-
Hiroki FUJINAGA , Katsumi TADAMURA
Session ID: 20-02-01
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
The advance in the performance of mobile devices such as smartphones and tablets allow us to use computing
cost-intensive functions such as the AR with practical response time. On the other hand, there are many examples of the use of
ICT as a means of attracting domestic and foreign tourists to local tourist attractions. In particular, there is a growing trend as
follows: reconstructing historical buildings that no longer exist today with computer graphics (CG) and using it with virtual
reality (VR) to allow visitors to freely observe them in a virtual space, or to combine them with the real landscape with
augmented reality (AR). In this paper, we propose a method for synthesizing historical buildings that do not exist today with
the real world through the marker-less AR, taking into account effects brought to the real world by CG objects. A prototype of
the proposed method was installed to a smartphone and its usefulness was confirmed by being applied to a miniature of the
Hagi Castle and its surroundings.
View full abstract
-
Tomoya Wakasugi, Bisser Raytchev, Kazufumi Kaneda, Masashi Baba
Session ID: 20-02-02
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
The color of objects captured in the sea is different from the color observed on the ground due to the influences of
scattering and absorption of light. It is important to restore the color of underwater images in order to obtain the real color and reflectance
of objects. In this research, we use a spectral renderer to generate underwater images under various light sources with different spectral
distributions considering the influence of scattering and absorption of light in water, and visualize the influence of scattering and
absorption of light in water. We also investigate the color restoration using RGB images converted from the spectral images of each
light source. In spectral rendering, we use three types of seawater with different scattering and absorption coefficients. In the color
restoration process, we use two kinds of scattering and absorption coefficients: coefficients determined by the peak sensitivity and the
sensitivity distribution of RGB sensors of a camera
View full abstract
-
Haruna CHIDA, Haruka TONOYA, Motohumi HATTORI, Kazuhisa YANAKA
Session ID: 20-02-03
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Local and corporate characters are popular and are useful for public relations of local governments and companies.
Opportunities to take pictures with characters in costumes have existed for some time. However, they are only found during
events and public occasions. In addition, from the perspective of infectious disease control, taking commemorative photos in
short distance is difficult. Therefore, if a 3D model is introduced on the screen of a smartphone using augmented reality (AR)
technology, then capturing photos together with characters is easy. We created two types of AR systems, namely, using a local
manhole as a marker and using a simple reverse geocoding service of GPS. In this approach, smartphone and Unity setup are
used. In the former case, Vuforia is used as well. We examined the cases, in which each method is suitable. If such systems are
used, then taking pictures with virtual characters by using only a smartphone is possible. Thus, sightseeing satisfaction is
increased.
View full abstract
-
Yuji SAKAMOTO
Session ID: 20-02-04
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
A holo-HMD using holography is expected as the principle of the next generation HMD. This paper introduces the
basic principles, advantages, and future potential of holo-HMDs.
View full abstract
-
Yuriko Maruyama, Kota Hashimoto, Bisser Raytchev, Kazufumi Kaneda, Tor ...
Session ID: 20-02-05
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
It is desired to develop a system that automatically detect the projector position without markers and project the internal
structure of a human body onto the body surface without restrictions on the projection direction in breast reconstruction surgery using
perforator flap in plastic surgery. In this paper, we propose a method to estimate the position and pose of the camera from the image of
the human body taken by the camera attached to the projector based on the body line using a particle swarm optimization method.
View full abstract
-
Risako II, Ryosuke FURUTA, Yukinobu TANIGUCHI
Session ID: 20-02-06
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
With the increase of the scale of dairy firms and the popularity of free-stall barns, individual
management of dairy cows has been getting more difficult. For efficient management, there is a method
of installing wide-angle cameras on the ceiling; however, it makes difficult to grasp the positions of the
dairy cows in the barn as the cameras increases. In this study, we aim to create a panoramic image that
satisfies the following requirements: the dairy cows must not be cropped, duplicated or missing, and the
stitches between the images must not be noticeable. We propose a method that extracts the individual
regions of the dairy cows and synthesizes them on the underlying panorama image. We conduct a user
evaluation experiment and compare the proposed method with multi-screen and simple composite images.
View full abstract
-
Kazuki WAKIDA, Ryosuke FURUTA, Yukinobu TANIGUCHI
Session ID: 20-02-07
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In order to monitor health of dairy cows, we previously proposed a system for individual identification from
cameras installed on the ceiling of barns, which uses deep learning to identify individual cows with spot patterns as a clue. In
this system, to automate the construction of the database of spot patterns, we have developed a method for generating a 3D model
from multiple RGB-D data sets. However, the generated images from the 3D model have seams that do not exist in real cows.
To solve the problem, we propose a method for synthesizing seamless 3D models. The method defines a distance metric to
determine the neighborhood point considering the normal direction. Using the distance, we blend two 3D point clouds by
averaging nearest neighbor points and synthesize a seamless point cloud. To confirm the effectiveness of the proposed method,
we evaluate the performance of individual identification that uses the images generated by the proposed method.
