Computational platform for multimodal affective data analysis

• Autor(a) principal: Cardoso, Vasco Rodrigues
• Data de Publicação: 2021
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10773/33879
• topic: Affective computing; Biosignals; Multimodal; Emotional classification; Web platform; API; Work queues; Information extraction; Affective data analysis

Bio signals can be used to quantify several types of information about a person. These signals, by themselves, can reveal, directly or indirectly, interesting aspects about their source, thus having a wide range of applications. An emerging application is the use of these signals as input for technological systems that detect emotions, in the area of Affective Computing. Emotion identification systems are gaining more and more attention: it is expected that, in the future, many products will be developed based on the emotional analysis of their users, in order to provide more personalized and appropriate experiences, thus obtaining the best possible results. To this end, computational tools are needed that allow to analyze and study bio signals in an easy way, abstracting the underlying computational processes. This work proposes and demonstrates a computational platform that allows its users to analyze multimodal data for emotions detection, generating a summary report in which three base states are identified: happiness, fear and neutrality. The platform can use electrocardiogram (ECG), electrodermal activity (EDA), and electromyogram (EMG) data. The platform enables the user to perform feature extraction and selection, and emotional classification, on these types of bio signals. The platform is available in a Web environment and is oriented towards the end user, who does not need to be an expert either in bio signals or in processing methods. As some analyses may be demanding and therefore not immediate, the platform makes use of the abstraction of work queues for task submission and tracking. An initial version of the platform has been developed and explored with concrete problems of students performing emotion analysis in their academic projects.