Analysis of Facial Emotion Recognition Technology and Its Effectiveness in Human Interaction

ABSTRACT
Facial expressions are connected to experiencing emotions according to Facial-Feedback Theory of Emotion developed by Charles Darwin and William James. Knowing how to read and interpret facial expressions helps us understand how human beings connect with one another in mobile-mediated communication (MMC) and improve our own abilities at managing emotions. There have been multiple types of research conducted that study and compare factors that influence facial emotion recognition in humans. As one of the most prominent channel in human interactions, facial expression is a good indicator of a person’s emotions. Recent technological advancements have had a drastic impact on human interaction. Besides face-to-face chat, video chatting become more and more popular in the digital age. This research provides to understand how video, voice and music influence emotion perception in the MMC context among different user groups. In this research, we evaluated challenges and effectiveness between two primary emotions: happiness and sadness in video chatting. This study analyzed the emotional influences of voice and music on video chatting and its effectiveness on sensing others emotional energy. Participants were from 18–42 years old and 98% of participants are in between 18–32 years old. A series of visual stimuli were created with four models were from Gen Z; Millennial; Gen X; and Baby Boomer. In addition, the research examines the comfort level with different relationships (e.g. friends; acquaintances; and strangers) in the video chatting environment.

ISBN 978-3-319-94600-9
DIO 10.1007/978-3-319-94601-6_23
PUBLISHER Springer, Cham
PUBLICATION DATE 28 June 2018
AUTHORS Qiuwen Li & Young Ae Kim
CATEGORIES Emotional Design
KEYWORDS Human interaction, Positive emotion, Negative emotion

Method

Procedures

The philosophical foundation for the method of this study was based in the virtual human interaction. Therefore, the fundamental assumption of the research design is that the virtual static and motion stimuli are primary and necessary to understanding how an individual senses others emotion through the digital platform as well as how an individual fail to sense others emotions. Additionally, the participant encouraged expressing comfort level of different relationships and acknowledged possible ambiguities and contradictions in the dialogue.

Participant

Participants were 51 college students (20 men, 31 women) and 18 years of age (mean age = 25) or older from two universities (University of South Dakota and South Dakota State University) located in South Dakota. There were volunteer participants and no credit and compensation were given for participating this study. 84.3% participants had lived in the Midwestern United States (e.g. Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, North and South Dakota, Nebraska, Ohio, and Wisconsin). 15.7% participants were born in other places, but then moved to Midwestern and attended universities there. The ethnicity of participants is 84.3% of Caucasian, 7.84% of Asian, and 7.84% of His-panic.

Experimental Stimuli

A total of 103 survey questions are divided into four sections – (1) background information (7 questions), (2) emotional response (80 questions), (3) internet usage (9 questions), and (4) the comfort level in the video chat based on different relationships (14 questions). In particular, the images and videos that used in the emotional response section of the survey, were created with four models from each generational group – Gen Z(Female), Millennial (Male), Gen X (Male), and Baby Boomer (Female). Gender distribution of model was equal to two males and two females. There is a specific tone of colors that are a favor to the different gender. For example, warm tone is preferred by women than men (Hillock 2003). Thus, colors were strategically removed from stimuli to ensure the direct influences from the visual, voice, and music only.

Procedure

The experiment was approved through the Institutional Review Board (IRB) of the University of South Dakota. In compliance with the IRB, all participant signed a consent form notifying them of their rights as a subject before the experiment began. A series of visual stimuli were used with voice and music to measure the effectiveness on sensing others emotional energy with eight questions. A sample question is “What emotion do you read?”. It is scored on a 2-point Likert scale. The voice and music have positive and negative influence to support or distract the participant from sensing others emotion from given static image and video. The first-person point of view was used in the positive voice - “I am so excited.” and the second-person point of view was used in the negative voice - “You hurt my feeling.” in order to create the personal connection from the participants to the given visual stimuli. The positive and negative music was carefully chosen from the most well-known that have been used as sad and happy situation. “The lonely Man” from the incredible Hulk was used as sad music (negative) Analysis of Facial Emotion Recognition Technology and Its Effectiveness 221 and “Merrie Melodies” from the Looney Tunes was used as a happy music (positive) in this survey. The comfort level in the video chat based on relationships was scored on a six-point Likert scale ranging from very uncomfortable (1) to strongly comfortable (6). A sample question is “How do you feel when you do video chat?”.

The survey averaged 30 min in length and took place at two college locations - University of South Dakota and South Dakota State University. All participants were asked to complete their survey through the Qualtrics and no discussion was allowed during the survey. There was an assistant in the lab to assist conducting survey and technical support.

Single Project

Visual stimuli happy (left) and sad (right) emotional expressions

RESULTS

Data Screening. Prior to analysis, was explored to examine several crucial preliminary topics, missing data, potential outliers, and normality. There was no missing data and a total of 51 participants completed the experimental procedure. With all data point included, it was normally distributed. The data was analyzed to compare the main effect of voice and music (two independent variables) on sensing others emotional energy.

The Impact of Visual Stimuli Both Static Image and Video to Sense the Feeling of Others. The study found that both static images and videos helps participants to sense the feeling of others and taking them on as own significantly. The main effect of both visual stimuli yielded an F(1, 50) = 0.009, p < .02, indicating a statistical significance. In particular, video stimuli help participants to understand others feeling with 93% of chance, which is easier than static stimuli with 88% correctly reading others feeling. In sum, participants correctly read the feeling of others regardless of the visual stimuli type; however, the video stimuli help participants to sense the feeling of others at much better rate (Table 1).

