-
Takuma YOSHIZAWA, Takayoshi MORI, Tsukasa KIKUCHI
Session ID: 19-04-001
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
It is very important to express bubbles and splashes in water to enhance the realism of water in CG animation.In
this research, bubbles are generated and rise by the wake of an object falling when it falls into water In addition to modeling the
phenomenon, we propose a procedural animation method that enables motion control in which the generated bubble rises while
imitating the shape specified by the user. We propose that it is possible to generate an animation that can control the rising motion
of bubbles that can be applied as one of the effects in the entertainment.
View full abstract
-
Yuki Kato, Tsukasa Kikuchi
Session ID: 19-04-002
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
This study proposes a procedural animation method of instantaneous state transition by cooling water. The fluid
simulation of this method is performed using the FLIP (Fluid-Implicit Particle) method, in particular, using the
viscous term of the Navier-Stokes equation, which is the governing equation, and updating the temperature change
at each time step. To stop the movement of the fluid and create a stationary state.
Temperature data is updated by using volume data (information group with volume) and particle data (information
group of point aggregates. Temperature transition is added to both data, and state transition with free shape is
possible by interference. Become.
In addition, by allowing only a part of the fluid to interfere, it is possible to represent a state in which solid and
liquid coexist in one simulation.
View full abstract
-
-Light Painting drawn by Traces of Figure Skating-
Noriki AMANO
Session ID: 19-04-003
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
We are studying the visualization of sports and exploring its art. As a part of this, we will visualize figure skating
and aim to create a light painting based on it. Specifically, the skating of figure skating is simulated by a drone equipped with
LED lights, and a light painting with the sky as a canvas is created based on the trajectory. For this reason, a drone program
that simulates the technique of figure skating, such as a triple axel, is implemented and combined. This research is a unique
attempt to fuse sports and art, not only to expand the application area of drones, but also to propose a new expression method
of visualization.
View full abstract
-
Kai LENZ, Masataka Ishizuka, Masanori KAKIMOTO
Session ID: 19-04-004
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In video works, the time series of the world in the story and the order of depiction may be different. There are
several techniques for describing time series changes, and these are important elements that characterize a work. In this study,
we analyzed many works with time-series changes, and derived four types: reminiscence type, repetition type, group drama
type, and modified type. Then, we devised a visualization method of the time series composition, and succeeded in visualizing
the work features through the time series visualization graph.
View full abstract
-
–Distortion correction by cube maps–
Shunsuke KIRA, Masanori NAKAYAMA, Issei FUJISHIRO
Session ID: 19-04-005
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Deformation plays an essential role in expressing a large variety of materials in the reality. Due to
the progress of computer graphics, deformation simulation also needs to be more accurate. In this study, we propose
a new method of deformation simulation on an arbitrary closed surface using displacement mapping. We divert cube
maps to displacement mapping so as to correct distortion between the texture space and the physical space. In this
article, we simulate wave propagation as an example of the deformation.
View full abstract
-
Kenta Yamaguchi, Youngha Chang, Nobuhiko Mukai
Session ID: 19-04-006
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In this paper, we propose a method to realize falling and breaking down of flowers in a Ikebana simulator. Falling down of
flowers is judged by comparing the gravity on the flower and the friction between the flower and pins. On the other hand, breaking down of
flowers is judged by comparing the elastic limit of the flower and the stress, which is calculated by the bending of the flower.
View full abstract
-
Mayuka TSUJI, Yoshihiro WATANABE, Takenori ISHIKAWA, Daiki KIJIMA, Kei ...
Session ID: 19-04-007
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
When cosmetics are applied to the skin,it is difficult to distinguish the applied area from the unapplied area with
the naked eye,resulting in problems such as uneven application. Therefore,in this study,focusing on the fact that the intrinsic
spectrum differs depending on the substance,spectral information acquired at 0.5 nm intervals between wavelengths of from
400 to 720 nm is used to calculate the difference between the spectrum of the skin and that of the cosmetic,and an appropriate
wavelength is searched for in the visualization of the portion to which the cosmetic is applied. In addition,real-time coating
situation imaging is enabled by selecting three wavelength regions.
