-
Article type: Cover
Pages
Cover1-
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Article type: Index
Pages
Toc1-
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Ryo NAGATSUKA, Yuichi TAGUCHI, Takeshi NAEMURA
Article type: Article
Session ID: HI2009-1/ME2009-1/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper presents a method that controls appearance of each object in a 3D TV system using a camera array and an integral photography display. The method allows a user to easily extract objects in multi-view images, and to independently control the positions of the extracted objects on the display by using different rendering parameters for diffrent objects. It enables interactive editing of the 3D scene such as emphasizing the depth of a certain object and changing the relative positions of objects, rather than presenting the captured scene as it is.
View full abstract
-
Kohei Omura, Masayuki Sugawara, Yuji Nojiri
Article type: Article
Session ID: HI2009-3/ME2009-3/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
The motion blur caused by aperture time can be improved by reducing the rate of a camera shutter opening. However, with reducing the aperture ratio, the "jerkiness" which is another artifact about the movement perception appears. We evaluated the jerkiness of natural images for various frame rates and aperture ratios. The subjects consisted of nonspecialists. In to 240 frame/sec, we got the result that there was a picture quality improvement by raising a frame rate.
View full abstract
-
Yukihiro BANDOH, Seishi TAKAMURA, Kazuto KAMIKURA, Yoshiyuki YASHIMA
Article type: Article
Session ID: HI2009-4/ME2009-4/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Over the past decade video acquisition rate has brokenthrough over 1000[Hz]. Current display systems, unfortunately, have a maximum frame-rate about 60 [frames/sec]. The display systems are definitely slower than the camera system. When a high frame-rate video (over 1000 [frames/sec]) is played in real time, however, it is necessary to down-sample the frame-rate of the video. Conventional methods keep constant the interval of neighbouring frames after down-sampling. It is based on the assumption that non-constant intervals would yield jerky motion, i.e. perceivable discontinuity in the optical flow field. However, this assumption is not true for high frame-rate video. Such video has many frames around the "constant interval" sub-sampled frame that do not yield jerky motion since the optical flow field displacement within the frames is too small to be perceived. In other words, constant frame interval is not always needed when down-sampling high frame-rate video. Relaxing the interval requirement allows greater flexibility in selecting the frames to be sub-sampled, which allows the number of bits in a sub-sampled sequence to be reduced, subject to realizing the same image quality. In this paper, we propose a temporal down-sampling method for high frame-rate video. The proposed method features the adaptive selection of down-sampled frames so as to minimize inter-frame prediction error. Our method can improve the PSNR of prediction signal by 0.13[dB] to 0.23[dB] compared to simple sub-sampling with constant interval.
View full abstract
-
yoshinori INABA, jumpei KOCHI, hanae ISHI, jiro GOHBA, shigeru AKAMATS ...
Article type: Article
Session ID: HI2009-5/ME2009-5/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this work, preliminary experiments were conducted on a novel method seeking physical parameters of 3D objects that might strongly contribute to their visual impressions through computational investigation of the impression ratings made on learning samples. Car body was selected as a class of 3D objects to be investigated, and a set of the 3D shape data for various car types was obtained by measuring plastic models of different car lines with a range-finder. Morphable 3D model of car bodies describing the variation of their appearance in terms of a smaller number of parameters was obtained by applying PCA to a set of high dimensional vectors representing the 3D shapes. A preliminary method was proposed to transform the visual impression of the 3D object by manipulating the parameters defined in the morphable 3D model. Results of psychological experiments conducted by applying the method of paired comparison indicated possible validity of the proposed method. A new scheme is also proposed to properly re-sample a novel object having even a peculiar shape so that such object could also be represented by the morphable 3D model.
View full abstract
-
Zhen LI, Atsushi UEMURA, Hitoshi KIYA
Article type: Article
Session ID: HI2009-6/ME2009-6/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we propose a FFT based full search BMA (block matching algorithm) with SSD (sum of squared difference) criterion. The proposed method focus on the relationship between the cyclical cross-correlation and the SSD criterion. By using FFT to calculate the cross-correlation between data in different sizes, the proposed method has a large increase in speed of block matching. Especially, when the intended block is made up of real signals we can finish matching two intended blocks in one time. In the simulation of motion estimation, it shows that the proposed method is more accurate than the conventional method with SAD criterion and is faster than the conventional method with SSD criterion. Compare to the direct SSD full search, the proposed method's processing speed is 10 to 700 times faster.