Table 1. Comparison between static image vs. video stimuli influences

Single Project

The Impact of Music with Visual Stimuli to Sense the Feeling of Others. The study found that both negative and positive music were influenced participants for their decision making to understand others feeling. The main effect for music on sensing the feeling of others with static image yielded an F ratio of F(1, 50) = 0.3, p < .001 and with video yielded an F ratio of F(1, 50) = 0.3, p < .001, indicating a statistical significance. The interaction of music and visual stimuli was significant and the result indicated that the music influenced to sense the feeling of others for all participants. For example, the music was distracted participants from reading correct feeling of others when sad music with positive visual stimuli displayed together. However, the video stimuli help to read the feeling of others with music influences whether it is a sad music (negative) or happy music (positive) than the static image with music. In sum, the music distorts the perception of reading feeling of others when the expression of visual stimuli (e.g. happy) is opposite of the musical expression (e.g. sad) and increases the perception of reading feeling of others when the expression of visual stimuli (e.g. happy) is same as the musical expression (e.g. happy)

(Table 2). Comparison between static image with music vs. video with music influences

Single Project

The Impact of Voice with Visual Stimuli to Sense the Feeling of Others. The study found that both negative and positive voice were influenced participants for their decision making to understand others feeling. The main effect for voice on sensing the feeling of others with static image yielded an F ratio of F(1, 50) = 0.15, p < .001 and with video yielded an F ratio of F(1, 50) = 0.2, p < .001, indicating a statistical significance. The interaction of voice and visual stimuli was significant and the result indicated that the voice influenced to sense the feeling of others for all participants. For example, the voice was distracted participants from reading correct feeling of others when sad music with positive visual stimuli displayed together. However, the video stimuli help to read the feeling of others with voice influences whether it is a sad voice (negative) or happy voice (positive) than the static image with voice. In sum, the voice distorts the perception of reading feeling of others when the expression of visual stimuli (e.g. happy) is opposite of the voice expression (e.g. sad) and increases the perception of reading feeling of others when the expression of visual stimuli (e.g. happy) is same as the voice expression (e.g. happy)

(Table 3). Comparison between static image with music vs. video with music influences

Single Project

The Difference Between Voice and Music Influence. The study found that the voice helps to sense the feeling of others significantly higher with p < .001 than the music. Interaction of voice and visual stimuli was more significant than music and visual stimuli. The result indicated that the voice influenced to sense the feeling of others for all participants significantly better even pairing with opposite emotional expression between two – e.g. negative voice (You hurt my feeling) and happy static image and video stimuli (Table 4 and Fig 5).

(Table 4). Comparison between voice and music influences on sensing feeling of others

Single Project Single Project

Fig. 5. Visual stimuli happy (left) and sad (right) emotional expressions

References.

Baumgardner, S.R., Crothers, M.K.: Positive Psychology. Pearson Education Inc., Upper Saddle River (2009)
Cacioppo, J.T., Hawkley, L.C., Thisted, R.A.: Perceived social isolation makes me sad: 5- year cross-lagged analyses of loneliness and depressive symptomatology in the Chicago health, aging, and social relations study. Psychol. Aging 25(2), 453–463 (2010)
Collin, Catherine: The Psychology Book: Big Ideas Simply Explained. DK Publishing, New York (2016)
Darwin, C., Prodger, P.: The Expression of the Emotions in Man and Animals. Oxford University Press, New York (1998)
Ekman, P., Davidson, R.J., Friesen, W.V.: The Duchenne smile: emotional expression and brain physiology: II. J. Pers. Soc. Psychol. 58(2), 342 (1990)
Gabrielsson, A.: Emotions in strong experiences with music. In: Juslin, P.N., Sloboda, J.A. (eds.) Music and Emotion: Theory and Research, pp. 431–449. Oxford University Press, New York (2001)
Hatfield, E., Cacioppo, J.T., Rapson, R.L.: Emotional contagion. Curr. Dir. Psychol. Sci. 2 (3), 96–99 (1993)
Henry, J.D., Von Hippel, W., Molenberghs, P., Lee, T., Sachdev, P.S.: Clinical assessment of social cognitive function in neurological disorders. Nat. Rev. Neurol. 12(1), 28–39 (2016)
Hallock, J.: Colour assignment. (2003)
Fox, J.K.: Factors of Emotion Recognition in Faces: Three Perspectives (2004). http://www. jyi.org
Larsen, J.T., McGraw, A.P., Cacioppo, J.T.: Can people feel happy and sad at the same time? J. Pers. social Psychol. 81(4), 684–696 (2001)
Rainie, L., Zickuhr, K.: Video calling and video chat [Web blog post 2010]. http://www. pewinternet.org/2010/10/13/video-calling-and-video-chat/
Barrett, L.F.: How To Understand Emotions? [Web video 2017]. https://www.youtube.com/ watch?v=GV7ui_KQXcE
Morland, R.L., Beach, S.R.: Exposure effects in the classroom: the development of affinity among students. J. Exper. Soc. Psychol. 28, 255–276 (1992)
Rattue, P.: (Tuesday 14 February 2012). Sad Music Can Bring On Real Sadness [Web blog post]. https://www.medicalnewstoday.com/articles/241653.php
Rozin, P., Berman, L., Royzman, E.: Biases in use of positive and negative words across twenty natural languages. Cognit. Emot. 24(3), 536–538 (2010)
Soussignan, R.: A test of the facial feedback hypothesis. American Psychological Association (2002)
Russell, J.A., Carroll, J.M.: On the bipolarity of positive and negative affect. Psychol. Bull. 125(1), 3–30 (1999)