ITE Technical Report
Online ISSN : 2424-1970
Print ISSN : 1342-6893
ISSN-L : 1342-6893
38.46
Showing 1-22 articles out of 22 articles from the selected issue
  • Type: Cover
    Pages Cover1-
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (10K)
  • Type: Index
    Pages Toc1-
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (72K)
  • Nobuhiko MASAKI, Tatsuya YOSHIZAWA
    Type: Article
    Session ID: HI2014-60
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    SFM (Structure From Motion) is a visual phenomenon, as a well-known ill-problem, that the perception of the three-dimensional spatial structure, such as shape and depth, can be produced from two-dimensional visual images. It has not been clear how a mechanism of this perception can solve this problem. For instance, there is some ambiguity concerning the depth structure in this perception. In this study, we tested whether chromatic contrast can determine the depth structure, to investigate what kind of visual cues contributes to produce the appropriate depth structure in this perception. We found that the chromatic contrast strongly engaged with the determination of the depth structure.
    Download PDF (299K)
  • Tasuku Takeshita, Katsunori Okajima, Sei-ichi Tsujimura
    Type: Article
    Session ID: HI2014-61
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    There are two cone-opponent mechanisms, |M-L| and |L+M-S| cone-opponent mechanisms that contribute to red-green and blue-yellow perception, respectively. At a dim light level in mesopic vision, rods contribute color perception as well as cones. Here we have investigated a contribution of rod signals to the cone-opponent mechanisms. We used a four primary stimulation system that enables independent stimulation to each receptor to control stimulation of the three types of cones and rods. We applied these stimuli for a test stimulus at intense photopic light level and at dim mesopic light level. It was found that rods contribute to |M-L| cone-opponent color signals at the mesopic light level.
    Download PDF (393K)
  • Takuma Morimoto, Steven L. Buck, Keiji Uchikawa
    Type: Article
    Session ID: HI2014-62
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    When yellow targets are surrounded by higher-luminance stimuli, they are perceived as brown. Previous studies examined some of the basic features of brown, and revealed surround stimuli are indispensable to perceive brown. However, we know little about what features of surrounds induce brown perception. Thus, in this study, we employed homogenous and complex surrounds to examine the influences of surrounds on brown perception in terms of luminance statistics. Result showed that luminance of the test stimulus to produce both pure yellow and pure brown is determined by mean luminance of whole surrounds rather than minimum or maximum luminance.
    Download PDF (7088K)
  • Takehiro NAGAI, Toshiki MATSUSHIMA, Kowa KOIDA, Yusuke TANI, Michiteru ...
    Type: Article
    Session ID: HI2014-63
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    We measured reaction times and effects of stimulus durations for material category discrimination tasks in psychophysical experiments. In addition, we analyzed these temporal properties focusing on the relationship to perceptual surface qualities on objects such as glossiness and roughness. In the results, visual surface features (e.g., glossiness and translucency) more strongly affected the temporal properties when reaction times and stimulus durations were short, while non-visual surface features (e.g., roughness and heaviness) more strongly affected the temporal properties when they were long. These results suggest that the mechanisms underlying visual and non-visual feature processing may differ in terms of relationship to higher-order material recognition, or processing time and processing load.
    Download PDF (3615K)
  • Ichiro KURIKI, Hongfei Xie, Yoshiyuki YAMADA, Rumi TOKUNAGA, Kazumichi ...
    Type: Article
    Session ID: HI2014-65
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Visual feature information are once separately processed at each modality; color, shape, motion, and etc. The present study investigated how and where they are integrated into one unified percept by using both psychophysics and fMRI.
    Download PDF (628K)
  • Tatsuya YOSHIZAWA, Shouhei FUJIWARA
    Type: Article
    Session ID: HI2014-66
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In the motion perception, real motion cannot be discriminated from apparent motion whose motion energy is not enough within the window of visibility in the spatiotemporal frequency domain. To perceive the apparent motion, it is necessary to solve the so-called "corresponding problem". In the corresponding problem, the corresponding stimuli between frames are generally defined by physical edges, such as luminance border in order to produce the motion energy. If this is the only corresponding information, the subjective contour could not figure out the corresponding problem. To understand this issue, we tested whether it is possible to perceive the motion using the stimuli defined by the subjective contours exclusively located between the frames. We found the same performance of the motion discrimination in the subjective contour stimuli as those in the real contour, indicating that the motion energy can be generated by not only physical but the subjective border.
