Facial expressions: emotions recognition based on selected landmarks

• Autor(a) principal: Lopes, Daniela Soares de Pina
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositórios Científicos de Acesso Aberto de Portugal (RCAAP)
• Texto Completo: http://hdl.handle.net/10773/36998
• topic: Affective computing; Facial expressions recognition; Facial landmarks; Image processing; Machine learning

Automatic emotion recognition based on facial expressions is an active and challenging research field, owning to its academic and commercial potential. Most of the already existing facial expression recognition systems use a holistic approach to withdraw emotional characteristics from facial images, i.e., they adopt methods that make use of all facial attributes in order to recognize emotions. This project aims to identify the regions of the face that characterize each emotion within a set of basic emotions. For this, research was conducted to study the facial movements that occur from the neutral expression to the apex of a given emotion, resorting to the analysis of the evolution of the euclidean distances between facial landmarks throughout the frames of a video. From this analysis, there were selected sets of facial landmarks essential to identify each facial expression. The model used to perform the facial expressions classification was the Support Vector Machine (SVM). During this project, it was possible to verify that the number of facial landmarks can be reduced to a minimum value, promoting better computational efficiency, the usage of simpler algorithms, and also, in case it is necessary to store information, preserve people’s privacy since it diminishes the probability of re-identifying the intervening subjects with only the selected features.