View full abstract
-
Hidekazu TSUJIAI
Session ID: 19-04-008
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
I thought about using less equipment to reduce the complexity when displaying the projection mapping. Although
a projector is required, we tried to reduce the number of equipment and reduce maintenance by using a PC. In fact, I report the
results of PC-less projection mapping at Mitsui Outlet Park Hokuriku Oyabe at two locations in the hall and in the gondola of
the Ferris wheel. In particular, in the construction of the ferris wheel, the fact that it could be set up without stopping the ferris
wheel could expand the range of use of the projector.
View full abstract
-
Shu Ichinose, Yuta Kurino, Shunsuke Suwa, Tsukasa Fujikawa, Kanae Wada ...
Session ID: 19-04-009
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Conventional interactive projection mapping has a disadvantage that is “difficulty of understanding the relevance
to users’ behaviors and corresponded projection images.” The disadvantage is due to misunderstandings between users’ mental
models and producers’ ones. Therefore, we introduce users’ natural behaviors based on their mental models into our interactive
effects of projection mapping. Our solution may solve the above problem. Especially, we verify the effectiveness of our
interactive projection mapping through a practice event at a Japanese shrine.
View full abstract
-
Chishun KANAYAMA
Session ID: 19-04-010
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
We propose a virtual calligraphy system that enables soft tufts of the writing brush and the faint expression of
brushstrokes due to the change of the ink quantity contained in the brush tufts. In this system, a pen tablet that supports tilt
detection is used for handwriting input. The virtual brush tuft consists of seven bundles of hair that divide the center and its
periphery into six, and each hair bundle is individually deformed and moved by pen pressure and movement. Sampling points
for transmitting black ink from brush tuft is arranged in the hair bundle area in contact with the paper surface, and the locus of
the point moved with the deformation and movement of the hair bundle is drawn by a line segment. The brightness of the line is
changed according to the change in the ink quantity. A threshold value is set for the brightness of the drawn line, and brushstroke
is generated by drawing pixels having a density equal to or higher than a threshold on a bitmap image assuming a paper surface
in black.
View full abstract
-
Kaito TANAKA, Ryota ONISHI, Hiroaki SAWANO, Seiji HOTTA
Session ID: 19-04-011
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
A video content, adding a movement and a sound source to a comic image, is known as a motion
comic, and a video production technology is necessary. The authors have developed an automatic creation
method of the motion comic with an image processing, and the timing of switching panels has been manually
determined. In this study, we propose a panel switching method based on a tempo feature of the sound source.
We analyzed tempos of three sound sources, and created the motion comic based on analysis results.
View full abstract
-
Ken AOKI, Yumeka ISHII
Session ID: 19-04-012
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Today, text based short messaging tools are using widely on SNSs. But sometime emotional expression is not
transmitted correctly because difference on the nuance of words. So, this paper proposes a emotional expression support tool for
text which adds an emotional expression to the text by selecting a speech bubble. For recommending a speech bubble which fits
the emotional expression of the text, we constract 3-D emotional space coordinate and fitness function for evaluating the fitness
of a speech bubble and the coordinate value
View full abstract
-
Michitaka NAKAI, Masanobu ENDOH
Session ID: 19-04-013
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
A sense of exhilaration is one of the evaluation terms often used in game play. We suggested that players will
feel a sense of exhilaration if they can know a successful experience. We implemented the two versions testing game and
conducted an experiment on a successful representation and sense of exhilaration. As a result, players who played the game
with a successful representation felt more exhilarating than players who played the game without a successful representation.
View full abstract
-
Yusuke MUKAI, Hiroki TAKAHASHI
Session ID: 19-04-014
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In AR technology, it is important to synthesize virtual objects in real space so that users perceive the synthesis
image as free of artificiality. Especially, it is the most significant to keep their optical consistency because it can reduce
artificiality of the image. In this paper, we estimate external camera parameters from a single image using a cylindrical object
and propose a method for estimating a light source direction in three dimensional space by using inverse perspective projection
transformation. In addition, we estimated light source directions using the proposed method in a real environment and discussed
the merits and demerits of the proposed method and a conventional method by comparing estimation accuracies of both methods.