    Download PDF (488K)
  • Mayu KAKEGAWA, Ryo KOMIYAMA, Yuko MASAKURA, Masayuki KIKUCHI
    Type: Article
    Session ID: HI2014-67
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    We analyzed brain activities obtained from 2ch NIRS measuring prefrontal cortex. Especially we focused our analysis on the genre discrimination from brain activities, and relation between Kansei evaluation and NIRS data. As a pre-processing, we executed baseline correction for the NIRS data. We adopted Random Forest for discriminating music genres. We calculated correlation between distances of arbitrary two musical pieces on the brain data space and on the Kansei data space, resulted in at most 3.0% correlation. We obtained relatively high discrimination rates for genre discrimination as a whole, though some genres resulted in low performances. Based on above results, we concluded that it is possible to know macro characteristics about listening music from prefrontal brain activities. However, it remained as future problem to find effective features establishing correspondence between objective brain data and subjective Kansei data.
    Download PDF (572K)
  • Shunpei Kaji, Yasuo Watanabe
    Type: Article
    Session ID: HI2014-68
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Recently whether we sense depth from videos on the screen was researched from various points of view. In real space, we sense depth from congestion caused by the movement of both eyes, and the images reflected on the retina of the human eyes. Keys to sense the depth by the images reflected on the retina appear such as binocular parallax which is the difference of the images of both eyes and motion parallax caused by the motion of the images, Relative Velocity of the images, size of the images, shapes of the objects, occlusion of the objects, brightness, the shadows, blur, the size of the field of view, etc. This paper confirmed by experiment and consideration which depth sensitivity caused by movement of the images on the screen was occurred, motion parallax or Depth Sensitivity caused by optical flow. And motion videos were evaluated by depth sensitivity obtained by result, additionally the evaluation was confirmed by experiment whether correct or not.
    Download PDF (493K)
  • Yasuaki TAMADA, Takuto SADAKARI, Kiyoshi IWASAKI, Masayuki SATO
    Type: Article
    Session ID: HI2014-69
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    It has been shown that motion parallax is an effective depth cue for humans, however most of previous studies examined motion parallax accompanied by observer's head movement only in the right and left direction. In the present study, we quantified apparent depth specified by motion parallax when the observer moved horizontally, vertically or in both directions to examine the effects of head-movement direction. The results showed that perceived depth depended neither on the head-movement direction nor the surface orientation but on the spatial frequency of depth corrugation, i.e. larger depth was perceived for lower spatial frequency. Moreover, five of ten observers perceived opposite depth to the geometrical prediction for convex surfaces with more than 1.25 of peak gradient.
    Download PDF (3200K)
  • Yuma HONDA, Takeya YOSHIKAWA, Ichizo MORITA, Akira KANEKO, Kazuki YOSH ...
    Type: Article
    Session ID: HI2014-70
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Stereoscopic images have been spreading rapidly, such as in films and television. However, these images can have adverse effects such as eye fatigue, headache, and 3D sickness. The causes for these symptoms remains little understood. One explanation has been that the occurrence of double images is uncomfortable and because some person cannot recognize the disparity in the images. In this experiment, we set-up a hypothesis that the ability of stereoscopic recognition for 3D images may be developed by training. We have conducted an experiment using 110 subjects of various ages to examine the hypothesis.
    Download PDF (6486K)
  • Akira KANEKO, Yuma HONDA, Takehito KOIMA, Ichizo MORITA, Kazuki YOSHIK ...
    Type: Article
    Session ID: HI2014-71
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Under bright or dark environmental conditions, a man can control the amount of light coming into the retina due to changing the pupil size. In the present experiment, we used 148 subjects including wide range of ages (13 to 84 years of age). Subjects were asked to view and fuse the 3D stereoscopic images on tablets under both bright and dark environmental conditions. We measured subjects' pupil size during their viewing 3D popping out images with 4 levels of parallax. Using the recognition rate of 3D fusion for each parallax, we evaluated the depth of field of subjects. We also studied the relationship between the pupil size and depth of field for 3D images.
    Download PDF (3585K)
  • Takeya YOSHIKAWA, Yuma HONDA, Takehito KOJIMA, Ichizo MORITA, Kazuki Y ...
