Emotion elicitation with virtual reality

Over the last few years virtual reality has become a fundamental aspect in the creation of any multimedia content. Just think of how many video games come out every day on the most different platforms, or how many advertisements use this technique to better capture the viewer. This technique has also begun to be used in laboratories, which offers the possibility of simulating all real/unreal situations to investigate human behavior in well controlled designs in the laboratory. A fundamental aspect of human behavior is that concerning emotions, and, in particular, the Facial Expression Recognition (FER) is an important aspect to develop system that can automatically recognize emotions. And it is precisely in this context that the work of this thesis fits. In fact, the aim of the experiment was to create five virtual environments that had the purpose of elicit a certain emotion in the people who would have navigated them. In particular, two scenarios should have elicit fear, two anger / frustration and the last disgust. While the user navigated these five environments he/she was filmed by an RGB camera and a coded-light depth integrated sensor that had the aim of recording all the facial expression that would emerge during the experiment and to provide a 3D map of the acquisition. The purpose of the experiment is multiple. The first goal is to enrich an ecologically-valid facial expression database (created from previous works using non-immersive stimuli, likes images, taken from the IAPS and GAPED databases) by eliciting this three human emotions with the virtual environments. The database thus obtained was then used to test and train a neural network capable of recognizing facial expression and connecting it to the corresponding emotion. The creation of this neural network would make it possible to create applications in the context of Human-Computer Interaction (which concerns the methods of interaction between human beings and computers) or in the context of Emotional Design (which is a design approach to create objects able to leave a positive experience in users who try them). Moreover, the same virtual reality environments can be used to become aware of emotions, a very useful aspect for creativity and therefore productivity. Before navigating the virtual environments, each lasting 90 seconds, each participant was asked to fill out a questionnaire on their current emotional state and after each environment, every person was asked to evaluate, in term of arousal and valence through the Self-Assessment Manikin (SAM) scale, the tried environment and to label the emotions felt, in order to understand if the environment had obtained the desired effect. Furthermore, before participating in the experiment, each person was asked to fill out a questionnaire on empathy and alexithymia to understand their emotional profile. The experiment, carried out in collaboration with the Politecnico di Milano, was conducted at the 3D Lab of the Politecnico di Torino with a sample of 28 students. The results obtained showed how the virtual environments were able to arouse the desired emotion. The whole project, from the design of the virtual environments to the experimentation, is presented.