View full abstract
-
Ryoji HASHIMOTO, Hiroshi TSUTSUI, Takao ONOYE, Tomohiro IKAI
Article type: Article
Session ID: HI2009-7/ME2009-7/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Distributed Video Coding (DVC), which is a new image compression paradigm, attracts a lot of attention from video researchers. While computational cost of DVC encoding is lower than that of MPEG-2, low coding efficiency has been an issue for practical DVC applications. In this paper, a likelihood estimation method for transform domain DVC is proposed, which uses Cauchy distribution as a virtual channel model. In the decoder, virtual channel is estimated by utilizing error between forward and backward predicted images for each frequency component. Likelihood is obtained from estimated virtual channel. Simulation results show that the proposed method can estimate the error rate with 1.6% error on average.
View full abstract
-
Seishi TAKAMURA, Masaaki MATSUMURA, Yoshiyuki YASHIMA
Article type: Article
Session ID: HI2009-8/ME2009-8/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Evolutionary methods based on genetic programming (GP) enable dynamic algorithm generation, and have been successfully applied to many areas such as plant control, robot control, and stock market prediction. However, conventional image/video coding methods such as JPEG and H.264 all use fixed (non-dynamic) algorithms without exception. In this article, we introduce a GP-based image predictor that is specifically evolved for each input image. Preliminary results demonstrate 1.4% and 1.8% entropy reduction (overhead included) against the optimal linear predictor and CALIC's gradient adjusted predictor, respectively.
View full abstract
-
Atsumu TANAKA, Tomonori IIDA, Kiyoshi NISHIKAWA, Hitoshi KIYA
Article type: Article
Session ID: HI2009-9/ME2009-9/AI
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We propose a one-pass rate control method given a target PSNR for JPEG 2000. The proposed method is based on PCRD-opt algorithm, therefore computational costs for rate controling would not increase compared to those of the rate control method given a target bit-rate. Besides, bit-rate could be controllable under the condition that an upper limit of bit-rate and a target PSNR are simultaneously given. By applying the proposed method to a video sequence compression, we could prevent to generate low PSNR frames. Also, by employing the proposed method with a rate control method given a target bit-rate, it is possible to efficiently reduce the amount of coded data.
View full abstract
-
Sunmi KIM, Hirokazu TANAKA, Takahiro OGAWA, Miki HASEYAMA
Article type: Article
Session ID: HI2009-10/ME2009-10/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we propose an error resilient and error concealment method for 3D wavelet video coding over wireless transmission. The proposed method consists of the following two schemes; the first one is a 3D wavelet encoding with dispersive grouping, and the second one is an estimation scheme for dispersive grouped elementary streams using a Minimum Mean Square Error (MMSE) algorithm at the decoder. Experimental results indicate that the proposed algorithm provides better performance from aspects of PSNR (Peak Signal-to-Noise Ratio) and visual quality.
View full abstract
-
Mamoru DOKE, Narichika HAMAGUCHI, Hiroyuki KANEKO, Seiki INOUE, Yukio ...
Article type: Article
Session ID: HI2009-11/ME2009-11/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We have been researching and developing the "TV4U (TV for you)" system, as a novel TV system, which uses computer networks as the infrastructure and covers all functions which are TV program production, delivery and viewing. We have already realized the sequence of functions which are easy TV program production, delivery and viewing. So far, we have especially focused on development of upgrading of TV program production function. This time, we have reconsidered the "TVBrowser" which is TV program viewing system on TV4U. And we have developed a new TVBrowser, which is a sophisticated TV program browser, as a new TV program viewing system provides user-intuitive interfaces and convenient and useful services such as an automatic introduction program generation and user-to-user communication.
View full abstract
-
Kazunori MATSUMOTO, Masaki NAITO, Keiichiro HOASHI, Jian Ming Wu, Yasu ...
Article type: Article
Session ID: HI2009-12/ME2009-12/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
IEICE (The Institute of Electronics, Information and Communication Engineers) provides a word template file for the Technical Report of IEICE.
