Study and evaluation of human perception on virtual humans and crowds

• Ano de defesa: 2020
• Autor(a) principal: Araujo, Victor Flávio de Andrade
• Orientador(a): Musse, Soraia Raupp
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Pontifícia Universidade Católica do Rio Grande do Sul
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação
• Departamento: Escola Politécnica
• País: Brasil
• Palavras-chave em Português: Percepção humana; Multidões; Interações; Características culturais; Personagens de CG
• Palavras-chave em Inglês: Human perception; Crowds; Interactions; Cultural features; CG characters
• Área do conhecimento CNPq: CIENCIA DA COMPUTACAO::TEORIA DA COMPUTACAO
• Link de acesso: http://tede2.pucrs.br/tede2/handle/tede/9279

The behavior of the human being depends on his perception of the world. Behaviors have been studied for a long time, in different aspects, in particular research in Graphic and Visual Processing. In addition to being studied, behaviors can be simulated, as well as analyzed in videos, and in both cases studies of real behaviors are of great value. The Crowd Simulation area is in this context, where the objective is to simulate and provide realistic behaviors. A correlate study is the area of human perception regarding virtual humans, groups and crowds. Human perception is essential to identify whether another person is crying, or is angry, or any other characteristic regarding the shape, movement or cultural aspects of the simulated virtual humans. With all the research on human behavior and the evolution of Computer Graphics (CG), the entertainment industry has created very realistic characters and faces, sometimes causing strangeness and involuntary feelings in human perception. In this sense, this work aims to carry out studies and evaluations of human perceptions about human beings and virtual crowds. We divided these analyzes into three parts: i) Perception of interactions in crowds, ii) Perception of geometric and cultural features in crowds, iii) and Perception of CG characters. Regarding the first part, we simulate interactions between virtual agents through their spatial perceptions, identify interactions (using the same parameters of interactions of virtual agents in the simulations) between people in real video sequences, and made interactive visualizations to facilitate the analysis of the data generated by interactions (both, in simulations and videos). In the second part, we analyze human perception through a questionnaire about geometrically extracted data (distance between people, people’s speed, angular variation and density), in addition to non-geometric data, such as personalities and emotions. In the third part, we analyze the human perception of characters created with CG (movies, games, animations, etc.) in order to answer questions raised by the effects of Uncanny Valley, also through questionnaires.