View full abstract
-
Shun OUE, Yuiko SAKUTA, Makoto HIRAHARA, Shigeru AKAMATSU
Session ID: 20-02-08
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
We propose a hidden Markov model (HMM)–based method to categorize eye movement pattern in terms of
machine learning technique, which was applied to the cognitive process of impression evaluation of paintings. In our experiment,
we tracked the gaze transition pattern of subjects while they evaluated their personal impressions of traditional Japanese Ukiyoe paintings. For each class of positive/negative impressions, we estimated the HMM parameters from the training samples of eye
movements. For the eye movement data of the test samples, we conducted a classification test based on the differences of the
log-likelihood values obtained from each HMM. For the cases of receiving negative impression, we identified a common pattern
of eye movement from a broader perspective, but no such holistic patterns of eye movement were found for the cases of receiving
positive impression.
View full abstract
-
-Reflection of the latest public information about the train operation -
Hiroki TAKATA, Katsumi TADAMURA
Session ID: 20-02-09
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Japan had been steadily increasing the number of foreign tourists visiting Japan with the aim of becoming a
tourism-oriented country, before the COVID-19 pandemic. However, most of the foreign visitors to Japan are concentrated in
the so-called the golden route in metropolitan areas and Mt. Fuji, and the growth of tourists is struggling in the minor regional
cities. One of the reasons for this is the lack of multilingual displays for tourists. In rural areas where the public transportation
network is poor, obtaining information on the operation status of railway is essential for smooth sightseeing. In this paper, we
propose an application that helps foreign tourists visiting Japan obtain useful information in their native language when they
travel through a local railway. The prototype of the proposed application was implemented in a smartphone and its usefulness
was confirmed by actual use at a local train station.
View full abstract
-
Surena Kawahara, Teruaki Hirano, Daisuke Takahashi, Kaoru Nakayashiki ...
Session ID: 20-02-10
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
The first step in typing learning is to acquire a home position. However, beginners tend to type with their own
fingering, which is not the home position. Once such a habit has been developed, it is necessary to correct it with a strong
consciousness to correct it by oneself. The purpose of this paper is to fix the correct fingering according to the home row position.
To that end, we propose two systems. The first is a fingering discrimination system that automatically discriminates whether the
input is correct. The second is a typing learning game that allows subjects to easily learn correct fingering. In the experiment, 14
subjects who played the typing learning game a total of 3 times were asked to input a total of 1800 words, and the fingering
improvement effect of the subjects averaged 26.5% was measured.
View full abstract
-
Mei KODAMA, Yukino MISHIMA
Session ID: 20-02-11
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
A cached video content management method using conent priority estimation for multi-quality
video delivery systems has been studied. In particular, since it has been evaluated for the effect in the same
cache server only, there is a problem in the practical system. Moreover, the video utilization system with user
behavior has not been considered yet. Then, we propose a cached video content management method with
prediction of video usage. In this study, we focus on a management method using video usage characteristics,
e.g., video category, series of videos, and number of videos. We design the usage model based on simple daily
human behavior and preference. The simple prediction method considering the series videos is proposed in the
cache servers. By the simulation experiments, we evaluate the proposed system in each cache hit rate. The
improvement of the cached video management method for human movement is discussed.
View full abstract
-
Kazuma Saitou, Sho Ooi, Mutuo Sano, Keisuke Sasaki
Session ID: 20-02-12
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In recent years, the accuracy of alerting and detecting dangerous behavior has improved with the spread of AI,
and work accidents are on the decline. Those are caused by human errors such as lack of worker's attention and mistakes in
information transmission. There foreign workers in recent years, and it is an important issue for the system to grasp worker's
conditions. This study used video analysis technology to detect and predict abnormal conditions during factory work.
Specifically, this study simulated person whose cognitive function declined depending with or without a cognitive load task,
obtained the abnormal state by the behavior trajectories between the normal state and the abnormal state, and estimated the
abnormal state using long short term memory (LSTM).
View full abstract
-
Yun SONG, Sho OOI , Haruo NOMA, Mutsuo SANO
Session ID: 20-02-13
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Patients with brain dysfunction suffer from unstable cognitive states, usually presented as loss of attention,
concentration, and memory. Such symptoms lead to difficulties in their daily lives, while rehabilitation is often time and money
consuming. The most efficient rehabilitation for them is performing daily activities, in the case of this study, cooking, which
requires certain degree of attention and possibly has various dangerous occasions. This study obtained behaviors during cooking
in egocentric vision, used a computer vision algorithm, and proposed a model to detect dangerous moments. The model handled
frames of the cooking video with YOLOv3 and Open Pose for detection of utensils and hands respectively. Lucas-Kanade optical
flow algorithm was also implemented for concentration loss detection given that the egocentric vision could provide information
of head motions. It was also trained on a video data set recorded in egocentric vision in kitchen to measure its accuracy and
performance. Two videos were tested and the result showed increased detail and diversity of this approach compared with
previous methods. However, accuracy was still under question and optimization was required.
