-
Article type: Cover
Pages
Cover1-
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Article type: Index
Pages
Toc1-
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Yoshiya Iwaki, Hiroki Takahashi
Article type: Article
Session ID: ME2011-1
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Discharged headlights with HID (High Intensity Discharge) lamps are widely used by many cars, recently. Those headlights make driver's view clear because it output brightness and while color light. But when you drive at rainy day, your view is dazzled by oncoming cars have the headlights. Therefore, we propose a method to detect marking lines on the road surface and lanes at nighttime. Furthermore, the method is able to detect and track lanes by online processing. While the processing, it uses two images, current and previous images, for reducing processing times. Finally, we compare our method with former method about precision and processing times, and discuss these data.
View full abstract
-
Kenji SUDO, Masahide KANEKO
Article type: Article
Session ID: ME2011-2
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes a method of real-time detection and tracking of facial feature points from the facial image sequence taken by a web camera, and the automatic real-time synthesis of facial caricature animation. Firstly the pupils are detected in the initial frame by using the circular separation filter. Next the area of each facial part is localized from the positions of pupils, and facial feature points are detected in each area by image processing methods such as YIQ transform, Canny edge operators, and so on. These detection processes are carried out in every frame and facial caricature animation is generated based on the detected feature points. In order to carry out the process from capturing moving facial images to generation of facial caricature animation in real-time, GPGPU and parallel processing methods are introduced to accelerate the detection of facial feature points.
View full abstract
-
S. KUSAMA, D. NAKAJIMA, H. ISONO
Article type: Article
Session ID: ME2011-3
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
The authors have proposed a method for providing support to hearing-impaired students during classroom lectures using a combination of PC note-taking and mind mapping. Whereas earlier methods of PC note-taking involved information assurance using only textual information, the present mind mapping technique is a advantageous in that the overall content of the lecture can be summarized in illustrative form, thus facilitating student comprehension. This approach was found to be effective in providing assistance for hearing-impaired students attending lectures in a university setting.
View full abstract
-
Toyohiro MINAMIGAWA, Atsushi YAMASHITA, Toru KANEKO
Article type: Article
Session ID: ME2011-4
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In recent years, there has been growing the need for observation systems by underwater robots in the aquatic environment research. Image sensing is in general a very effective way for observation, however, in subsea environment, bubbles, small fish and other floating matter may sometimes interfere with the field of view as scene occluders. This paper proposes a method for removing scene occluders from images by detecting moving objects with high speed camera. The method identifies different objects with feature values such as motion, size and color, and removes the scene occluders by using sequential images.
View full abstract
-
Eigo MIYAZAKI, Kazuo OHZEKI
Article type: Article
Session ID: ME2011-5
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In recent years, the development of multimedia technologies have been developed, such as smaller and faster mass storage media high definition video transmission. Terrestrial digital broadcasting and TV recording has its practical use as an example. While a large capacity storage, but users can get information and say it is not considered. We analyzed video data to watch sports video data efficiently stored for a long time, has been studied to provide users with useful information. That highlight generation. Among the major sports are football, produce the effect obtained by highlighting seems to be great.
View full abstract
-
Tomoki SAKA, Toshiyuki GOTOH, Seiichiro KAGEI, Tae IWASAWA
Article type: Article
Session ID: ME2011-6
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Cancer cells induce capillary growth (angiogenesis) into the tumor for its expansion. Gd contrast agents are commonly used for dynamic contrast enhanced MRI to obtain perfusion maps, such as BV (blood volume), BF (blood flow) and MTT (mean transit time). However, it has a problem of the nonlinear dependence of MR intensity on contrast agent density on high-speed MR imaging. To solve this problem, a method for calibration using Gd-MR intensity dependent model is proposed in this paper. Then a method for analyzing the blood flow parameters with a model evaluating the effects of the intravascular and the extravascular exchange of contrast agent is also discussed.
View full abstract
-
Koji Soma, Daishi Watabe, Zichong Kong, Hideyasu Sai, Katsuhiro Sakai, ...
Article type: Article
Session ID: ME2011-7
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Nozomi ONODA, Hiroto INOUE, Mie SATO, Masao KASUGA
Article type: Article
Session ID: ME2011-8
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, design of characters has become diversified with the spread of the computer. As a representation of the characters, there is a design of characters with pictorial depth cues. This paper arranges pictorial depth cues and collects a large quantity of characters with pictorial depth cues. Then, on the basis of subjective evaluation experimentation, we classify the general representations of characters with pictorial depth cues and analyze features of solidity and a sense of depth that are given by those characters.
