-
Kohki Kurokawa, Kyoko Hasegawa, Yuichi Sakano, Roberto Lopez-Gulliver, ...
Pages
33C-4-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We study human's shape recognition when observing transparent objects using stereoscopic vision with binocular parallax and
motion parallax. We confirmed that simultaneous use of the two parallaxes improves the recognition.
View full abstract
-
Koki Sakamaki, Haruki Mizushina, Shiro Suyama
Pages
33C-5-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
W
e have proposed a new transparent DFD display by changing the combination of the image plane
s
to be fused
and evaluated the perceived depth dependence
s
.
Our
proposed
transparent
DFD display can provide widely changeable
perceived
depth
s
of transparent im
age
s
with
various combinations
of layered displays
.
View full abstract
-
Pages
33D-0-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
-
Yuichi NAKAZAWA, Kiyotaka FUKUMOTO, Yoshinobu EBISAWA
Pages
33D-1-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
To enlarge
a
measurable
angle
range of
a
line of sight
in our precise remote gaze detection system based on
the
corneal reflection
-
pupil center,
we propose the
gaze detection technique that
add
ed the gaze direction detection method based on
the
pupil
ellipticity
to
our conventional
platform
.
View full abstract
-
Yuhei MIYAMOTO, Yuki MIYAMOTO, Kiyotaka FUKUMOTO, Yoshinobu EBISAWA
Pages
33D-2-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We have already developed a long-distance gaze detection system allowing a movement of a single user in a wide
depth range of 2.5 - 3.5 m and in a horizontal range of ±
0.5 m at a depth of 3.5 m. In the present study, we enabled a wide range
of region of gaze detection (94 cm by 86 cm) and a simultaneous gaze detection of two subjects.
View full abstract
-
Kiyotaka FUKUMOTO, Yusuke SONE, Yoshinobu EBISAWA
Pages
33D-3-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
In our previous study, the automatic lighting power control method of the ring type of near-infrared LED light
source for robust detection of the pupil in the difference image of the bright and dark pupil images was proposed for robust gaze
detection. In the present study, we developed the light source having two LED rings for bright pupil image generation and the
power control method of the source. They improved the robustness of corneal reflection detection especially.
View full abstract
-
Naoto OZAKI, Shingo HONDA, Toshimitsu TANAKA, Yuji SAGAWA
Pages
33D-4-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
A
Hiragan
a character is input based on the Japanese syllabary table.
T
he row of the table is sel
ected by a stroke
gesture, and
the column is by a tap.
These operations are executed at edges of a display of a smartwatch in order to reduce
occupation area of the syst
em.
View full abstract
-
Kotaro SAKA, Toshimitsu TANAKA, Yuji SAGAWA
Pages
33D-5-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We have been developing the c
haracter input method for smart
watches.
A user can choose a character by one stroke gestures
and one tap, however, the screen is often hidden by user’s finger during the operations.
To eliminate this problem, we devel
oped the touch
board optimized to the method.
View full abstract
-
Pages
34B-0-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
-
Ryo NAKAO, Yuji TESHIMA, Osamu SHIKU, Ken-ichi KOBORI
Pages
34B-1-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
Point Clouds which are
created using
3D
-
Scanner contain a lot of defects due
to the measurement environment
.
In this study, we
propose
a method of detecting defects
with
Point
Morphology Operations.
We verif
y
the effectiveness of
the
method by
some
experiment
al results
.
View full abstract
-
Yuta YAMABE, Shuhei KODAMA, Tomoaki MORIYA, Yuki MORIMOTO, Tokiichiro ...
Pages
34B-2-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
A painterly style image generation method based on scalable Poisson disk sampling method (SPDS method) has
been proposed. However, it required long computational time and also its results are biased. We already proposed a fast sampling
method by paving disks with random radii. In this article, we evaluate stochastic characteristics of our method. Experimental
results verified our method arranged disks randomly but uniformly compared with the previous SPDS method.
View full abstract
-
Shiori OBANA, Kazuma SHINODA, Madoka HASEGAWA, Hirotsugu YAMAMOTO, Mas ...
Pages
34B-3-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
Near-infrared spectroscopy is utilized for the sugar content measurement of fruits. However, only a few studies focusing on the non-destructive analysis using a hyperspectral camera have been conducted to date. In this study, we predicted the sugar content of apples by using a hyperspectral camera without damaging the apple surface.
View full abstract
-
Ryosuke ISHIZUKA, CHANG Youngha, Nobuhiko MUKAI
Pages
34B-4-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We have performed an aortic valve simulation based on blood flow velocity, which uses a particle method and
analyzes the behavior of particles by solving Navier-Stokes and Cauchy-Navier equations for blood and the aortic valve,
respectively. The simulation uses the velocity of blood flow as the input and calculates the pressure inside the blood vessel as
the output.
View full abstract
-
Naoyuki TANAKA, Toshimitsu TANAKA, Yuji SAGAWA
Pages
34B-5-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
In this research, to represent a windshield in the heavy rain to separate the windshield to mesh, adding up the
rainfall for each partition. In addition, to implement the drainage by the wiper to provide the mesh to move along with the
wiper in front of the wiper.
