Face detection : the role of race and color

Abstract

Human face perception literature found that face detection is a prerequisite process which must first occur before any other face perception processes, such as face recognition and identification can occur. Despite the importance of face detection, little research effort has been placed into understanding the processes underlying face detection. As such, Lewis & Ellis (2003) proposed a basic model representing processes involved in face detection. These processes include visual search and face template matching. In face detection literature, tasks which involves both visual search and face template matching are considered as face detection tasks, while tasks which lacks visual search are considered as face categorization tasks. One of the most established finding amongst face recognition literature is the other-race effect, whereby people are better at recognizing faces from their own race. Drawing from findings across both literatures, this study aims to find out if people are better at detecting own-race faces, and if they are better at detecting human faces with color information. This study employed a rapid serial visual presentation design whereby participants had to react whenever a face image was detected. The face images are of Chinese, Indian and Malay Singaporean origin. Furthermore, only Singaporean Chinese participants were recruited. Results suggest that same-race faces are detected faster when compared to other-race faces in face categorization task. The advantage seems to be moderated by contact experience. It was also found that presence of color information result in minimal improvement of face detection performance in a face categorization task.