View full abstract
-
Masato Hashimoto, Hisae Shimamura, Hiroto Inoue, Mie Sato, Masao Kasug ...
Article type: Article
Session ID: ME2011-9
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Content creators have to consider the large number of pattern of camera paths and emotional effects for setting the camera work on the CG scene. This task will be a burden on them. Therefore, the study on the technology that generates appropriate camera work automatically has been conducted. In this paper, we focused on the optical flow, and fundamentally studied automatic generation of appropriate camera works on CG scene. We conducted the subjective evaluation experiment. From the result of the experiment, we considered camera works that give a natural impression to the viewers.
View full abstract
-
Kazuhiro MATSUSHIMA, Mie SATO, Masao KASUGA, Naoki HASHIMOTO
Article type: Article
Session ID: ME2011-10
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Immersive projection display gives us a feeling of immersion. However, immersive projection display is largescale and expensive. So, it is impossible to use personally. In this paper, we propose a small-scale device that can present immersive images in an indoor space. In our image projection, a fish-eye lens is used to expand the projected plane and surrounding walls in the space are used as a screen.
View full abstract
-
Ayano HOSOGAI, Mie SATO, Masao KASUGA, Naoki HASHIMOTO
Article type: Article
Session ID: ME2011-11
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, image projection has been becoming popular since projectors are miniaturized and lower-priced. When images are projected on non-white screen, patterns of the projected plane influence the projected images. This study examines luminance correction to reduce the influence of the projected plane from the projected images under dynamic environment.
View full abstract
-
Michimi INOUE, Mie SATO, Masao KASUGA, Naoki HASHIMOTO
Article type: Article
Session ID: ME2011-12
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, HDR (High Dynamic Range) display has been active studied. HDR display has high dynamic range of luminance and high gradation. Related studies tried to make a HDR display by using multiple projectors. However, it was difficult to make smooth gradation change on projected planes with the multiple projectors. Therefore, the aim of this study is to propose a HDR display with four projectors that can change gradation smoothly by converting four projector's response functions into a response function of the HDR display.
View full abstract
-
Motoaki KURAMOCHI, Takuya OGAWA, Shoji YAMAMOTO
Article type: Article
Session ID: ME2011-13
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We have developed the manufacture and design support system by using projection display based on argument reality. At the practical implementation, ARToolKit is useful for matching the position between virtual and real object. However, the marker necessary to use the ARToolKit is interruption for design support system, since the marker overlaps with the design or image. Therefore, in this paper, we examined the technique for matching the object to the projection image without the marker by using the natural feature detection. As the result, a visual interruption in the design support was able to be removed dramatically compared with the conventional method, while minimum edge for the feature point emphasis was needed.
View full abstract
-
Ken ISHIBASHI, Kazunori MIYATA
Article type: Article
Session ID: ME2011-14
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Conventional research on Non-photorealistic Rendering proposed various rendering methods based on some techniques and knowledge such as brush stroke and the use of color. However, Painterly images require the addition of Kansei information, because Kansei information is an integral part of a painting. In this paper, we propose a novel method of converting realistic photographs to graphic designs, using filtering and color conversion by specifying an abstract word. Our aim is to generate painterly image with Kansei information to express user's Kansei information.
View full abstract
-
Sho WATANABE, Masahide KANEKO
Article type: Article
Session ID: ME2011-15
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes a method to make a mobile robot with auditory sensors (microphones) sense humans in an occluded area where visual sensors don't work well. Conventional mobile robots mainly use visual sensors such as monocular camera, stereo camera, and omni-directional camera, or laser range finders in recent years to recognize surrounding objects and humans. These sensors use visible or infrared light, thus when there is an object which obstruct the light, they cannot recognize humans behind the obstacle. To solve this problem, this paper utilizes auditory information to sense humans in occluded areas where the visual sensors can't get information. Even if some objects or walls obstruct the view, a robot can hear reflected or diffracted sounds. The proposed method is examined in several types of occluded environments. Experimental results show that the proposed method can recognize humans in occluded area and can avoid the unexpected collision with them.
View full abstract
-
Aya NAGAOKA, Masashi OKUDAIRA
Article type: Article
Session ID: ME2011-16
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this article, a method is proposed to generate music based on color and motion information of associated video. The Tempo and key are selected in advance. First chord is set C/Am. The successive chords are determined based on the difference of the sum of R, G, and B values of the dominant color pixel of the frame and that of the first frame. Each melody step is decided based on the displace amount of characteristic points. Music generated with videos are felt acceptable.
View full abstract
-
Naoki UMETSU, Hisae SHIMAMURA, Hiroto INOUE, Mie SATO, Masao KASUGA, H ...