View full abstract
-
Akihito WATANABE, Katsuhiro MASAKI, Mie SATO, Masao KASUGA
Article type: Article
Session ID: HI2009-14/ME2009-14/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Recently, various internet businesses are expanded. Especially, the business form that combines the store sales and the net sales attracts attention. It is said that such a business form obtains the synergy effect, and it is called click-and-mortar. This research aims to extract features of the store sales and the net sales. Furthermore, based on a extracted features, we seek a marketing method in hope of synergy effect expansion of the click-and-mortar.
View full abstract
-
Yoshihiro NOMOTO, Kyohei HAYASHI, Yutaka ISHIBASHI
Article type: Article
Session ID: HI2009-15/ME2009-15/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this report, we discuss a video server system that performs streaming transmissions of Consumer Generated Video (CGV) files. The CGV files are placed in iSCSI (i.e., SCSI over TCP) storages at each owner's home, and the iSCSI storages are connected from one video server on demand. To guarantee QoS, a managed IP network is used between the video servers and iSCSI storages. The managed IP network means the Next Generation Network (NGN). Since iSCSI uses the TCP protocol, the network delay influences the iSCSI throughput largely. The network delays between video servers and iSCSI storages depend on the locations of the servers and storages, and the network topology. Although the iSCSI storages have different network delays for the video servers, it is important to design the system so that the streaming transmission quality does not depend on the network delays. To ensure the fairness of streaming transmissions, we adopt a technique to dynamically control the number of TCP connections in the iSCSI protocol. Simulation results show that the CGV file streaming transmission quality at 3Mbps is stable on the condition that the Round Trip Times (RTT) between servers and storages are less than 100ms.
View full abstract
-
Tatsurou FUJIWARA, Masashi KAMEDA
Article type: Article
Session ID: HI2009-16/ME2009-16/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Sketch is a fundamental skill for drawing pictures. For learning sketch, it is most important that information between teacher and students is exchanged interactively during students draw pictures. We have proposed the interactive pencil-drawing learning support system for beginners. This paper presents the system configuration to realize the learning of proportion. It is clarified in our experimental results that the proposed system is effective to learn the sketch by one's self instead of the conventional sketch learning methods.
View full abstract
-
Atsushi ITAKURA, Hideo MAKINO, Wataru HIOKI, Ikuo ISHII, Hideki KOMAGA ...
Article type: Article
Session ID: HI2009-17/ME2009-17/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In order to provide assistance following a natural disaster, it is necessary to acquire accurate information. Aerial photography can provide such information. We propose a method for quickly acquiring images across the entire area, using a camera with a fish-eye lens and a radio controlled helicopter. However, fish-eye lenses also distort the images and this distortion makes interpretation of the information very difficult. Therefore, to improve interpretation we develop a perspective projection image in real-time. The image also rotates because the airframe isn't stable. A method for removing the rotation of the image and presenting an easily interpreted picture is discussed.
View full abstract
-
Kenji HARA, Atsuhiko MAEDA, Hirohito INAGAKI, Youichi TAKASHIMA, Minor ...
Article type: Article
Session ID: HI2009-18/ME2009-18/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We propose a facial color beautification method for video communication. The video-captured face often disappoints us, especially the subjects themselves, because the lighting environment is often poor. Inadequate lighting makes the face look depressed or tired. The best solution is illuminating the face directly with good light, but it is not always possible to realize a good environment in everyday situations. Our solution is to retouch the captured video. Our method identifies the distribution of skin color and the eye luminance of a detected face, and converts the color distribution of the entire image by histogram transformation. This method provides three strong points, which are 1) converting video in real-time, 2) beautifying the facial color even if it is captured with inadequate light, and 3) creating a variety of face "styles" by controlling the transformation parameters.
View full abstract
-
Yasutaka HATAKEYAMA, Takahiro OGAWA, Satoshi ASAMIZU, Miki HASEYAMA
Article type: Article
Session ID: HI2009-19/ME2009-19/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
An accurate video retrieval method based on features of video materials is presented in this paper. The proposed method applies canonical correlation analysis to features obtained from video materials to enable definetion of similarities between different features in the new variate space. Furthermore, features obtained from the video materials are regarded as Web pages and their adjacency matrix us calculated based on the defined similarities. Then, by applying link analysis to the obtained adjacency matrix, Web community extraction based on features of video materials can be realized. Consequently, the proposed method can achieve accurate result of retrieval based on features of video materials.
