映像情報メディア学会誌
Online ISSN : 1881-6908
Print ISSN : 1342-6907
ISSN-L : 1342-6907
60 巻, 12 号
選択された号の論文の32件中1~32を表示しています
ふぉーかす
講演 アントレプレナー・エンジニアリング・フォーラム基調講演
小特集
ディジタルコンテンツ制作を支える新技術
技術解説
  • 中川 正雄
    2006 年 60 巻 12 号 p. 1908-1913
    発行日: 2006/12/01
    公開日: 2008/12/01
    ジャーナル フリー
    いつでもどこでも情報を受け取ることのできる無線通信の普及にはすごいものがある. 電波, 特に波長が10cm程度以上の波は, ビルの谷間でも, 室内でも, どこからともなく入り込み, いつでもどこでも情報を交換するという我々の要求を満たしてくれる. こうしたモバイル技術の次に来るものとして, ユビキタス技術が期待されている. これはあらゆる物が情報を発する技術であり, 物の管理のためだけでなく, 街やオフィスなどを, 人間にとって便利に, 安全に, 楽しくするための技術である. こうしたことを可能にする可視光通信技術について述べていく. 見える光は電波と同じ波の一種なので, 当然ながら, 情報を伝えることができる. 可視光通信の可能性をユビキタスなインフラとしてのメリットや社会的メリットから探っていく.
講座
コンピュータビジョンは人のビジョンにどこまで近づいたか
~難点とそれを回避する実践的アプローチについて~
てれび・さろん
~時代を創った技術~
知っておきたいキーワード
映像現場訪問記
会議レポート
海外文献紹介
ニュース
論文・研究速報
論文小特集 ヒューマンインフォメーション
小特集論文
  • 松井 利一
    2006 年 60 巻 12 号 p. 1947-1955
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    We studied whether the cooperative motion detection model is effective in reproducing human vision's motion detection and accommodation characteristics for many kinds of apparent motion stimuli. The stimuli consisted of two kinds of apparent motion sine-waves with different motion directions, and the model's responses to the stimuli were compared with actual responses in human vision. The responses corresponded well in two ways. (1) There was a critical spatial frequency at which the motion perception changed from one state to the other, and it increased as the stimuli approached in a continuous motion. (2) Under the critical spatial frequency, two kinds of motion were detected in two kinds of accommodation states depending on which motion stimulus was paid attention to, while over the critical frequency, one motion was detected in those states. These results suggest that the cooperative model is an accurate model of the motion detection mechanism including the accommodation mechanism.
  • 渡部 正行, 名手 久貴, 高木 康博
    2006 年 60 巻 12 号 p. 1956-1963
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    Recently developed high-density directional displays have the ability to evoke the accommodation responses and provide smooth motion parallax. In the this work, we explain how smooth motion parallax is produced by the high-density directional displays. We also examine the effects of the motion parallax on object depth perception using 64-directional displays. Cylindroids with different ellipticities that have coarse random dot patterns or checker patterns on their surfaces were displayed, and observers were asked to describe the perceived ellipticities. The density of the random dots was so coarse that more than half of the observers could not correctly perceive the ellipticities using only binocular parallax. The use of the motion parallax improved thier ellipticity perception greatly. Moreover, almost all observers could perceive the ellipticities to a certain precision only with the motion parallax. The binocular disparity and motion parallax produced by the high-density directional display were so effective that the use of checker patterns on the cylindroid surfaces results in little improvement.
  • 白澤 洋一, 三浦 信, 西田 眞, 景山 陽一, 栗栖 怜史
    2006 年 60 巻 12 号 p. 1964-1970
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    Speech speed is irregular in each speaker; however, it is important to adjust the structure of the speech data and allow for regional differences (accents or dialects) when identifying a person by their lip motion features. We propose a method to identify individuals by analyzing three kinds of lip motion features: lip width, lip length, and lip angle. The analysis is made on the basis of these features' relative values obtained from the primary and object frame. The proposed method's procedure involves three steps: automatic extraction of the lip motion features from the primary frame, change of the standardized data structure, and individual identification by fuzzy reasoning. We analyzed the lip motion features of 17 subjects using the proposed method, and achieved a successful identification rate of 100%. The results of the cumulative match rate (CMR) suggest that the proposed approach can be used to identify individuals more accurately than can DP matching.
  • 阿部 清彦, 佐藤 寛修, 大山 実, 大井 尚一
    2006 年 60 巻 12 号 p. 1971-1979
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    We present an eye-gaze input system for people with severe physical disabilities, such as those caused by amyotrophic lateral sclerosis. This system uses a personal computer and a home video camera to detect eye-gaze under natural light. The system detects both vertical and horizontal eye-gaze using simple image analysis, and does not require special image processing units or sensors. We propose an operation support system for physically disabled computer users as an application of the eye-gaze input system. Our new computer operation support system uses a direct selection method by switching indicator groups. This system enables the input of Japanese text (both kana and kanji), alphabets, and symbols at a faster pace. In addition, users can access all Windows operations using the system, thereby enabling many commercially produced software applications to be operated by eye-gaze.