Article type: Article
Session ID: ME2011-17
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Today, there are various contents based on VR (Virtual Reality) technology. In the development of VR contents, it is important to present objects in the virtual space to be recognized easily. However, the method for designing the 3D soundscape in VR contents have not been studied enough. This study is a fundamental investigation for developing a 3D audio navigation system adapted to head motion and position. From a result of experiment, we thought that a position of target sound source was easily recognized when the target and other sound source stood in proper position.
View full abstract
-
Manami OKAMOTO, Mie SATO, Masao KASUGA
Article type: Article
Session ID: ME2011-18
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this study, we aimed to develop the system that can support to cultivate elementary schoolchildren's sensitiveness in education. The system is based on "Imagination expansion effect."Imagination expansion effect is obtained by adding sound information into movie contents. This system is applied to creating movie contents for elementary school education. In this paper, we proposed an automatic music classification system as a part of automatically movie generating system. From our experimental result, we investigated feature quantities of music that is necessary in the case of classifying the music into emotion expressions.
View full abstract
-
Kota TAKAI, Kodai CHIBA, Takeshi FUJIMURA, Junya HIRATA, Tatsushi GODA ...
Article type: Article
Session ID: ME2011-19
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Technologies re-creating piano performance in the form of CG animation are eagerly anticipated by people working in various fields, such as content production and music education. We are promoting research into the analysis and CG expression of realistic and natural-looking piano fingering. This paper proposes a method that automatically generates an interactive animation of a pianist playing the piano from a musical score in real-time. The optimal piano fingering and the trajectories of all control points in the bony frameworks of hands based on the fingering are determined by using optimized algorithms. The piano-playing motion generated automatically is humanized at the performance skill level that users expect, and the real-appearing skin texture is simulated by the GPU (graphics processing units) rendering in real-time. These methods can be applied to many other areas, such as designing a training support system for piano practice or for contents creation.
View full abstract
-
Tatsushi GODA, Shinichi FURUYA, Haruhiro KATAYOSE, Hiroyosi MIWA, Nori ...
Article type: Article
Session ID: ME2011-20
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Our goal was to find the organization of finger and arm movements. Six experts and six amateurs were asked to strike two keys alternately with their thumbs and little fingers. Kinematics and the muscle activities of the finger, thumb, and forearm were recorded by a motion capture system and electromyography. The results showed that experts' had greater elbow, lower joint of the thumb, and little finger movements than amateurs. Additionally, we found that experts demonstrated less muscle activity when compared with amateurs. These results indicate that expert pianists enhance the efficiency of their movements by taking advantage of the proximal joint.
View full abstract
-
Tomohiro Isono, Yutaka Nakano, Toshiyuki Yoshida
Article type: Article
Session ID: ME2011-21
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Shape From Focus/Defocus (SFF/SFD) has been known as a depth estimation technique by using focus information. The technique estimates the depth for each point of a target object based on a focus measure extracted for each block on a set of multi-focus images. The authors have proposed a 3-D whole shape reconstruction technique, which first obtains a set of multi-view and multi-focus images for a target object and then integrates the shape models obtained from each viewpoint with the SFF/SFD. In this technique, the depth estimation error is expected to be reduced by using not only a single set of multi-focus images from a single viewpoint but also another set from another viewpoint together. Although we have realized such an idea in our previous report, the criteria for estimating depth values was not necessary optimal. This report thus introduces a new criteria which uses the derivative of an approximating function for the set of the focus measures, and gives some experimental results to show its effectiveness.
View full abstract
-
Kyosuke MORIAI, Naoki HASHIMOTO
Article type: Article
Session ID: ME2011-22
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, aerial 3D displays which can project stereo-images in the air have been studied actively. Although they can provide realistic and intuitive 3D images, they require large and special devices which prevent daily use in our accessible environments. So, in this research, we propose a method of making an aerial 3D display with ordinary space-filling objects. We capture the 3D architecture of the objects and use them as space-filling materials for the aerial stereo-image projection with arbitrary space and size. In this paper, we especially show the result of 3D measurement for achieving our novel display system.
View full abstract
-
Takuma NAKAMURA, Naoki HASHIMOTO
Article type: Article
Session ID: ME2011-23
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
A dress-up is greatly attractive for many people, in addition to be apart of our daily life. Even though we want to enjoy many kinds of clothes, it is actually difficult with physical and social limitations. Therefore, in this research, we propose a virtual dress-up system with image projection technology. This system changes the color of real clothes into another one with a dynamic luminance correction technique. In order to achieve that, we precisely estimate the response function of a projector, and identify and control the color of the target clothes.