View full abstract
-
Riku NAKAZATO, Yuki ENDO, Yoshihiro KANAMORI, Jun MITANI
Session ID: 19-04-015
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
The hole problem in origami is, given a crease pattern with an area with the creases missing in it, a problem to
consider how to fill in the area with suitable creases. An algorithm for solving the hole problem is known. However, crease
patterns output by the algorithm have not been enumerated. In addition, when flat-foldable origami is targeted, it is not
guaranteed that the crease patterns output by the algorithm are capable mountain-valley assignments that satisfy conditions of
local flat-foldability (valid MV assignments). In our study, we have enumerated crease patterns output by the algorithm, and
have implemented a function to determine whether they have valid MV assignments. As a result, we have enumerated crease
patterns output by inputting polygons with four and six vertices and have made it possible to determine whether those crease
patterns have valid MV assignments.
View full abstract
-
Kaoru OHASHI, Yuki ENDO, Yoshihiro KANAMORI, Jun MITANI
Session ID: 19-04-016
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Curved folding is a way of folding a sheet of paper along a curved crease. By using curved foldingƅone can make
beautiful shapes which have curves and curved surfaces. A shape created by curved folding is composed by straight lines called
ruling. Because this ruling has an important property that they do not intersect on a sheet of paper, existing system to design
curved folding requires users to avoid this intersection manually. In this paper, we introduce a method to avoid intersection of
rulings automatically. We define an objective function that outputs degree of intersection and a penalty function which keeps the
shape from large deformation. By solving these functions as an optimization problem, we implemented a system which helps us
to design curved folding more quickly and easily
View full abstract
-
Haruka TAKAHASHI, Yuki ENDO, Yoshihiro KANAMORI, Jun MITANI
Session ID: 19-04-017
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
We propose a method for estimating depth from single-view terrain landscapes using convolutional neural
networks (CNNs). The existing depth estimation methods using CNNs mainly target at near-distance views (e.g., indoor scenes),
and often suffer from large-scale scenes with significant depth range. We attribute this performance degradation to the difficulty
of training due to considerable depth variations. In this work, we integrate a CNN for classifying depth distributions inferred
from input images on top of depth estimation to alleviate the influence of depth variations during training and inference. We first
classify training data into several depth classes and train a depth-estimation CNN for each class. During inference, we obtain a
final depth map by calculating a weighted average of outputs from class-wise CNNs, whose weights are obtained as likelihoods
inferred by the depth-classification CNN. We attempt more accurate depth estimation for terrain landscapes with this proposed
framework.
View full abstract
-
–Character recognition of seal scripts–
Shohei NINOMIYA, Masanori NAKAYAMA, Atsushi MIYAZAWA, Issei FUJISHIRO
Session ID: 19-04-020
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Calligraphic scripts are classified broadly into five types: seal, clerical, cursive, running, and standard. In this
study, we propose a convolutional neural network–based method for character recognition of seal scripts. Utilizing the feature
that kanji is composed of parts such as the radicals, we can recognize unreadable characters used in calligraphic works through
image recognition of each part of the characters. Our method provides a general purpose processing environment of character
recognition with small numbers of data and classes.
View full abstract
-
Shoma KATO, Hiroaki SAWANO, Akira KITO, Norio ISHII, Yuri SUZUKI
Session ID: 19-04-021
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
It is important to analyze a service in a table tennis, and a table tennis player confirms a service
hitting style and a ball course by an opponent with a video. Since the tendencies of the service and the course
are different in a game situation, it is necessary to watch all of the video and to record one. In this study,
we propose a method for estimating a score transition in a broadcast table tennis video. A score region in a
scoreboard is detected with an image processing, and the score transition is estimated with a brightness change
in the detected region. We evaluated the proposed method with five table tennis videos. Experimental result
shows that the accuracy of the score transition detection is 90 %.
View full abstract
-
Moe TSUJI, Tsukasa KIKUCHI
Session ID: 19-04-022
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Spread of the Internet and smartphone make it easy to get information on the internet and watch videos.
