observation by video recording and measuring psychological burden through a
questionnaire (Obayashi & Masuyama, 2018; Obayashi, Masuyama, et al., 2018).
The involvement of the observer’s subjective judgment is an issue with methods
using observation as the judgment criteria can possibly differ depending on the
observer. It is also a problem that the observer’s judgments may not match the user’s
actual level of satisfaction.
Meanwhile, there are also investigations through user self-report with measures
of depression such as a scale based on facial expressions or geriatric depression scale
(Shibata & Wada, 2011). Although these do have the advantage of directly measuring internal aspects of users, if the survey respondents have dementia, there may be
cases in which the reliability of self-reports is questionable.
As for studies using physiological indicators, in an investigation by Iijima,
Shibano, Murakami, and Yamaguchi (2010), amylase activity in saliva was used
to measure stress. It was found that stress reduced after communication in the form
of exercise with a robot. A similar investigation using physiological indicators by
Shibata and Wada (2011) utilized measurement of urinary hormone levels. One
challenge of studies utilizing physiological methods is the difficulty of measurement
in real time. User satisfaction is thought to change over time with physiological
changes occurring gradually after the psychological changes and potentially
disappearing after a set amount of time has passed.
In light of the above perspectives, there are thought to be advantages to using
facial expression analysis tools through noncontact AI as a method for directly
measuring CHRs/DU s’ psychological states and satisfaction with their lives in
real time. Utilizing AI eliminates the influence of an observer’s subjective perspective and allows real-time measurement of psychological states from users’ own facial
expressions.
6.1 Advantages of Facial Expression Analysis Using AI
Previous studies regarding facial expression recognition have typically used still
images; however, as facial expressions depend on the movement of the muscles in
the face, they must be captured in a fixed time series. For example, it has been
reported that not only do the muscles involved differ between spontaneous and
intentional facial expressions but the temporal changes when parts of the face move
differ as well (Namba, Makihara, Nakao, & Miyatani, 2014; Takahashi, 2002).
Using artificial intelligence (AI), which automatically analyzes the movements of
parts of the face related to facial expression, can be used to understand facial
expression patterns in a time series.
Information Technology/Artificial Intelligence Innovations Needed for Better. . .
45
questionnaire (Obayashi & Masuyama, 2018; Obayashi, Masuyama, et al., 2018).
The involvement of the observer’s subjective judgment is an issue with methods
using observation as the judgment criteria can possibly differ depending on the
observer. It is also a problem that the observer’s judgments may not match the user’s
actual level of satisfaction.
Meanwhile, there are also investigations through user self-report with measures
of depression such as a scale based on facial expressions or geriatric depression scale
(Shibata & Wada, 2011). Although these do have the advantage of directly measuring internal aspects of users, if the survey respondents have dementia, there may be
cases in which the reliability of self-reports is questionable.
As for studies using physiological indicators, in an investigation by Iijima,
Shibano, Murakami, and Yamaguchi (2010), amylase activity in saliva was used
to measure stress. It was found that stress reduced after communication in the form
of exercise with a robot. A similar investigation using physiological indicators by
Shibata and Wada (2011) utilized measurement of urinary hormone levels. One
challenge of studies utilizing physiological methods is the difficulty of measurement
in real time. User satisfaction is thought to change over time with physiological
changes occurring gradually after the psychological changes and potentially
disappearing after a set amount of time has passed.
In light of the above perspectives, there are thought to be advantages to using
facial expression analysis tools through noncontact AI as a method for directly
measuring CHRs/DU s’ psychological states and satisfaction with their lives in
real time. Utilizing AI eliminates the influence of an observer’s subjective perspective and allows real-time measurement of psychological states from users’ own facial
expressions.
6.1 Advantages of Facial Expression Analysis Using AI
Previous studies regarding facial expression recognition have typically used still
images; however, as facial expressions depend on the movement of the muscles in
the face, they must be captured in a fixed time series. For example, it has been
reported that not only do the muscles involved differ between spontaneous and
intentional facial expressions but the temporal changes when parts of the face move
differ as well (Namba, Makihara, Nakao, & Miyatani, 2014; Takahashi, 2002).
Using artificial intelligence (AI), which automatically analyzes the movements of
parts of the face related to facial expression, can be used to understand facial
expression patterns in a time series.
Information Technology/Artificial Intelligence Innovations Needed for Better. . .
45