Datasets, working papers, and peer-reviewed articles produced by MultiObs researchers.
Showing 1–2 of 2
De Freitas, Sergio Antonio Andrade; da Silva, André Corrêa; Silva, Milena de Faria; Batista, Renan Carneiro; Segundo, Washington
Veríscimo, Erico de Souza; Bernardes Júnior, João Luiz; Digiampietri, Luciano A.
User experience (UX) is fundamental for the acceptance and use of information systems. Although there are well-known and widely used UX evaluation techniques for traditional interfaces, a literature review revealed several gaps regarding UX evaluation in non-immersive 3D interaction. One significant gap is the predominant use of pragmatic criteria in assessments, while another is the lack of an approach that evaluates hedonic aspects using facial emotion recognition. This work proposes an approach for automatically evaluating user experience in non-immersive three-dimensional environments, focusing on its hedonic aspects based on facial emotion recognition. An experimental protocol was developed and approved by the Ethics Committee. The experiment was conducted with 52 participants. Throughout the testing period (before, during, and after the interaction), participants' faces were recorded using a low-cost camera. The experiment involved participants playing a game and answering questionnaires, including categorization and mood profile instruments, as well as the UEQ-S and the PLEX Framework. The Face-api.js library was used for facial emotion recognition. The hypothesis that automatic facial emotion recognition can support user experience evaluation was confirmed. This method enabled the estimation of UEQ-S and PLEX questionnaire responses with an average error of approximately ±1 point using only emotion extraction through an artificial intelligence model. Given that UX evaluation is crucial for the acceptance of new software or functionality, this work contributes to improving system quality and acceptance.