Meanwhile, “1-minute video” is in fashion among social media. Nevertheless, the selection method of information in terms of
1-minute video has not been clarified. The purpose of this study was to clarify what the selection method of information in
terms of 1-minute video. Besides, it is applied for video using infographics. Infographics is good at transmitting information.
Finally, we propose a new method of transmitting information
View full abstract
-
-Sensitivity Analysis of Multi-Agent System to Space Structural Metamorphosis-
Mitsuhiro ODAKA
Session ID: 19-04-023
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
-
-Digital Short Film-
Ryuichi TAKAYAMA
Session ID: 19-04-024
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Short Film for Digital System
View full abstract
-
Kazuyuki MIYAMOTO, Dai HIJIKATA, Koji IIGURA
Session ID: 19-04-025
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In this presentation, we introduce the art-informatics, which digitizes human activities in the art field and
analyzes them using the information technology. The art-informatics is not a well-established field, but is a research field / area
that the authors are aiming to launch. In this paper, we will introduce it and report the results of preliminary experiments
conducted to explore the possibility of launch. These experiments targeted at a number of lectures and research meetings
organized and held by people involved in Akita University of Art , to which the authors belong.
View full abstract
-
Takashi IMAGIRE
Session ID: 19-04-026
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
GitHub is a famous version control service in software development. It seems to be widely used in class, but in
fields related to image information, the results are often visually important, so there is much leeway for fully utilize the GitHub
than general software development. Here, we will explore the deepening of issues and progress management methods in CG
programming education by sharing the utilization strategies of GitHub used in multiple lectures over multiple years..
View full abstract
-
Daisuke MIYAZAKI, Osamu YUUKI, Kohe TOKOI, Norimasa YOSHIDA, Yoshihiro ...
Session ID: 19-04-027
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In the last few decades, CG technology has evolved significantly. A wide variety of technologies have been
developed, and it is becoming difficult to keep up with technological advances. Therefore, it is necessary to systematize existing
CG technology comprehensively and to organize knowledge and technology. Educating the next generation of CG engineers
using our collected knowledge is important for further development of CG technology. In this report, we introduce how we have
collected CG knowledge and how we have gathered them as a document. The document will be published as a book so that many
people can use it for CG education.
View full abstract
-
Shun OUE, Moeka HENMI, Ai MORIKUBO, Akito YAMADA, Yuko TASHIRO
Session ID: 19-04-028
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
We report the 19th Visual Information Processing Camp (VIP Camp) held from 21th to 23th, September 2019.
The VIP Camp is taken place for the students who study visual information processing and other related research fields. The VIP
Camp is planned and operated every year by voluntary university students from all over Japan. The primary purpose of the Camp
is to provide opportunities for practicing their research presentations and face-to-face discussion on them. The catchword of the
Camp this year is “Next Stage”. As represented by the catchword, we arranged poster sessions with fast-forward presentations,
and a group work session in which the participants discuss research papers presented “What is a good poster presentation?” In
this paper, we report the organizing activities for the 19th VIP Camp and discuss possibilities to improve the future VIP Camps
by analyzing the answers in participant questionnaires.
View full abstract
-
Taisei INOUE, Norimasa YOSHIDA, Motonori ISHIBASHI
Session ID: 19-04-029
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Measuring human characteristics quickly and easily is important in self-improvement and performance
improvement. In previous work, we showed the relationship between a line drawing task based on a point cloud using a
tablet device and the degree of action slip (difficulty in attention control). However, what kind of point clouds can be used to
classify the degree of action slip is not clear. In this study, we describe a plan for extracting specific features of point clouds
that can be used to classify the degree of action slips.
View full abstract
-
Kazuki Nishikawa, Toshiyuki Imai, Kohe Tokoi
Session ID: 19-04-030
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In playing cards, cards are shuffled for randomization. Randomization in playing cards means that when removing cards,
all cards means that when removing cards, all cards are selected with equal probability. However, randomization is not
performed sufficiently, such as predicting the card to be removed next if the shuffle method is bad.
The purpose of this study is to evaluate whether randomization of shuffle is being performed. Simulate the four major
shuffles on a computer and try them individually or in combination to analyze their characteristics.