View full abstract
-
Kousuke MATSUSHIMA, Kouhei SENBA, Keiichi UCHIMURA, Shoji ESAKI
Article type: Article
Session ID: HI2009-20/ME2009-20/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Pedestrian-vehicle accidents account for about one third of the total automotive related fatality and injury. So the inclusion of pedestrian protection systems such as external airbags is being consider as a solution to preventing pedestrian fatality and injury. However, such systems require knowledge of pedestrian presence for correct activation. In this work, it proposes the pedestrian recognition technique for considering three-dimensional information based on the image obtained from right and left cameras set up in the vehicle. These techniques are applied to the road scene images, and the confirmation and the verification of effectiveness are performed.
View full abstract
-
Jinge Wang, Shintaro Ono, Katsushi Ikeuchi
Article type: Article
Session ID: HI2009-21/ME2009-21/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We propose a new expression of space-time image which is called "Temporal Height Image (THI)". Assigning a gray value which is in proportion as height to all objects in the space-time volume, we can get THI by looking at the space-time volume above. THI is supposed to be used at urban area mainly. The THI has similar concept as EPI (Epipolar Plane Image), whereas, EPI can't extract edge of object stably. Here, THI can overcome the shortage of EPI by using height of buildings effectively. In this paper, we constructed THIs from actual urban image and 3D-city model, and then aligned these two THIs by DP-matching to verify the effectivity of THI. Then, using the matching result, we tried taking texture mapping onto building models.
View full abstract
-
Ryota Matsuhisa, Shintaro Ono, Hiroshi Kawasaki, Atsuhiko Banno, Katsu ...
Article type: Article
Session ID: HI2009-22/ME2009-22/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Estimation of Vehicle Motion has generally done by GPS, accelerometer and so on. However, higher accurate method is necessary for measurement of vehicle characteristics, Automatic Driving and so on. In this paper, we estimate vehicle motion and surrounding environments by Structure from Motion (SFM) method with omnidirectional image sequences. SFM can estimate those parameters accurately as pixel resolution. However, SFM method is usually unstable dependent on initial parameters and noise. In this paper, we propose a SFM method for omnidirectional image sequences to achieve high accuracy and robustness.
View full abstract
-
Akihiko NISHIMURA, Jeyeon KIM, Takaaki HASEGAWA
Article type: Article
Session ID: HI2009-23/ME2009-23/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper describes a proposal of vehicle positioning system by M-CubITS in plural traffic lane roads, and shows the effectiveness of This system by experiments. The system recognizes its vehicle's position in plural traffic lanes using the driving lane recognition system (lateral positioning) and M-CubITS (longitudinal positioning). We carry out offline experiments on a third parts road. Results show that recognition rate of each frame is 90.0% (without lane changing information), 96.2% (with lane changing information), and that achieves 100% by using time-series data.
View full abstract
-
Michio MIYAMOTO, Hiroyuki YOMO, Oyunchimeg SHAGDAR, Takashi OHYAMA, Me ...
Article type: Article
Session ID: HI2009-24/ME2009-24/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In order to appropriately evaluate performance of inter-vehicle communication systems which support safety driving, it is important to perform evaluation under realistic accident scenarios with suitable application requirements. In this paper, assuming realistic intersection collision scenarios, we perform simulation evaluation of CSMA based inter-vehicle networks based on experimental guide-line defined by ITS Info-communications Forums. We compare performance of CSMA based systems with that of our proposed MM-SA (Multi-Carrier Multi-Code Spread Aloha) system employing CDMA (Code Division Multiple Access) scheme. The simulation results show that MM-SA has superior performance to CSMA based systems thanks to high robustness to high level of interference as well as to hidden terminal problem.
View full abstract
-
Yuichi HAGITO, Takaaki HASEGAWA
Article type: Article
Session ID: HI2009-25/ME2009-25/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper describes a proposal of the multi-hop communication scheme using heavy vehicles in the multi-class zone ITS communication scheme and its performance evaluation. The proposed scheme improves the communication performance by introducing the repeating technique of the synchronization signal and frame control signal using heavy vehicles and multi-hop communication technique of the first and second category intersection RVC (Road to Vehicle Communications) using heavy vehicles to the vehicle that cannot perform RVC owing to shadowing by heavy vehicles. Performance evaluations are carried out by a microscopic traffic flow simulator including the communications, and improvement of communication performances by introduction of the proposed scheme are shown. In addition, it is shown that the required frequency bandwidth achieving 90% of packet arrival rate is 38.2MHz, and the required frequency bandwidth achieving 97% of packet arrival rate is 41.6MHz with the main roadway consists of five lanes and the intersecting roadway consists of three lanes.
