Researchers from the Faculty of Science and Technology of (FCTUC) are developing Artificial Intelligence (AI) technologies capable of automatically identifying the emotions present in music, revealed the higher education institution.
In a statement sent to the Lusa agency, the explained that the The project’s objective is to “improve systems that allow music to be classified according to the emotion it conveys, a feature that is increasingly relevant on streaming platforms.”
The investigation takes place within the scope of the “Music Emotion Recognition: Next Generation” (), project coordinated by FCTUC professor which has a team of researchers from the Center for Informatics and Systems at the University of Coimbra (), from FCTUC, and which is financed by the Foundation for Science and Technology with 196 thousand euros.
According to Rui Pedro Paiva, the Merge intends to take a leap forward in sector research“combining audio and song lyrics to create systems capable of automatically classifying emotions, using artificial intelligence, audio signal processing and natural language.”
From technology to results
One of the main innovations is the bimodal approach, which integrates audio and lyric dataallowing you to extract musical descriptors related to rhythm, tonality, expressiveness and percussion.
According to the FCTUC, the project also includes new public databases of emotionally annotated music, fundamental to strengthening research in this area.
Among the results already obtained is an application prototype that positions each song on an “emotional map” with two axes, namely valence (positive or negative emotions) and activation (emotional intensity).
“This allows you to create personalized playlists or explore music according to the user’s mood,” explained the college
In addition to scientific advancement, the project aims to demonstrate the innovations through a standalone application and a ‘web’ platformcontributing to the development of more accurate and useful systems in ‘Music Emotion Recognition’.
Investigation team
The Merge team includes researchers Matthew Davies, Pedro Lima Louro, Renato Panda, Ricardo Malheiro and Ricardo Santos.
The project also included the contribution of doctoral student Tiago Ribeiro and several master’s students, who participated in the development of the scientific work, including Alice Mangara, Hugo Redinho, Guilherme Almeida, Guilherme Branco, Luís Seco, Mariana Paulino, Pedro Sá, Samuel Machado, Simone Chieppa, Tiago Ribeiro and Tomás Ferreira.