The randomized sequence is based on data obtained by rearranging cards by random numbers. Compare this with sorted
data based on major shuffles to assess whether randomization has taken place.
View full abstract
-
Moe TSUJI, Tsukasa KIKUCHI
Session ID: 19-04-031
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Spread of the Internet and smartphone make it easy to get information on the internet and watch videos.
Meanwhile, “1-minute video” is in fashion among social media. Nevertheless, the selection method of information in terms of
1-minute video has not been clarified. The purpose of this study was to clarify what the selection method of information in
terms of 1-minute video. Besides, it is applied for video using infographics. Infographics is good at transmitting information.
Finally, we propose a new method of transmitting information
View full abstract
-
Naoto ENDO, Subroto Prasetyo, Tsukasa KIKUCHI
Session ID: 19-04-032
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
About mosaic processing technique by the facial recognition using the machine learning
View full abstract
-
Taiki Shimomura, Takashi Kanai
Session ID: 19-04-033
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In the liquid simulation, the density or velocity information stored as a volume data can be visualized by volume
rendering. However, a time-varying volume data tends to be spatially and temporally very large. In this paper, we proposed a
spatio-temporal compression method by introducing motion compression prediction in addition to the conventional VDB method.
We evaluated a compression radio and compression error by using PSNR and showed that our method can decrease compression
error by adding differences created from motion compensation prediction. However, we found that our method caused a worse
compression ratio than VDB when extracting all differences. We then improved our method not to extract vertices which changes
slightly, and showed better compression ratio than VDB by such the improvement
View full abstract
-
Akari Abe, Masaki Abe, Taichi Watanabe
Session ID: 19-04-034
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Empathy is one of the elements required to maintain immersion feeling in content. A player is immersed in
events in the game by feeling the emotion of game of character. Recent video games require NPC to behave like human being.
Empathy may be impaired by deviant behavior of the NPCs. The purpose of this study is to maintain the empathy felt by player.
We propose a method to maintain the consistency of actions of NPC action by constructing the human relationship from both
player’s action and action outcomes. The proposed method determine NPSs behavior by giving NPC a psychological model
and human relationship
View full abstract
-
Natsuki YAMAMOTO, Masaki ABE, Taichi WATANABE
Session ID: 19-04-035
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In recent years, with the spread of smartphones, many people use smartphones every day.
In many cases, a smartphone is operated using a touch UI, but a Fat-Finger problem occurs in which an
adjacent touch UI is erroneously pressed. In addition, various Fat-Finger issues have been studied, and
there have been many studies on smartphone software keyboards and smart watches, but few studies on
smartphone websites. This study uses a magnifier and pressure detection on a smartphone website, solves
the Fat-Finger problem, and proposes a magnifier that is easier to handle than existing magnifiers.
View full abstract
-
Takumi EBARA, Koji MIKAMI, Yoshihisa KANEMATSU
Session ID: 19-04-036
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
The quality of spatial projection of 3D stereoscopic images called holograms and spatial projections has been
improved due to the development of projection methods and devices. On the other hand, methods for expressing high-quality
projected images from multiple directions still have many issues. Therefore, in this research, we extended the method using
half-mirror that can view projected images from multiple directions, and proposed a method aimed at improving the quality of
projected images using screen image synchronization of multiple tablets. We implemented a system using Unity and Photon,
and compared it with the content by the existing method and obtained good results.
View full abstract
-
Maho FURUKAWA, Masaki ABE, Taichi WATANABE
Session ID: 19-04-037
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In recent years, game AI is regarded as important. We focused on goal-based AI used in many works. In this
method, a long-term goal is determined first, and actions are constructed from the goal. However, this method needs to reconstruct
the behavior when the situation changes, and if it is reconstructed frequently, the optimal behavior cannot be obtained as a result.
On the other hand, there is a utility base as a method to select the optimal action by local situation evaluation. However, this
method only performs local situation evaluation, and has a problem that it is extremely inefficient for achieving a long-term goal.
In this study, we propose a method that combines goal-based AI and utility-based AI with the aim of realizing efficient
achievement of long-term goals that can respond to local situation changes.
