1.  Implementation of mood-based feature extraction. 

2. Implementation of (self-)similarity metrics. 

3. Implementation of the graphical user interface. 

4.  System evaluation: accuracy, computational efficiency and subjective tests.  

5.  Software tests. 

6. Documentation (in each stage) and thesis writing. 

7.  Writing of a scientific article. 

(8. Design, implementation and evaluation of new approaches to the problem. (if time 

permits)) 


Condições

No salary available.

Orientador

Rui Pedro Paiva
ruipedro@dei.uc.pt 📩