Titulo Estágio
Music-Induced Emotion Recognition by Brain Activity Analysis (BRAIN-SONG)
Áreas de especialidade
Sistemas Inteligentes
Local do Estágio
Laboratórios do Grupo de Computação Adaptativa do CISUC
Enquadramento
Music emotion recognition (MER) is a research area of increasing interest in Music Information Retrieval (MIR). It has a wide range of applications in fields such as automatic music classification, playlist generation and similarity analysis.
In fact, recent studies identify music emotion as an important criterion used by people in music retrieval and organization. Moreover, music psychology and education acknowledge the emotion component of music as the one most strongly associated with music expressivity.
However, if one consider just MIR, some questions arise: Is the user really enjoying the selection of a given music? How is its brain reacting?
Objetivo
This thesis proposal aims to complement existent work on MER by extracting information from users cerebral activity, and to inspect how a given music selection is affecting its brain. Features from cerebral waves, measured by a portable encephalogram (EEG) device connected to an Android cell phone will be considered. These features should be able to describe basic emotions (excitement, tension, sadness, contentment, etc.) of the user with the selected music.
Plano de Trabalhos - Semestre 1
1. State-of-the-art review
• Objectives: Review about music emotion recognition, and about EEG emotion correlates.
• Start: September 2014
• End: October 2014
• Expected Results: A review about the state-of-the-art.
2. Development of EEG features related to emotions
• Objectives: Development of EEG features able to discriminate, for example: excitement, engagement, and frustration should be developed, mainly from the EEG frequency response.
• Start: September 2014
• End: December 2014
Expected Results: Android code able to compute features in real-time.
Plano de Trabalhos - Semestre 2
3. Development of a classification system able to discriminate different feature patterns
• Objectives: Development of pattern recognition tools able to classify features to belonging to one of the emotional states.
• Start: Janeiro 2015
• End: April 2015
Expected Results: Android code able to classify features in real-time.
4. Development of an emotion-based classification that accounts for cerebral activity
• Objectives: Development of an emotion-based classification that takes in account MER but also the cerebral classification developed in 3.
• Start: March 2015
• End: May 2015
Expected Results: A report comparing the traditional emotion-based classification with the new approach.
5. Thesis writing
• Start: May 2015
• End: June 2015
Condições
Competencies:
-Pattern Recognition Techniques
-Android Programming
Observações
Co-supervisor:
Professor Rui Pedro Paiva
email: ruipedro@dei.uc.pt
Remuneração: Estágio não remunerado.
Orientador
César Teixeira
cteixei@dei.uc.pt 📩