Affective computing in generative art installations: The case of emo-land

Abstract

This thesis explores the relationship between human emotion and Artificial Intelligence (AI) in affective and interactive environments. The goal is to reduce individuals' distractions in public places, an area largely neglected by current AI and IoT technologies. Emotions play a significant role in people's behavior and decision- making, and attention. On the other hand, the built environment has a profound impact on emotions, and well-being. Therefore, this research investigates how affective and interactive installations can influence users' attention and mitigate distractions in public spaces. This study stands at the intersection of several critical domains, drawing upon theories of generative art, affective computing, calm technology, and ambient media to forge a comprehensive framework for the design of installations within public spaces. While extensive research has explored the realm of interactivity, the intricate relationship between generative art installations, affective computing, and architectural elements has remained ambiguous. In this context, the research not only sheds light on these connections but also elucidates how technological advancements and computational capabilities can infuse deeper connections and fresh dimensions of interactivity, serving as a means to counteract the pervasive distractions that have proliferated in the age of ubiquitous screen technology. This study thus emerges as a pivotal contribution in the ongoing quest to harness technology and design for the purpose of restoring attention, fostering meaningful connections, and redefining the human experience within public environments. The research methodology involves the development of a computational system that incorporates generative techniques applied through installations in public spaces. Facial expressions are used as a parameter to detect emotions through a Convolutional Neural Network (CNN) model trained for Facial Emotion Recognition (FER). An interactive visual art interface called Emo-Land is designed and developed using AI- models and projection mapping. The complexity of patterns in the installation adapts based on the user's emotional state. By discussing the principles of Calm Technology and their application to the Emo- Land installation, the study tries to create an installation that let the users to restore their attention. This research aligns with the third wave of Human-Computer Interaction (HCI), emphasizing human values, meaning-making, and socially situated and embodied action. The evaluation of the Emo-Land's usability garnered an overall positive response, with users perceiving it as easy to use and expressing contentment with their interaction experience. The assessment encompassed dimensions of easey to use, efficiency, and satisfaction. Notably, users found the installation's design user-friendly and engaging, though some improvements are required to enhance its efficiency further. In terms of users' experiences, the majority of questions received favorable feedback, underscoring a positive ambiance and enhancing the spatial experience. Users' emotions were influenced by the installation. Biophilic patterns, while contributing to attention restoration, posed challenges in fully grasping their meanings. Nevertheless, participants positively regarded the participatory interaction aspect of the installation, highlighting its potential for enhancing user experiences. The findings of this study contribute to the understanding of how affective and interactive installations influence users' attention in public places and reduce distractions. By incorporating principles of Calm Technology and utilizing biophilic patterns, the research offers new perspectives on designing generative installations that foster an ambient commons in public places.