View full abstract
-
Rena AIZAWA, Akinori ITO
Session ID: 19-04-038
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In this research, we analyzed music videos of “Maji LOVE 1000%”, “God's S.T.A.R.”, and “Immortal Inferno”
in “Utano☆Princesama”. In the analysis, the components of video and music were classified into attributes such as Degree
name, lyrics type, and characters appearing in the video into categorical data, and related rules were extracted using apriori's
algorithm.
View full abstract
-
Kiyoka YAMAMOTO, Masaki ABE, Taichi WATANABE
Session ID: 19-04-039
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Video games are becoming use vast terrain due to improving hardware performance. The subdivision process
increase the number of polygons, however it takes time when the camera distance is suddenly get close. In this research, we
propose a method for polygons division processing and texture dynamic generation. It calculate the area that needs to be
divided based on the camera distance.
View full abstract
-
Seiya Sakurai, Norimasa Yoshida
Session ID: 19-04-040
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In conventional deep learning, it is difficult to obtain the desired accuracy unless the time interval is constant.
A method called ODE-Net (Ordinary Differential Equation Network) that can adapt to data with an irregular time interval has
been proposed. In this study, in order to examine the performance of ODE-Net, the logarithmic spiral used in the previous
study is extended to log-aesthetic curves, and the approximation is performed with conditions such as how to sample the curve
and from which region to perform approximation. The results show that a good approximation can be obtained by learning
log-aesthetic curves toward approaching a circle, and by using arc length instead of tangential angle as a parameter.
View full abstract
-
Sayaka USHIMARU, Narumi ARAYA, Hidekazu TSUJIAI
Session ID: 19-04-041
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Hybrid image projection mapping means a technique to reflect a picture to a solid thing using a projector by the data which I
made with a computer. Using a technique of the projection mapping, I reflected Hybrid image this time. The hybrid image means the image
which a viewpoint turns into by a perspective. We can enjoy the difference of an image to look at near and the image to look at from far away.
I performed projection mapping in Mitsui Outlet Park of Oyabe city and actually reflected several kinds of images using a projector
on the wall in the shop.
View full abstract
-
Tomoki KIZAWA, Yuki ENDO, Yoshihiro KANAMORI, Jun MITANI
Session ID: 19-04-042
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Many image segmentation methods formulate the problem of foreground extraction as an eigenproblem of a
Laplacian matrix defined by the input image, as done in the classical approach, Normalized Cut. The problem here is that the
computational cost of such eigenproblem increases rapidly as the resolution of the input image becomes large. In this work, we
accelerate Laplacian-based image segmentation by separately solving small eigenproblems defined by one-by-N image strips
that are extracted in horizontal, vertical, and diagonal directions in the input image. Their eigenproblems can be solved fast
because the resultant Laplacian matrices are tridiagonal. We further subdivide the image strips according to the consequent
eigenvectors into segments and assign fore-/background labels to the segments in order to obtain a binary mask. We demonstrate
that the proposed method performs fast, even for high-resolution images.
View full abstract
-
Yuya Horikoshi, Tokiichiro Takahashi
Session ID: 19-04-043
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
We have realized an expression method that draws the outline of a fast-moving object vaguely, which is often used
in comics. In our expression method, particles are placed on each surface of the 3DCG object, and are generated at regular
intervals. Our method draws speed lines placed at the contours of the 3DCG object by connecting the generated particles with
straight lines. Then, our system draws the contours by using the back-face method. By superimposing and synthesizing the speed
line and the contours, the outline of the object moving at high speed could be drawn vaguely. The evaluation results showed that
our expression method was useful and competitive to the conventional methods using in comics.
View full abstract
-
Saori Okamoto, Hisakazu Hada
Session ID: 19-04-044
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
During playing a game, when the character you control falls from a high place or is pulled
by something and moves at high speed, there is a feeling of floating and pulling over the screen. When
performing such an action, I think that the sense of presence may be improved by actually blowing the
wind. The aim is to create a simple game in which the character falls in Unity, and use a fan to blow the
wind while the character is falling, thereby improving the sense of reality
View full abstract
-
Akihiro Tomaki, Masanobu Endoh
Session ID: 19-04-045
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
One of the elements that makes the game fun is sense of agency. That is a feeling that player's actions are
consistent with the results. In online games, there is a Lag that the operation and result do not match due to communication
delay. At this time, we thought that the sense of agency was lost and the fun of the game was lost too. We implemented an
experimental game and verify the relationship between lag and sense of agency. As a result, the player felt discomfort due to
the lag, and the recognition of the delay occurred at the lag during 7.9 to 14.8 frames with a 95% confidence interval.
