All issues
Predecessor
25.29
Displaying 1-29 of 29 articles from this issue
- |<
- <
- 1
- >
- >|
-
Article type: Cover
Pages Cover1-
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (15K) -
Article type: Index
Pages Toc1-
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (121K) -
Satoshi KINOSHITA, Suguru SAITO, Masayuki NAKAJIMAArticle type: Article
Session ID: HIR2001-29
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSContours which obtained by image processing are sometimes not continuous. So it should be connected for postprocessing. There are curve fitting methods to interpolate curves, but these are difficult to apply to arbitrary curves. We propose an algorithm which interpolates curves by calculating possibility of curve existence and joining pixels which have high possibility. Because this method calculates the possibilty of existance uniformly in the whole image, it doesn't need additional information of the image. This algorithm uses filter, so we investigated the optimum filter size for the curvature in the image. We also applied this algorithm to real image and confirmed it can interpolate Curves.View full abstractDownload PDF (607K) -
Fumio Wada, Mitsuru Iwata, Shun'ichi TanoArticle type: Article
Session ID: HIR2001-30
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSOur goal is an implementation of an adaptive system for presenting information in high-cognitive load environment. Users in such environments include car drivers, aircraft pilots and plant operators. We have examined models to estimate user's cognitive loads for planning and creating a great variety of information presentation methods, and set up three hypotheses. This paper describes an experiment to measure subjects' response to a combination of information presentation methods in order to verify these hypotheses. The results suggest that a combination of information presentation methods influence the ability of user's cognition and there is some possibility that hypotheses are valid.View full abstractDownload PDF (1028K) -
Toshikazu MatsuiArticle type: Article
Session ID: HIR2001-31
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSA new objective image quality evaluation method for sharpness has been developed by a pertinent application of a cooperative human vision model which can simulate various kinds of perceptual responses theoretically. In order to examine its effectiveness, the method is applied to sharpness evaluation for generalized knife-edge images composed of white and black lines with variable line-width and the theoretical explanation of optimum viewing-distance characteristics and their underlying visual factor clarified in previous experiments. The following results are obtained: (1) the experimental relation between image transmission system's band-width and sharpness is reproduced; (2)the existence of optimum viewing-distance is derived; (3)the optimum viewing-distance takes a constant value depending on the presented image's tine-width at wide band-width conditions and increases as the band-width is narrowed; (4)the visual factor for our sharpness judgments is explained. These results suggest that this evaluation method is effective as a sharpness metric based on human judgments.View full abstractDownload PDF (808K) -
Junji KAWASAKI, Norihisa MORI, Kyoko KATO, Taizo IIJIMAArticle type: Article
Session ID: HIR2001-32
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSwe presented a theory and experimental results for layered model, which extends external world, retina and brain of the visual model of the linen. These papers propose the objective evaluation to coincide with the subjective evaluation of the leman in over to evaluate the relative merits of the various modulation methods. using Girl, Moon surface, Milk drop, Column, the measure of approximation presents the objective evaluation to coincide with the subjective evaluation by ITU-R and the method of successive categories.View full abstractDownload PDF (1958K) -
Junji KAWASAKl, Kentaro TACHIBANA, Naoto KITAMURA, Kyoko KATO, Taizo I ...Article type: Article
Session ID: HIR2001-33
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWhen we see a binary black and white image, it appears, to our visual sense, to become clearer because of a pseudo halftone. We presented a theory and experimental results for layered model which extende external world, retina and brain of the visual model, the equivalence approximation method. This paper propose the modifyed equivalence approximation method which modify the equivalence approximation method, it is difficult to find window size. We obtain experimentally the reconstracted images the measure of approximation of density four division method, ordered dither method, simple two value method and least mean error method. The measure of approximation presents the objective estimation scale to coincide with the subjective estimation of the four modulation images by the modifyed equivalence approximation method of the fixed window size. We presents the image estimation method which is no significant difference between its method and epuivalence approximatioon method.View full abstractDownload PDF (1176K) -
Naotaka Kamiya, Mehdi N. ShiraziArticle type: Article
Session ID: HIR2001-34
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIntracortical inhibitory cells play an important role in shaping the simple and complex cells' orientation selectivity. There is, on the other hand, an ample experimental evidence supporting that the intracortical inhibitory cells are orientation selective as well. Recently a computational model for the orientation-specific inhibition has been introduced. The model consists of three parts : (a)a two-layered hierarchical Markov random field ; (b)a computational goal formulated on the Maxirnum-A-Posteriori estimation principle ; (c)a deterministic parallel algorithm to achieve the computational goal. In this article, we introduce a physiologically-plausible firing-rate-coded neural network to show how this computational model can be implemented.View full abstractDownload PDF (716K) -