    Type: Article
    Session ID: HI2014-72
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In viewing 3D images, the important point for human vision is convergence and lens accommodation. In this study, we focused on lens accommodation. We measured lens accommodation while viewing 3D and 2D images which moved sine curve and in with step with changing conditions. The conditions that changed were brightness, type of image. We compared the data and examined the influence of these conditions on lens accommodation.
    Download PDF (5419K)
  • Akihiro SUGIURA, Kunihiko TANAKA, Shun WAKATABE, Chika MATSUMOTO, Kazu ...
    Type: Article
    Session ID: HI2014-73
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    While continuously observing objects in motion, humans may develop an illusionary sense of moving in the same direction as the objects, despite being quiescent. This phenomenon is termed as vection. In this study, we investigate the effect of viewing 3D movies on vection. The body sway of subjects while viewing 2D and 3D 3-min movies was recorded. From the results of this study, we obtained three findings. First, a longer viewing time provides higher synchronization accuracy with the movie. Second, the synchronization accuracy while viewing 3D movies is higher than that while viewing 2D movies. Third, there is a possibility that the entire space motion in the movie affects vection.
    Download PDF (4250K)
  • Yuriko Yoshino, Hyoungseop Kim, Joo Kooi Tan, Seiji Ishikawa, Seiichi ...
    Type: Article
    Session ID: HI2014-74
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Recently, some computer aided diagnoses (CAD) systems has been introduced until now. Temporal subtraction technique, as one of the CAD systems, can emphasize temporal changes of interested regions by subtracting a previous image from a current one, which are images of a same patient. A previous image is deformed to suit a current image considering a variation of pose and inhalation differences when a patient accepts CT inspection at different times and subtracted from a current image. However, accurate registration is difficult because organizations such as a vessel are distributed three-dimensionally in a lung region. It is caused that each shift vector (shift amount and direction of a pixel) is less consistency to surrounding shift vectors in some area of a lung region by mis-registration and many artifacts are remained on temporal images. In this paper, we propose a method of registration based on GGVF considering information of structures in a lung regions and smoothing of shift vectors, and aim to improve registration and reduce artifacts in a subtraction image.
    Download PDF (3732K)
  • Yusuke FUKUSHIMA, Toshihiko YAMASAKI, Kiyoharu AIZAWA
    Type: Article
    Session ID: HI2014-75
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In this paper, we present methods to assess oral presentations by estimating the topic and estimating impressions the audience would have. We demonstrate that some impressions strongly depend on their syntax and contents. Moreover, we conduct an evaluation of topic classifications using latent semantic analysis (LSI), latent dirichlet allocation (LDA) and Naive Bayes classifier.
    Download PDF (651K)
  • Eiji WATANABE, Takuya HIGASHI, Takashi OZEKI, Takeshi KOHAMA
    Type: Article
    Session ID: HI2014-76
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    In this report, we discuss the influence on the behavior by the lecturer. First, we extract the face movements by lecturer and students from movies. Next, the relations among the above features concerning on the behaviors by the lecturer and students can be approximated by multi-layered neural networks. Finally, we extract the influence on the behavior by the lecturer based on the internal representations of neural networks.
    Download PDF (4182K)
  • MIeko KOGI, Kaori SAKAMOTO, Tatsuya YOSHIZAWA
    Type: Article
    Session ID: HI2014-77
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    It is very important to analyze human chromosomes for the definitive diagnosis, in order to predict congenital abnormality and to prenatally diagnose genetic abnormality of fetuses. This analysis generally has been carried out by well-trained experts with a long-term experience. Since there is a shortage of those experts in the recent clinical field due to the increasing of the diagnosis required chromosomal analyses, it is mandatory to employ more experts. However, it is required that the experts must supervise unskilled trainees to practice to be trained experts despite this circumstance. To increase the number of the experts with a limited time, we therefore developed a new self-training education system for learning the karyotyping skill. This system can improve trainees' skills with their own paces and levels of the skills without supervising by the experts. To evaluate this system, here we tested the performance of how beginners can improve their skill by using this system. Results showed that all trainees could improve their karyotyping skills significantly after 60 repeats of the training. This result indicates that this system has a great performance to learn and to improve the karyotyping skill for any beginners.
    Download PDF (475K)
  • Type: Appendix
    Pages App1-
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (81K)
  • Type: Appendix
    Pages App2-
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (81K)
  • Type: Appendix
    Pages App3-
    Published: November 21, 2014
    Released: September 22, 2017
    CONFERENCE PROCEEDINGS FREE ACCESS
    Download PDF (81K)
feedback
Top