View full abstract
-
Katsuyoshi SATO, Kyoichi IIGUSA, Hiroshi HARADA
Article type: Article
Session ID: HI2009-26/ME2009-26/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Interest in ITS (Intelligent Transport Systems) technology has been growing in recent years, because it may be used to alleviate problems such as traffic jams and traffic accidents. Wireless communications technologies form the basis of enabling ITS to be implemented and among related technologies, road-vehicle communication (RVC) and inter-vehicle communication (IVC) play important roles. In ITS, Inter-Vehicle Communications (IVC) have an important place in assisting safe, effective, and economical driving, as well as supporting automated driving such as Advanced Cruise-Assist Highway Systems (AHS). In these applications, the transmission quality of IVC can be a matter of life or death for many users. One of the application examples is the collision avoidance system in the NLOS environment at intersections. Because the strong diffraction is expected, 720MHz band is allocated for this ITS application. The propagation characteristics in this band for IVC is not necessarily clear enough now. Therefore, it is important to know this characteristics in this frequency band to develop the reliable system. This paper shows propagation characteristics for IVC in 720MHz band at intersection structure.
View full abstract
-
Daisuke TAKASAO, Tomohiro YENDO, Masayuki TANIMOTO, Toshiaki FUJII, Yo ...
Article type: Article
Session ID: HI2009-27/ME2009-27/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we propose a road-to-vehicle visible communication system for ITS. In this system, a LED traffic light is used as transmitter and a photodiode is used as receiver. Moreover a telephotographic lens is used to receive a light from long distance, and a video camera is used for tracking a traffic light. Therefore we need to detect traffic light position. Thus we proposed high-speed detection method of traffic light that used the feature such as color information and shape information, etc. Then we had done experiment and it was verified that was a high-speed method.
View full abstract
-
Satoshi OKADA, Tomohiro YENDO, Masayuki TANIMOTO, Toshiaki FUJII, Yosh ...
Article type: Article
Session ID: HI2009-28/ME2009-28/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we propose a road-to-vehicle visible communication system for ITS. In this system, a LED traffic light is used as transmitter and a photodiode is used as receiver. There are several problems associated with applying visible light communication to the field of ITS. It is necessary to receive information from long distance. And tracking the transmitter for a certain moving distance of the vehicle is also important. We applied an imaging optics to receive information over long distance, and two cameras are used to solve the relationship between the transmitter and the receiver position changes with time, and vibrational correction technique is also fixed to the system. We developed algorithms to track the transmitter. The experiments were conducted to confirm the proposals.
View full abstract
-
Kyota AOKI, Shunsuke MORO
Article type: Article
Session ID: HI2009-29/ME2009-29/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
There are many shot-change detection methods. They work well in videos without large motions and fast camera works. But, those methods can't withstand recent news videos or live sport videos that include large motions and fast camera works. This paper proposes more robust method for detecting shot changes and confirms its performances with experiments. The method is a composition of motion estimation and the best color based shot-change detection method.
View full abstract
-
Norihiro KAKUKOU, Takahiro OGAWA, Miki HASEYAMA
Article type: Article
Session ID: HI2009-30/ME2009-30/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper extends a particle filter and applies it for a flow estimation method based on a Helmholtz decomposition theorem. The proposed method utilizes a state transition model including two state variables affecting each other and an observation model affected by a previous observation. These models do not satisfy traditional particle filter's assumptions that the current state variable depends only on itself at the previous time and the current observation depends only on the current state variable. Therefore, the proposed method utilizes the new assumptions satisfying the above models for the extension of the traditional particle filter. Furthermore, the modified one is applied for the flow estimation method based on the Helmholtz decomposition theorem. The flows whose directions are forward and backward are utilized as the two state variables. For each state variable, the state transition model is defined from properties of the flows. The observation model and an observation density are defined from a gradient-based method and a model of the Helmholtz decomposition theorem extended based on a transitional component. The modified particle filter with these definitions can realize the flow estimation based on gradients of intensities, rotation, divergence, and translation in such a way that the estimation errors included in the previous flows do not affect its scheme. Consequently, an accurate flow estimation can be achieved.