Tetsuya Yagi, Tadatsugu Udono, Seiji KamedaArticle type: Article
Session ID: HIR2001-35
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThe image processing is one of the most difficult problems for the current computer. On contrast, the visual system of living organs carries out the processing easily. In this study, the depth of the image was computed with an algorithm inspired by that of the visual cortex using the silicon retina which emulates the function and tile structure of the vertebrate retina. The system and algorithm described in the present study leads to a novel paradigm of the robot vision.View full abstractDownload PDF (1363K) -
Takahiro Watanabe, Akihiro Fujii, Kouji Takagi, Yasuhiro Cho, Yoshifum ...Article type: Article
Session ID: HIR2001-36
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThis paper presents the personal identification system based on iris patterns which is captured automatically. This system consists of a wide-view camera and a narrow-view camera. In the first step of identification, a face image is captured by the wide-view camera and eye position is detected from the face image. In the second step, iris image is captured by the narrow-view camera using eye position found in the first step, and the person is identified using the iris pattern. Our experimental results show the availability of our system for personal identification.View full abstractDownload PDF (920K) -
Satoru ODO, Kiyoshi HOSHINOArticle type: Article
Session ID: HIR2001-37
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThe purpose of this study is to estimate three-dimensional motions of the sign language with a monocular camera. In the system, a three-dimensional human model with three-dimensional polyhedrons with seven degrees of freedom was prepared, and two-dimensional image sequences sith a CCD camera were applied. Moreover, optical flows in three frames before current time were used for the motion estimation, so that the system can be robust to self-occlusions by performer's own face and body. In the experiment, standstill, straight, circular, arc motions and their combinations were adopted for the estimation. As the results, the system realized the gesture estimation and tracking with high accuracy. In the experiment of reproducing three-dimensional motions with computer graphics, gesture reproductions were performed within 340ms per a frame without any specific computer hardwares. These results showed that the system is effective for three-dimensional motion estimation for sign language with a single CCD camera.View full abstractDownload PDF (654K) -
Liming Li, Yuki Hayashida, Tetsuya YagiArticle type: Article
Session ID: HIR2001-38
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSA partial vision is expected to be restored in patients with retinal degeneration using an artificial retinal implant to stimulate electrically the survived neurons. To design an appropriate retinal implant spatio-temporal poperties of thc retinal ganglion cell response to transretinal current. stimuli were studied. There were two typical chasses of response. One class showed constant. latency as the current amplitude was increased from 50〜200μA. Another class showed a variable latency among 10 trials. And the latency got shortrer as the current was increased. The threshod of the two classes of response ranged 10〜50nC. In most Cases the current stimuli used in this study excited ganglion cells in a local area. When I.he retina was stimulated by a charge of 10nC the ganglion cells at about 300 μm front the stimulation site were initiated by a probability of 50%.View full abstractDownload PDF (728K) -
Y. Terasawa, T. YagiArticle type: Article
Session ID: HIR2001-39
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSVisual prostheses are the one to recover the visual sense by means of engineering techniques. They have been developed in and out of our country. The purpose of this study is Lo determine the necessary specification of visual prostheses. Therefore we have developed a simulator which can translate a normal vision into a prosthetie vision. Then we selected the reading as one of the most typical task inourdaily lives. and estimated the reading ability under the condition of simulated prosthetic vision. We quantitatively evaluated the reading ability by using MNREAD. a commonly-used reading ability test. As a result. It was suggested that the range 4-8 is appropriate as the levels of electric stimulation.View full abstractDownload PDF (663K) -
Kiyoshi HoshinoArticle type: Article
Session ID: HIR2001-40
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn this study, the author used a humanoid robotic arm with rubber muscles whose muscle-skeleton structure was equal to that of tile human, for the purpose of understanding precise and complicated voluntary movements produced by human brain. First, responses which resembled primitive reflexes observed in the human infant were generated by the robotic arm, and output patterns of human CNS to each muscles were estimated. Then, magnitude and time-related parameters of the output patterns were changed, so as to experimentally examine how much precise and complicated the behavior of the robot may become in its voluntary movements.View full abstractDownload PDF (1491K) -
Dai HASEGAWA, Masakazu AKASHI, Yuichiro KUME, Junji SONEArticle type: Article
Session ID: HIR2001-41
