Geometric Simplification in Humanoid Robot Facial Design: A Cross-Cultural Study of Emotion Recognition

Abstract

This study investigates facial expression recognition in humanoid robot design, aiming to clarify cultural differences in emotional perception and to establish design guidelines for effective nonverbal communication. As robots are increasingly expected to coexist with humans in daily environments, the face functions as a crucial interface for conveying emotional states and social intentions. However, highly humanlike appearances often evoke psychological discomfort, known as the Uncanny Valley phenomenon. To address this issue, this research advocates a design approach that prioritizes artifact-like familiarity and psychological comfort through simplification rather than direct human imitation. We argue that robots should embody forms appropriate to their own functions, instead of replicating human appearance without critical adaptation.Drawing on the concept of characterization, this study focuses on simplified geometric facial designs. Characterization involves abstracting and deforming distinctive features while maintaining recognizability. In this research, facial elements are reduced to minimal components: eyes and a mouth constructed from geometric shapes derived from perfect circles. By systematically manipulating parameters such as aspect ratio, inclination, and curvature, we generated multiple symbolic expression patterns. These variations enabled controlled examination of how geometric deformation influences emotional interpretation.Previous research consisted of a comparative questionnaire experiment conducted with university students in Denmark and Japan. Participants classified perceived emotions according to Ekman’s six basic categories—happiness, anger, sadness, disgust, fear, and surprise—along with two neutral conditions. Results showed minimal cultural variation in the recognition of happiness, anger, sadness, and neutral expressions. In contrast, significant differences emerged for disgust, fear, and surprise, indicating that these emotions may depend more strongly on cultural or contextual factors. The findings suggest that simplified geometric expressions can communicate primary emotions across cultures, whereas secondary emotions require more context-sensitive design strategies.Building on these results, the present study conducts a detailed follow-up investigation focusing exclusively on Japanese university students. This phase analyzes the relationship between specific geometric parameters and emotional cognition in greater depth. By examining how variations in eye slant, mouth curvature, and proportional balance affect perceived emotional intensity, the study clarifies culturally specific tendencies in Japanese emotion recognition. Statistical analyses provide insight into how subtle geometric adjustments influence viewers’ psychological responses and categorical judgments.The ultimate objective of this research is to propose practical facial expression design guidelines for humanoid robots that reflect cultural nuances while maintaining universal interpretability. By extending established emotion models to symbolic and abstract facial representations, this study contributes to foundational discussions on robot morphology and human–artifact interaction. The findings offer designers empirically grounded principles for creating approachable robotic faces that foster psychological safety and intuitive communication. As robots transition from industrial machines to social partners integrated into everyday life, such design strategies will be essential for ensuring comfortable and sustainable human–robot coexistence.Furthermore, the study discusses implications for cross-cultural validation frameworks and iterative prototyping processes in robot design. Integrating quantitative findings with qualitative feedback enables designers to refine symbolic expressions systematically, balancing clarity, subtlety, and cultural resonance in real-world deployment scenarios. This approach strengthens evidence-based creative decision-making in practice.

Keywords: Humanoid Robots, Facial Expression Design, Emotion Recognition, Human–Robot Interaction (HRI)

DOI: 10.54941/ahfe1008226