View full abstract
-
Sho OOI, Kazuma SAITO, Mutsuo SANO, Keisuke SASAKI
Session ID: 20-02-14
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
There are many fluid human resources such as foreign workers in the factory's working, and continuous education
is difficult. In such a situation, failures such as small mistakes to big mistakes are called near-miss (hiyari-hats) and are often
caused by human error. In particular, a small mistake in the factory working is a possibility of a life-threatening accident. As a
countermeasure for such problems, accidents can be prevented by sharing information with employees; however, the manager
cannot grasp all employees' situations or the correct information cannot be transmitted to foreign workers. This study simulated
person whose cognitive function declined depending with or without a cognitive load task, analyzed differences between two
models' behaviors, and verified detection for the abnormal behavior
View full abstract
-
Kazuaki SUGAI, Kitahiro KANEDA , Keiichi IWAMURA
Session ID: 20-02-15
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
In today's modern society, where goods are actively bought and sold between individuals, the
distribution of counterfeit products such as food packages, branded product tags, and drug labels, which
are easy to imitate, has become a serious social problem. In order to solve the above problem, we propose
a method to judge counterfeit products from commonly used inkjet printied codes. For authenticity
judgment, it is difficult to physically copy individual physical copies of printed matter of an inkjet printer
(PUF property) is used. In this paper, we propose an authenticity judgment system that combines the LLAH
system, which performs high-speed image search as a new method, and A-KAZE, which matches the
feature amount of inkjet printed matter with high accuracy, and verifies the accuracy
View full abstract
-
-In Case of Checkerboard Pattern-
Hiroshi UNNO, Kazutake UEHIRA
Session ID: 20-02-16
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
This paper describes a display technique which enables us to invisibly attach information to a displayed image on a
flat panel display by using temporally and spatially brightness-modulated patterns and to extract the attached information form
the displayed image captured with a video camera. We used checkerboard pattern as the information in this technique. First, we
assessed invisibility to the human eye of the attached information to the displayed image in terms of length of one side of
checkerboard's square. We also examined the readability with a video camera of the attached information.
View full abstract
-
Natsumi KANEKO, Masahiro ISHIKAWA, Tsutomu INOUE, Eito KOZAWA, Hirokaz ...
Session ID: 20-02-17
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Much importance has been put on the early detection and early treatment of chronic kidney
disease, which has seen a steady annual increase in recent years. An eGFR is frequently used to evaluate
renal function. An eGFR is an index showing how many waste products the kidneys can excrete into urine,
and is calculated from the level of creatinine obtained from blood tests and age. However, there is
concern that the reliability of blood tests is low because the values greatly fluctuate depending on the
timing of testing. To that end, the development of a renal function estimation method using an MRI that
can image hypoxia, fibrosis, morphological changes, etc. In this study, renal function was estimated using
the texture of renal MRI.As a result, we report the obtained results with a correlation coefficient of 0.74
and RMSE 13.69.
View full abstract
-
Ryo IGURA, Ryosuke FURUTA, Go IRIE, Yukinobu TANIGUCHI
Session ID: 20-02-18
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
We previously proposed a location estimation method from a single floor image, where similar images to the query
image are retrieved from an image database with location information. The problem is that a large amount of efforts and costs
are required for collecting floor images to build the image database. In this paper, we propose a method to automatically create
the database from videos recorded while moving on the whole floors. The proposed method sequentially estimates the location
information of each frame by calculating extrinsic camera parameters. In addition, we propose a method to mitigate the
accumulating errors from the starting positions by using reference images with correct location information and floor maps.
View full abstract
-
Xinyun LI, Ryosuke FURUTA, Go IRIE, Yukinobu TANIGUCHI
Session ID: 20-02-19
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
Due to the increasing complexity of indoor facilities such as shopping malls and train stations, there is a need for a
technology to find the current location of a user using a smartphone or other devices, even in indoor areas where GPS signals
cannot be received. As a solution, indoor localization methods based on image recognition have been proposed. Although many
outdoor localization methods have been proposed in the past, indoor localization is difficult to get high accuracy only with one
image taken by the user, because there are many similar objects (walls, desks, etc.) and there are only a few cues that can be used
for localization. In this paper, we propose an approach using (i) multi-view images with four shooting directions (front, behind,
left and right) to capture more effective features, and (ii) features obtained by aggregating CNN feature maps with Generalized
Mean Pooling(GeM Pooling) to achieve higher accuracy than conventional methods.
View full abstract
-
Atsushi IMURA, Munetoshi IWAKIRI, Kiyoshi TANAKA
Session ID: 20-02-20
Published: 2021
Released on J-STAGE: March 31, 2023
CONFERENCE PROCEEDINGS
RESTRICTED ACCESS
The contour line that traces the shape and scale of the excavated remains that the investigator
(archaeologist) actually measured in the field is called architectural drawings, and a method of automatically
drawing this is being sought. However,when the conventional edge extraction method is applied to the ortho
image (RGB image) obtained from the remains, there is a problem that it is difficult to extract the upper and
the lower edge lines of the remains with different heights. To solve this problem, in this study, we pay attention
to the linear density features obtained by applying the Canny method to the depth image of the remains, and
propose a method to extract the upper and lower edges of the remains by quantitatively expressing the density.
The effectiveness of the proposed method is verified by showing an example of applying the proposed method to
actual remains data.
View full abstract