View full abstract
-
Katsu OU, Junzo KAMAHARA
Article type: Article
Session ID: HI2009-31/ME2009-31/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Promoted by the professional and diverse demands of image retrieval the technique of content-based image retrieval (CBIR) becomes more mature. In CBIR, Color characterization of an image is one of the most important feature and is widely used. This article describes a technique that divides the image by center weighted division and calculates the cumulative histogram of each divided images in L^*a^*b^* color space. Similarity measure uses the Euclidean Distance. According to the experimental results, Our proposed method requires additional computational burden, but it is help to improve the image retrieval effectiveness.
View full abstract
-
Keigo Kitamura, Toshihiko Yamasaki, Kiyoharu Aizawa
Article type: Article
Session ID: HI2009-32/ME2009-32/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We have been developing a food-logging sysytem which can distinguish food images from other images, analyze the food balance, and visualize the log. In this paper, image analysi of this system is expalained. This paper describes the improvement in accuracy using the Bag of Features model in the extraction of food images and the estimation of food balance in this system. The experiments shows that the feature detection by the DoG processing is the highest-precision. The accuracy of the extraction of food images has improved by 3% and one of the estimation of the balance has improved by 5%.
View full abstract
-
Nanae TANAKA, Masashi OKUDAIRA
Article type: Article
Session ID: HI2009-33/ME2009-33/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes a guidance method of positions and directions of textured pavement blocks on a cellular phone. Concerning image processing and navigation, low cost algorithm and easy-to-follow voice guidance are adopted. A hyperplane that discriminates pixels of the blocks and others are determined with Support Vector Machine in advance. On the cellular phone, first, discrimination of pixels in each 3×3-divided region of a photo is carried on using the hyperplane. Then, based on connection patterns of block-existence regions, degenerated navigation patterns for humans are decided. Experimental results with over 90 images show the effectiveness of the proposed method for standard images. It takes about three seconds for image processing.
View full abstract
-
Takaaki HASEGAWA
Article type: Article
Session ID: HI2009-34/ME2009-34/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper describes a diffusion scenario of electric vehicles from the viewpoint of systems innovation considering both human society aspects and technological aspects. Firstly, fundamentals of the systems innovation theory and the platform theory are mentioned. Secondly, discussion on mobility from the viewpoint of human-society layer and discussion of electrical vehicles from the viewpoint of elemental techniques are carried out. Finally, R&D, measures and policies are described as a scenario including standardization (universalization) of contactless power feeding and a reasonable payment scheme in a comprehensive way.
View full abstract
-
Yusuke TAKANO, Toshimasa ASO, Takaaki HASEGAWA
Article type: Article
Session ID: HI2009-35/ME2009-35/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper describes effects on both positioning errors and vehicle detection misses on performance of the Advanced Demand Signals II (ADS-II) scheme using image sensors. Firstly, positioning errors are modeled and introduced to a conventional microscopic traffic flow simulator for Route 463. Simulation results show that the effects on image sensors' positioning errors on performance of the ADS-II scheme are little at both the off-hour (4:00 a.m.) and the rush-hour (7:00 a.m.). Secondly, effects of vehicle detection misses are evaluated. Results show effects of these misses are little at the off-hour for practical application. On contrast, at the rush-hour, these misses degrade the average idling time per person of the ADS-II scheme by approximately 5%, but it is smaller than that of the coordinated control scheme by approximately 30%. Finally, setting heights of image sensors are examined. It is shown that the ADS-II scheme performs well when image sensors are greater than 6m.
View full abstract
-
Tatsunori HIRATA, Haruki KAWANAKA, Masahiro MIYAJI, Koji OGURI
Article type: Article
Session ID: HI2009-36/ME2009-36/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
When measuring and recognizing it with the camera, the presumption accuracy of the camera parameter decides and puts the accuracy of the output of the system. However, for example, the position and a direction may shift gradually by vibration of an installation instrument, a shock and aging, thermal expansion, etc. So the increase in a measurement error is not avoided with the original camera parameter. Moreover, when the camera like the pan, tilt, zoom, etc. is adjusting controlled, it is necessary to update the camera parameter. Therefore we think that we demand an unknown parameter as a restriction condition with the plane symmetry characteristic of the photography object. Generally the parameter estimation in a projection camera model serves as a nonlinear optimization problem. But as update type algorithm, we aim at the method that is computability easily in a short time.