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSMotion platforms have been employed to stimulate vestibular sensation in addition to visual displays for virtual reality. Sensational characteristics of self-motion affected by acceleration and visual stimulation were investigated to construct effective virtual environments. In this study, vertical self-motion sensations were psychophysically measured with motion base and visual patterns moving linearly and vertically. Different sizes of visual Geld and different types of visual stimuli were used in the experiments to extract parameters affecting the self-motion sensation. As results, self-motion sensations are enhanced by landscape image rather than rectangular stripe patters, and follows the directions of the stripe pattern movement.View full abstractDownload PDF (772K) -
Yoshino Ishikawa, Shihoko Kamisato, Kiyoshi HoshinoArticle type: Article
Session ID: HIR2001-42
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThe purpose of this study is to examine the relationship between physical motion characteristics and the subjective impressions produced in dancing. The structure of subjective impressions for dance motions with no regular rule of operation was investigated using Okinawan folk dance "Kachaasii". Only the movement of the leg was discussed here. First, the questionnaire method and the principal component analysis were adopted for investigation of the mental structure of the leg motion in dance. Next, the dance was reproduced with computer graphics, and the subjective impressions were examined. The results showed that the movement of leg has five subjective scales : preference, weight, width and kinetics.View full abstractDownload PDF (898K) -
Yuukou Horita, Akira Kushida, Tadakuni Murai, Yoshio NakashimaArticle type: Article
Session ID: HIR2001-43
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSTo consider the quality of service for stereoscopic image, it is necessary to develop a quality evaluation method for coded stereoscopic color image. We perform the sujective evaluation of the quality degration for the steredscopic image. By using the MOS derived from this evaluation experiment, we consider the influence of the visibility of the quality degradation. The evaluated images are created by combining the object domain and background domain. The degree of the image quality of two domains is controlled by using the different coding parameter. As the result, the total quality of the stereoscopic image is falling down when the quality of the object domain and background domain has some difference.View full abstractDownload PDF (744K) -
Haruki Mizushin, Keiji UchikawaArticle type: Article
Session ID: HIR2001-44
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIt has been shown in our previous research that thresholds of displacement detection across saccade in a whole visual field were different from those in a restrict visual field. There is a question whether this difference may depend on the presence of cue of absolute location(frame)within the visual field or the presence of pattern stimulus in the peripheral visual field. In this study, we carried out experiments to answer this question. The results indicate that an frame has an effect of suppressing threshold elevation and a tendency of thresholds depends on presence of stimulus pattern in a peripheral visual field.View full abstractDownload PDF (879K) -
Ko Sakai, Takuya MonmaArticle type: Article
Session ID: HIR2001-45
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWe investigate the perception of three-dimensional structure from shading when contour is not given or contour gives ambiguous convex/concaveness. It is crucial to separate contour and shading in order to examine 3D perception purely from shading, and the interaction between contour and shading. We developed stimuli whose 3D structure is generated based on Gaussian or white-noise so that their saliency of contour is controlled. Psychophysical experiments were carried out using the stimuli and the following results were obtained. (1)Contour is dominant for 3D perception comparing to shading even if contour is ambiguous, (2)even if stereopsis gives 3D structure, shading is dominant if contour is invisible, (3)not only monotonic gradation of grayness gives shape-from-shading but also the global changes in local contrast.View full abstractDownload PDF (1045K) -
Masayuki Kikuchi, Ko Sakai, Yuzo HiraiArticle type: Article
Session ID: HIR2001-46
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSField et al. found, through a psychophysical experiments detecting smooth contour composed of Gabor-patches from a randomly oriented and positioned Gabor patches, that low curvature of contour and less orientation-discrepancy of Gabor-patches from the contour facilitate contour saliency. This study brought binocular-disparity in this experimental paradigm and investigated relation between contour saliency and various geometric features such as number of depth-gradient discontinuities, and cncave or convex depth-structures.View full abstractDownload PDF (1154K) -
Daisuke KAKEHI, Satoshi SHIOIRI, Miho TAKEO, Hirohisa YAGUCHIArticle type: Article
Session ID: HIR2001-47
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThere are two binocular cues for perception of motion in depth. One is binocular disparity changes in time and the other is velocity differences between the len and right retinal images(inter-ocular velocity differences). We investigated the individual variation of the use of the two cues for perceiving motion in depth. We found that 18 of 25 observers(72%)could use inter-ocular velocity differences. 24 observers(96%)could use disparity change in time. These results suggest that inter-ocular velocity differences as well as disparity change in the is useful to see motion in depth.View full abstractDownload PDF (702K) -