View full abstract
-
Koji YAMADA, Taichiro HIRANO, Atsushi YAMASHITA, Toru KANEKO
Article type: Article
Session ID: ME2011-24
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Sensing in aquatic environments meets the difficulty that image distortion occurs by refraction of light due to the difference of refractive indices of air and water. This paper proposes a 3D measurement method of objects in water by an underwater robot equipped with fisheye stereo cameras which enables the robot to do its task safely with a wide field of view. The method removes the effect of refraction to realize accurate three-dimensional measurement. Experimental results show the effectiveness of the proposed method.
View full abstract
-
Naoto MACHIDA, Masaaki KITAOKA, Atsushi YAMASHITA, Toru KANEKO
Article type: Article
Session ID: ME2011-25
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
A biped robot needs navigation to the destination in achieving its task. This paper presents a method by which a biped robot recognizes its surrounding environment, plans the path to the destination, and moves there. The position and the posture of the biped robot are calculated by landmark observation with a stereo camera system installed on the robot. In order to make it possible for the robot to execute path planning, a grid map is generated by using the surrounding environment information obtained from the stereo camera images.
View full abstract
-
Yuki OKADA, Hitoshi KIYA
Article type: Article
Session ID: ME2011-26
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes a new 2D-3D conversion method to stereo images. A 2D-3D conversion method extracts a target object and generates stereo images by taking account into stereoscopic parallax. The proposed method inpaints the blank area of background caused by object extraction. Thanks to this feature, the proposed method serves the versatility in derivation of stereo images. This method is expected to implore the naturality of 3D scenes.
View full abstract
-
Ling Xiao, Masahide Kaneko
Article type: Article
Session ID: ME2011-27
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Hand raising actions can be considered as the initial stage of communications between humans and a robot including greeting, approval and alert. Their recognition will help to develop applications such as a service robot, remote education and assistance to physically handicapped people. This paper proposes a recognition method of hand raising actions for users sitting at arbitrary positions in the room, where multiple tables are arranged. Color of each user's clothes is captured by a camera at the entrance of the room and a laser range finder is installed at a corner of the room to obtain the range information. Both informations are utilized to detect the hand raising actions and to identify the hand raising user by a stereo camera mounted on a robot.
View full abstract
-
Tomoaki Itaya, Masahide Kaneko
Article type: Article
Session ID: ME2011-28
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Each audience will choose his/her standing position according to his/her own decision so that he/she will see a presenter better at demonstration sale and public speaking. But when the number of audience becomes large, it is difficult for each audience to pay attention to other audience's fields of view and some audience cannot keep good positions to see the presenter. This paper proposes a method to grasp positions of multiple audience by using an overhead camera and to coordinate their positions through the verbal interaction with a robot so that more number of audience can keep better standing positions to see the presenter. A robot simulates the best formation according to personal space of each audience. Then a robot coordinates the position of each audience through conversation considering the reaction of each audience.
View full abstract
-
Takahiro ISHIKAWA, Masashi OKUDAIRA
Article type: Article
Session ID: ME2011-29
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this article, a relevant website visualization method is proposed. First, linked websites are listed out. Then, number of key words coincide with those of the base site and co-linked sites is counted for each site. Those sites which have high score are arranged in a cylinder based on the keywords group so that the relevance between the base site of linked sites are clearly shown.
View full abstract
-
Yusuke AKIMOTO, Yo NAKANISHI, Hiroto INOUE, Mie SATO, Masao KASUGA
Article type: Article
Session ID: ME2011-30
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, movies with caption have been positively used for learning English. The general notation of caption is English or Japanese sentences. Our previous study proposed an annotated caption as a tool of learning English. This paper report effects of notating caption, including the annotated caption, on learning English.
View full abstract
-
Yo NAKANISHI, Yusuke AKIMOTO, Hiroto INOUE, Mie SATO, Masao KASUGA
Article type: Article
Session ID: ME2011-31
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, the movie with English caption has been used for learning English because the viewing environment of video contents is wide spread. Though, the ordinary caption is Japanese or English sentences. The annotated caption is considered effective in learning English. In this paper, we examine the eye movements when viewers watch the movies with annotated caption and study caption annotation based on eye tracking to learn English.
View full abstract
-
Hiroki SHIMADA, Kazuo OHZEKI
Article type: Article
Session ID: ME2011-32
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, the research that uses the picture processing is done by a lot of sports because of the analysis of the action, the strategy, and the form, etc. In it, it is a purpose to reduce time like the record and the analysis, etc. by the hand work. The purpose of the present study aims at the construction of the system that can process by a consecutive rally and gives the player powerful information.
