Pitch Angle: A Newfound Influential Trait for Image-Based Facial Beauty Perception. (November 2020)

• Record Type: Journal Article
• Authors: Guo, Hao; Han, Wenming; Chen, Fangmei; Sun, Fuming
• Is Part Of: Journal of physics. Volume 1651(2020)
• Publication Date: 2020-11
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1651/1/012144
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: One of the most crucial issues in facial beauty study is to find the traits, which determine the perceived attractiveness of a face. During the decades, many traits have been studied extensively, such as proportions, averageness, and symmetry. However, facial beauty seems to be more complex than these traits. In this paper, a face image is represented by a series of interpretable parameters, which are learnt by a CNN based 3DMM fitting algorithm. Based on the SCUT-FBP5500 dataset, we explore the relationship between those parameters and the human rated facial beauty scores. We observe that the pitch angle parameter has a strong correlation with the beauty score. In order to study the causal relationship between the pitch angle and facial beauty, face images with different pitch angles were generated as the stimuli and 39 volunteers were invited to a rating experiment. The results show that the change of the pitch angle has a significant effect on the perceived facial beauty. This finding may explain that head pose plays an important role in the self-portrait shot.