View full abstract
-
Pages
34C-0-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
-
Yifei ZHANG, Hiroshi WATANABE
Pages
34C-1-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
This paper presents experiments on trajectory data mining from subjective shot sports videos based on SLAM
(Simultaneous Localization and Mapping) and wearable camera calibration. The experiment result shows adequate robustness
to most common motion clutter in the outdoor long distance sports videos.
View full abstract
-
Tomoki OHOKA, Keita TAKAHASHI, Toshiaki FUJII
Pages
34C-2-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We propose a new depth estimation method using structured light with two projectors. The light
includes disparity information between the projectors. We also propose a method of correcting the light. This
method is needed to estimate accurate depth.
View full abstract
-
Zijian FAN, Haruki MIZUSHINA, Shiro SUYAMA
Pages
34C-3-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We propose a new method to add/subtract the luminances in layered TN-LCDs by using polarization operation.
This system is only composed of TN-LCDs and a half-wave plate. Our method can successfully achieve the complete control
of whole luminance even when one LCD has any luminance.
View full abstract
-
Tatsuhiko EGUCHI, Haruki MIZUSHINA, Shiro SUYAMA
Pages
34C-4-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We propose a new 3D display, called “Layered multi-view DFD (Depth-fused 3D) display”, which has various kinds of
parallax components at the view positions. Our new display has smooth image shift and good monocular perceived depth by adopting
various kinds of parallax components, even when the number of multi-view is decreased.
View full abstract
-
Hayato WATANABE, Masahiro KAWAKITA, Naoto OKAICHI, Hisayuki SASAKI, Ju ...
Pages
34C-5-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We proposed a projection-type integral 3D display with several Ultra High Definition (UHD) projectors. In this
method, we can improve a resolution and a viewing angle of the integral 3D images by the arrangements of projector’s
positions and the design of display optics.
View full abstract
-
Yasuhito SAWAHATA, Toshiya MORITA
Pages
34C-6-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
Integral 3D imaging has technical challenges in displaying deep depth information with high quality since the
depth reconstruction range is restrictive. To display a variety of scenes with high quality, we introduced 'depth-compression
expression,' in which the depth of scenes are compressed, or modified in the axial direction, so that the appearances of the
depth-compressed scenes are kept natural for viewers. By taking advantages of the characteristics of human vision, we found
that viewers did not feel unnaturalness even for the depth-compressed scenes with half the depth of the originals..
View full abstract
-
Pages
34D-0-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
-
Kei OGURA, Kodai KIKUCHI, Takeshi KAJIYAMA, Eiichi MIYASHITA
Pages
34D-1-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We developed a compression recorder for full specification 8K Super Hi-Vision (full-spec 8K) which has frame rate of 120Hz, color depth of 12 bit and no chroma subsampling. To realize both downsizing and real-time processing, we designed the recorder with new video processing boards, which achieved compact layout in the recorder housing, and proposed frame unit processing method to simplify redundant processing. The prototyped recorder demonstrated real-time recording and playback including fast-forward and jog-shuttle.
View full abstract
-
Norihiko ISHII, Yutaro KATANO, Tetsuhiko MUROI, Nobuhiro KINOSHITA
Pages
34D-2-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
Dot selection method in 5:9 modulation code of holographic data storage was considered. Two cases were picked up and distance in each white dot was calculated. The 4 cases which had 8 shorter distances in 9 dots were eliminated. Moreover the cases, which had short distance in 9 dots, were set short distance in 5 bit data. As the results of simulation, improvements of hard detection were 2 ~ 3 % and iteration characteristics of LDPC were also improved.
View full abstract
-
Tetsuhiko MUROI, Yutaro KATANO, Nobuhiro KINOSHITA, Norihiko ISHII
Pages
34D-3-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We developed a parallel signal processing method using 4 GPUs to increase data decoding speed. We confirmed
that LDPC coded 5:9 modulation of compressed 8K video signal with 85 Mbps could be decoded in real-time.
View full abstract
-
Yutaro KATANO, Tetsuhiko MUROI, Nobuhiro KINOSHITA, Norihiko ISHII
Pages
34D-4-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
During reconstruction of hologram, the reference beam is transmitted through the medium. We constructed an
optical equipment for holographic data storage in which the transmitted beam is collected and reused to reproduce another
angle-multiplexed hologram simultaneously. The estimated data-transfer rate from the results of experiments was over 1 Gbps.
View full abstract
-
Mitsunobu OKUDA, Mayumi KAWANA, Yasuyoshi MIYAMOTO
Pages
34D-5-
Published: 2016
Released on J-STAGE: January 23, 2020
CONFERENCE PROCEEDINGS
OPEN ACCESS
We have
proposed
new storage devices
utilizing
current
-
driven domain wall motion
along
magnetic
nanowire
with
ultra
-
high
data transfer rate.
In order to reduce the driving current of magnetic domains
, w
e
fabricated [Co/Pd
] nanowire with
Ru inter
layer
and
inves
tigated
the behavior of current
-
driven domains motion according to the change of magnetic property.
View full abstract