  • 渡邉 亮, 上田 裕市, 坂田 聡
    2006 年 60 巻 12 号 p. 1991-1999
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    This paper describes evaluations of the readability of visualized speech represented by integrating several speech parameters. In the first test, a closed subset of a small vocabulary (75 words: nouns) uttered by four male speakers was examined. Next, the vocabulary was expanded to a medium size (520 words: nouns, adjectives, verbs, and adverbs) uttered by 10 males. Finally, we conducted a systematic test to analyze the influence of vocabulary size, defective representation of a parameter, and real-time representation on the readability of utterances by 10 males and 10 females. The results showed that readability of the visualized speech was almost unaffected by differences among speakers. Moreover, words from the small vocabulary (75-120) were readable in real time because subjects could rapidly grasp the vocabulary through the repetition of randomized words uttered by various speakers during tests even without being told correct answers.
  • 藤井 真人, 伊藤 崇之, 大西 昇
    2006 年 60 巻 12 号 p. 2000-2007
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    Shading in two-dimensional images produces a compelling perception of three-dimensional shape. When mechanisms of perception are considered from an engineering standpoint, shape-from-shading is known to be an ill-posed problem that is difficult to solve. Constraints are needed to solve this kind of problem. Although estimation of reflectance properties from shading is also considered an ill-posed problem, mechanisms of estimation by the visual system are still poorly understood. Therefore, we proposed a psychological experiment intended to determine whether the visual system can estimate surface reflectance properties from shading. In the experiment, we expect to use contours as a cue for perceiving three-dimensional structures. From a preliminary experiment we obtained data that did not rule out that the ability of the visual system to estimate surface reflectance from shading.
小特集研究速報
  • 田邊 喜一
    2006 年 60 巻 12 号 p. 2008-2011
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    We analyzed the eye-blink activity of a subject while he or she memorized facial images, that were presented in a restricted visual-field environment. Eye-blink activity was observed during two memorizing activities: facial component memorization and full-face memorization. A subject memorized a facial component (eyes, nose, mouth) individually in the former and memorized the whole face in the latter. Blinks frequently occurred when the subject observed facial components during full-face memorization. However, differences in observation area negligibly affected blink-rate during facial component memorization. Moreover, when the subject was allowed to blink during facial component memorization, the restricted visual field moved in slowly and the subject had long fixation pauses. In contrast, this did not occur during facial component memorization when the subject was not allowed to blink.
  • 中山 太一郎, 岩井 大輔, 佐藤 宏介
    2006 年 60 巻 12 号 p. 2012-2014
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    We have been developing an information presentation system which projects graphical information onto real documents on a desk. The system assists a user to understand the documents' contents with optically superimposed helpful information. When the graphical annotation is projected over very busy and/or textured documents' surfaces, the readability of both the printed information on the documents and the projected information need to be legible. In this paper, we introduce two types of techniques to improve the readability of the projected information; automatic projection area selection and optically moving graphics. In order to confirm the improvement on readability, we have experimented the readability of these two techniques with psychological tests. The result indicates that the former technique is enough in practical uses.
  • 吹拔 敬彦
    2006 年 60 巻 12 号 p. 2015-2018
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    The so-called apparent motion for sampled motion in TV signals can be clearly analyzed as real motion, using the theory of spatio-temporal sampling. The reason is examined why the “apparent motion” was introduced to explain the phenomenon that the sampled motion looks like a smooth one.
  • 中島 彩, 海老澤 嘉伸, 塗壁 悠治
    2006 年 60 巻 12 号 p. 2019-2025
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    Pupil detection is useful in human interface and monitoring systems. Until now, pupil detection systems have detected pupils using the pupil brightness difference, produced by the positional difference of two light sources. The proposed method uses the difference in pupil brightness produced by light sources of different wavelengths. The proposed light source is much smaller than conventional ones, and its brightness does not depend on the distance between the user's face and the camera. Testing showed that pupil detection was possible using the proposed method.
 
論文
研究速報
  • 牧野 大輔, 井上 光平, 原 健二, 浦浜 喜一
    2006 年 60 巻 12 号 p. 2038-2041
    発行日: 2006/12/01
    公開日: 2008/03/07
    ジャーナル フリー
    We describe a non-photorealistic rendering method for generating paper mosaic images that takes into account and compensates for the effects of the thickness and texture of the paper onto which the image is projected. We modulate the brightness of the paper mosaic image on the basis of the emboss-filtered image of the overlapped pieces of paper. We also modulate the brightness of the paper mosaic image on the basis of the paper's texture. That is, we incorporate the effects of the paper's thickness and texture into the resultant paper mosaic image. We experimentally showed that our method increases the three-dimensional sense of depth in paper mosaic images.
feedback
Top