Kenji Susami, Takehiro Kanamaru, Kazuo Ishikawa, Toyohiko HatadaArticle type: Article
Session ID: HIR2001-48
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThe purpose of this study is to confirm the perceived frontal parallel plane in binocular stereopsis. In this experiment, we used a I i I square as a target. The target was located in the perpendicular direction and the horizontal direction from a fixation point at the distance of 2, 4, 6 and 8 degrees. The subject judged the perceived depth by adjusting the binocular disparity of the target and it stopped in the position where a target is felt at the same plane as a fixation point. Consequently, on the horizontal conditions, a clear tendency was not found since there was an individual difference. In the perpendicular conditions, in order to perceive the target to the same depth as the fixation point, binocular disparity of the upper part needed to be given in a direction of crossed disparity, and the lower part tended to incline in a direction of uncrossed disparity. These results suggest that when the stereoscopic images are presented on the surface of a display with zero binocular disparity, the perceived perpendicular surface inclines in the depth direction.View full abstractDownload PDF (741K) -
Kazuho Fukuda, Satoshi Shioiri, Seungbae Lee, Hirohisa YaguchiArticle type: Article
Session ID: HIR2001-49
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWe investigated the effect of motion on stereopsis using sinusoidal gratings as stimuli. Measurement of disparity threshold(i. e. minimum disparity for depth perception)showed that stimulus velocity influenced contrast characteristics of disparity threshold. The result also showed that existence of relative motion influenced disparity threshold. These results suggest the importance of not only local motion velocity but also surrounding motion for stereopsis.View full abstractDownload PDF (601K) -
Fumio Okuyama, Masahiro Katou, Seo Hwa Joung, Nagamasa Sakabe, Tomio S ...Article type: Article
Session ID: HIR2001-50
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThere are the problem of accuracy and input speed in text data input procedure using Automatic Speech Recognition software practically. Therefore in this paper we discuss these problem. For this experiment we obtained the accuracy with ten subjects reading general newspaper article using commercial automatic speech recognition software. And we measured the time for input that the subject input text data with speech recognition software and after that he corrected the input misses with keyboard. The result are that the average accuracy was about 90% and input speed was as fast as expert keypuncher.View full abstractDownload PDF (654K) -
Hwa Jeong Seo, Nagamasa Sakabe, Tomio Sekiya, Fumio Okuyama, Yeun Hwa ...Article type: Article
Session ID: HIR2001-51
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWe distinguished speaker dependent system, individual speaker registered enroll worlds, speaker independent system, random speaker did not register enroll words and Database system based on technical medical terms with 73 words in the IBM ViaVoice. We examined an input system on Electronic Medical Record with automatic speech recognition software for outpatietnt records. And, there are the best results, speech recognition rates(accuracy), we obtained so far, using the speaker dependent system, gave 95.8% accuracy, using the speaker independent system. gave 88.2% accuracy and the results obtained show that this model Lead to improved recognition performance as like Database system based on technical medical terms.View full abstractDownload PDF (578K) -
Daisuke Tsuru, Hiroki Takahashi, Masayuki NakajimaArticle type: Article
Session ID: HIR2001-52
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSHearing has features such that emphasize the presence of objects and perceive what cannot be visible. So it is important to utilize spatial acoustics for virtual environments to strengthen realistics sensations, whereas the study for realtime generation of spatial acoustics has not been done much. Our goal is to generate realtime spatial acoustics in complex virtual environments. First we propose an acoustic model which models diffraction and reflection, and considers their approximations of frequency characteristic and the way to mix diffraction and reflection. Then, we propose data structures which can determine sound ray paths efficiently. We use visible graph for searching diffract paths, and cell adjacency graph for reflect paths. These methods make it possible to generate impulse responses at complex sound field like ordinary architecture in realtime.View full abstractDownload PDF (764K) -
Kouji Yamashita, Kenji Terao, Tetsuo Kawamura, Fumio Okuyama, Hiromasa ...Article type: Article
Session ID: HIR2001-53
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSwe have developed a tele-rehabilitation support system. OT and PT evaluated it that this system is valuable to use for the tele-rehabilitation training. However, they pointed out that they can't touch patient directly and there is a blind spot form in the camera eye. We have thought that we have to develop qualitative measurement method for Range of Motion by image analysis. mere is possibility that ROM can be automatically detected in this system Also, we could get the data which becomes the standard of the process of the recovery. We will point out that this system is useful for the tele-rehabilitation.View full abstractDownload PDF (698K) -
Article type: Appendix
Pages App1-
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (88K) -
Article type: Cover
Pages Cover2-
Published: March 21, 2001
Released on J-STAGE: June 23, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (78K)
- |<
- <
- 1
- >
- >|