View full abstract
-
Hiroaki YUI, Atsushi YAMASHITA, Toru KANEKO
Article type: Article
Session ID: ME2011-33
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Sometimes camera images taken through window glass have reflections owing to reflectivity of the glass surface. This phenomenon is noticeable in nightscape images because reflection intensity is considerable compared to the outside objects. This paper proposes a method for removing reflections of objects which are parallel to the glass surface from nightscape images of distant view. The method uses a parallel stereo image pair, in which there are no disparities between corresponding points of outside objects in left and right images but are constant disparities for reflecting objects. Results for preliminary experiment are given to show the effectiveness of the proposed method.
View full abstract
-
Rie MIYAKI, Atsushi YAMASHITA, Toru KANEKO
Article type: Article
Session ID: ME2011-34
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Color perception varies from individual to individual. It is difficult for a person with defective cones in the retina to recognize the difference between specific colors. In this case, he or she may face difficulties in daily life. This paper proposes a method for presenting color information by using a projector camera system. The system automatically detects if there are color combinations of difficult discrimination in the scene. The system, it projects border lines or color names on real object surfaces which have those specific colors so that color vision defectives can discriminate the colors.
View full abstract
-
Fumiya TSURUMI, Atsushi YAMASHITA, Toru KANEKO
Article type: Article
Session ID: ME2011-35
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
When an image is taken through a wire net, the wire net influences the image quality of a subject. Therefore, it is necessary to restore the subject image without wire net overlap by image processing. A method for wire net removal has been already proposed using multiple, focus images. It removes the overlapping effects of defocused wire net. However, this method determines the essential parameter of blurring radius manually. This paper proposes a method for determining the parameter automatically to restore the subject image without wire net. Experimental results showed the effectiveness of the proposed method.
View full abstract
-
Takuya OGAWA, Motoaki KURAMOCHI, Shoji YAMAMOTO
Article type: Article
Session ID: ME2011-36
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we propose simple and robust estimation method of light position calculated by shadow of real objects in mixed realty. In general, the estimation of light position was performed by using the captured image of minor boll to keep the photometric consistency. However, it is difficult to obtain only the light position from processing the captured image, since the image of reflected by mirror boll has complexity caused by various ambient objects. Therefore, we estimate the light position by using shadows which consist of four pushpins around the square maker. The gravity and shape of shadows is available to calculate the light position accurately, while the shape of pushpins is already known. As a preliminary result of our proposed method, the shadow of the virtual object almost corresponded with real shadows.
View full abstract
-
Yumiko Susuki, Yutaka Nakano, Toshiyuki Yoshida
Article type: Article
Session ID: ME2011-37
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Some of documents, which have a fixed format and are updated periodically, are often modified very partially, resulting in very similar documents before and after the modification. The authors have proposed an automatic comparison and detection technique for such a modification in a pair of printed documents. However, the technique fails to detect modifications due mainly to a difference of region segmentation of corresponding characters between two document images. This paper thus proposes a new technique to overcome the problem, and improves the detection accuracy. The experimental result given in this paper illustrates that the proposed technique achieves a precision ratio of 97% under a recall rate of 100%, and that the errors due to the region segmentation can be completely removed.
View full abstract
-
Takuya NAKAJIMA, Takefumi TOZAWA, Koichi MORI, Kyoko HASEGAWA, Susumu ...
Article type: Article
Session ID: ME2011-38
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Implicit surface reconstruction is one effective way for the purpose of surface modeling from point data. Implicit surface model is suitable to smooth surface representation and precise construction details of the object. However, the model requires high memory cost depending on the complexity of the surface. In this paper, we propose a method for data compression of implicit surfaces using the technique of wavelet transform.
View full abstract
-
Sumio YANO
Article type: Article
Session ID: MB2011-39
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
The research statuses from stereoscopic to three-dimensional image systems are briefly described. As for stereoscopic images, the psychophysical effects are greater than these of monoscopic images. Other hand, there are some un-expected effects in the stereoscopic images. For improving these defects, the developments of multi-view stereoscopic and three-dimensional image systems are progressed. Multi-view stereoscopic image systems reduced partially defects of stereoscopic image systems. Three-dimensional image systems, which can display objects as the same of the real world, are developed. For examples, three kinds of three-dimensional image systems, whish are consistent with the characteristics of human visual fields, are introduced in this article.
View full abstract
-
Article type: Appendix
Pages
App1-
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Article type: Appendix
Pages
App2-
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Article type: Appendix
Pages
App3-
Published: February 12, 2011
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