View full abstract
-
Yike SUN, Takayuki ITOH
Session ID: 19-04-046
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Impressions of human face images are deeply related to the geometric features and skin colors of the faces. We
are focusing on the skin color and analyzing the relationship between the skin region of the face images and their impressions
that viewers had. Specifically, we are challenging to discover the correlations between the clustering results of human face
skins and their impression evaluation results. We generate fictitious human face images generated by a generative machine
learning technique and applied a region segmentation technique based on RGB color spaces to extract face skin parts from
these face images. Then, we calculate the color distribution of the images and cluster these images by applying the K-means
method. We introduce the technical details of the above processes and describe our plan on the subjective evaluation of the
skin images by applying the SD method with a five-point Likert scale to find correlations with clustering results.
View full abstract
-
Maaya HIGUCHI, Yuma YAMAZAKI, Hisakazu HADA
Session ID: 19-04-047
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
This installation is created with the aim of re-recognizing that the person who usually communicate on
SNS is a real human. When a visitor communicates on SNS, they can feel the gesture and reaction of the partner from
the waving of the curtain in front of them. Furthermore, they can feel the sign of the partner behind the curtain. The
movement of the curtain that created by this communication is produced by sending wind to the curtain.
View full abstract
-
Takuma UNO, Hironori HARA
Session ID: 19-04-048
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In previous research, it is clear that a 3D model in a virtual space feels a personal space. In the preliminary
experiments, the author always stipulated that habituation with 3D models was related to personal space. Therefore, in this
research, We investigated what effect habituation with 3D models had on personal space. Experiments show that subjects
who have not familiar the 3D model have more personal space.
View full abstract
-
Shoko Banzawa, Tsukasa Kikuchi
Session ID: 19-04-049
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
People face many illusions in everyday life. We tried to apply the expression by Barber pole illusion using 3DCG.
Using the characteristics of 3DCG, we looked for the part that caused the illusion while looking at the object from multiple
sides, and compared the differences in appearance by changing the conditions. As an application, we created a comparative
video about the visual effect of the Barber pole illusion on objects in the same screen. We evaluated the produced video and
verified whether the illusion was actually given.
View full abstract
-
Mizuki HASHIMOTO, Tsukasa KIKUCHI
Session ID: 19-04-050
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Recently, the movie came to be used in various scenes such as the elevators of projection mapping and attraction
facilities. Such a thing, it is thought that it is a purpose to please the person who saw it or to let you feel comfort. It is this
article, a purpose is to become able to express an abstract impression of “the comfort” intentionally. At first I collect design
elements about the movement from a movie work. I clarify the combination of design elements about “comfort” by using the
rough sets. I analyze a tendency from a design element used for those combinations.
View full abstract
-
Sachiko ICHIKAWA, Keiko OCHI, Yasunari Obuchi
Session ID: 19-04-051
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In this paper, we focused on a chalkboard scratching sound and investigated how its impression changes by the
context effect. To generate the context, we used familiar classical music and contemporary music with dissonant sounds. We
conducted a listening survey about the impression of uncomfortable sounds between those context sounds.
View full abstract
-
Naoya KINOSHITA, Keiko OCHI, Yasunari OBUCHI
Session ID: 19-04-052
Published: 2020
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
It is required to understand the theories of timbre making to use a synthesizer. In this paper, we propose a system
that searches a timbre by natural language, without using technical terms of the synthesizer. We conducted a set of experiments
to evaluate sounds of various wavetables using adjective pairs. Using the evaluation results, we developed a system that
searches a timbre from any word. Word2Vec converts the input word into a set of values representing similarity to the
expressive words. By calculating the similarity between the values of the input and the values of the wavetable, the system
finds the most proper timbre.
View full abstract