Study Of The Precise Detection Of Facial Expression Recognition Clues

In this thesis, eyes and the mouth were selected as the key features for research and two experiments were designed to explore the basic cues in identifying facial expressions.This research used the three parameters of gradient, arc and shape to describe the movement of eyes and the mouth. Simple lines stood for eyes and the mouth, with the slope of lines for gradients, the simple curve for arc and double curves for shape. To simplify the task, three parameters were recombined into two groups-one was to combine gradient with arc and the other was to combine gradient with shape.Experiment 1 had a 2x2x6 within-subject design. The three factors were symbol form (simple curve versus double curves), modulating areas (part versus whole) and words of 6 basic emotions (sadness, anger, surprise, fear, happiness, disgust).Experiment 2 had a 2x2x2 within-subject design. The three factors were subject variable (male versus female), symbol form (simple curve versus double curves), and the program of task (production versus duplication).From the two experiments, the author drew four main conclusions: 1) In identifying facial expressions, both the eyes and the mouth had almost the same effect when acting on its own or acting in the whole face. The three parameters showed the same results in modulating eyes and mouth simultaneously and respectively. 2) The results were not influenced by the way to simplify the task by recombining the various features into simple groups. 3) Generally, there were no significant differences in identifying facial expressions for men and women, but still slight differences existed. 4) Different basic cues were relied on in identifying each facial expression.