IEEE Systems, Man and Cybernetics Magazine - January 2020 - 16

images, which provided a neural algorithm for creating
As early as 2003, Freeman et al. [5] proposed an algoartistic images.
rithm to translate the lines of a painting into any style that
users want. The authors used a simple and traditional linSystem Architecture
ear-fitting interpolation method for training and testing.
The system architecture includes five modules: a user data
In [6], the authors developed a body interactive creacquisition module, a historical creation data set, a landative game called "Word Out." This game is aimed mainscape data acquisition module, a
ly at 4-7-year-old children to help
cloud processing module, and an
them learn to recognize and spell
artwork publishing module. First,
letters independently while playduring the system design process,
ing the ga me, stimulate their
The machine has
we used an expressive robot of
active learning, and fully tap into
creativity similar to
brain-wearable devices and weartheir imagination and creativity.
able clothes to collect data. Second,
During the game, the alphabet
that of humans for
we collected different styles of
first appears on the screen. After
drawing paintings and
paintings created by users and
t he user select s a word, t he
creating artwork.
employed CNNs to extract the style
body's contour is matched with
features of paintings [14]. The third
the words on the screen when the
part, i.e., the landscape data acquiuser manipulates the pose. The
sition module, is utilized when
game can also be completed by
users are outdoors and do not want to create. They can then
the cooperation of multiple people working together as
use the camera to take pictures, and the system extracts the
creatively as possible.
content features of the artworks. Then, an AI algorithm was
Chatain et al. [7] designed an application for users to
deployed in the cloud. After completing the intelligent decidesign and create games themselves. First, users draw on
sion fusion and creating artwork in the cloud, the results
paper the elements of the game that they want to develop,
were sent to the intelligent terminal and displayed. This sysincluding which types of instances the game has. Then,
tem architecture is shown in Figure 1.
they take a picture of the paper through the tablet camera
and integrate it into the application. Next, users can design
the level, and the event-based visual programming lanData Collection of Brain- and
guage is used to define the logic and rules of the game.
Body-Wearable Devices
Finally, users can test their own curated and designed
games. At any stage of the game's entire development proEEG Data
cess, users can skip or test to fully utilize their imaginaFor the CreativeBioMan creativity game, we used an indetion and creativity.
pendently designed, brain-wearable device to collect users'
The related work of the creative games introduced above
brain wave data. In our research findings, varying neural
can enhance the creativity of users, but these games do not
activity generated a different brain wave pattern [13]. For
take into account the users' emotional attributes and perexample, when a human was idle or in a state of artistic cresonality characteristics. With the development of deep
ation, the brain wave was a theta wave at a frequency
learning and emotional computing, the characteristics of
between 4 and 8 Hz.
more users are integrated into the creative game, and additional complex deep learning algorithms are adopted to
Real-Time Artistic Creation
make the creative game system more intelligent, which has
As for the drafting of real-time artistic works created by
attracted our attention.
artists, in this article, it is called the content feature data of
Conversely, for their research methods, Fink et al. [8]
artworks. The drawing board configured for the wearablecollected EEG signals of professional dancers during
clothing expressive robot can be used to collect data in real
improvisational dancing. They observed that, during the
time. Accordingly, the image data can be transmitted to the
generation of alternative uses, professional dancers
cloud in real time. The wearable-clothing expressive robot
showed stronger alpha synchronization in posterior pariis a good tool for recording the creativity of artists who
etal brain regions than did novice dancers. Fortino et al.
have divergent thinking, are good at improvisation, and
[9]-[11] proposed a completely new architecture that supwant to capture inspiration at a certain moment. When artports the development of novel, smart wearable systems
ists travel and are inspired to paint a picture but there are
for cyberphysical, pervasive computing environments.
no painting supplies available, the wearable-clothing
And, in [12], the authors introduced an augmented convoexpressive robot can complete the creation for them.
lutional neural network (CNN) architecture that bridged
the gap between generative algorithms and pixel-labeling
Multimodal Emotion Data
NNs. Gatys et al. [15] used neural representations to sepaAn affective interaction through wearable computing
rate and recombine the content and style of arbitrary
and cloud technology (AIWAC) box is an embedded
16

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE Janu ar y 2020

IEEE Systems, Man and Cybernetics Magazine - January 2020

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - January 2020