View full abstract
-
Tomoaki HIROOKA, Wataru MITSUHASHI
Article type: Article
Session ID: HI2009-37/ME2009-37/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We propose a method for detecting and tracking vehicles in multilane traffic flow by using a camera installed on an overpass. The image intensities of stationary background scenes such as road surfaces and road markings are estimated through Kalman Filter based approach, and the feature points are detected only on vehicles by using Kanade-Lucus-Tomasi tracker. The Normalized Cuts based cluster analysis is utilized for segmenting the feature points into individual vehicles. Since a prior knowledge of vehicles as to shape, velocity and image intensity is unnecessary, the proposed method is robust to environmental changes in traffic scene, thereby increasing the possibility of an automatic measurement of traffic flow.
View full abstract
-
Kentaro MATSUO, Masayuki MIYOSHI, Tsuyoshi HAMADA, Yuichiro SHIBATA, T ...
Article type: Article
Session ID: HI2009-38/ME2009-38/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
The Phase Only Correlation (POC) method demonstrates high robustness and subpixel accuracy in the pattern matching and the image registration. However, there is a disadvantage in computational speed because of the calculation of 2D-FFT etc. We have proposed a novel approach to accelerate POC method using GPU to solve the calculation cost problem. Using our GPU-based POC implementation, each POC calculation can be done within 2.36 seconds for 256×256 pixels, within 7.92 seconds for 512×512 pixels, and 27.65 seconds for 1024×1024 pixels.
View full abstract
-
Atsushi KATO, Yuji SAKAMOTO
Article type: Article
Session ID: HI2009-39/ME2009-39/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Electro-Holography using a reflective LCD is expected as the 3-dimensional movie display technology in the future. We realized the reflective LCD Electro-Holography allowing for the wide viewing zone and visual field using the Fourier transform optical system. In addition, we estimated the size of reconstructed image, the distance for looking in both eyes, the region that place an object in our optical system theoretically. We performed the optical experiments and confirmed their results with reconstructed images.
View full abstract
-
Hiroyasu Negishi, Daisuke Kisara
Article type: Article
Session ID: HI2009-40/ME2009-40/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
As a difference of embedded graphics products, the corresponding of rich content is expected. An attention to parametric curve drawing of the vector graphics and drawing quality that done not depend resolution of the vector graphics raises. The vector graphics needs a large amount of calculation but embedded CPU has poor performance. The speed up of the vector graphics is expected using hardware. This paper proposed the methods of speed up of the vector graphics using 3D graphics hardware in existing products.
View full abstract
-
Takeshi YAMAGUCHI, Tomohiko FUJII, Hiroshi YOSHIKAWA
Article type: Article
Session ID: HI2009-41/ME2009-41/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we have investigated the computer-generated disk hologram (CGDH). Since general flat format hologram has the limited viewable area, we usually cannot see the other side of the reconstructed object. Therefore, we proposed the CGDH to realize the 360 degrees viewable hologram. In our previous study, the CGDH was realized. However, since the relation between the vertical viewing angle and reconstructed image size is trade-off, the size of the reconstructed image and vertical viewing angle is not enough. Therefore, to improve both parameters, we modified the fringe printer to output the high resolution fringe pattern.
View full abstract
-
Sho MATSUDA, Tomohiko FUJII, Takeshi YAMAGUCHI, Hiroshi YOSHIKAWA
Article type: Article
Session ID: HI2009-42/ME2009-42/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Since a general hologram has a limited viewable area, we usually cannot see the other side of a reconstructed object. There are some holograms that can solve this problem. A disc hologram is well known to be viewable in 360deg. However, Computer-Generated Disc Hologram (CGDH) requires huge calculation amount. By using CPU, for example, it takes 115.6 hours to calculate a high resolution CGDH (167,936x167,936 pixels). In this paper, to reduce the calculation time of CGDH, we employ Graphics Processing Unit (GPU). Since GPU has many shading unit and parallel processing structure, GPU can work the repetitive process well. As a result, we have achieved that CGDH can be generated at 3.2 hours with high resolution. It is 36.1 times faster than CPU.
View full abstract
-
Wonkuk LIM, Miki HASEYAMA
Article type: Article
Session ID: HI2009-43/ME2009-43/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we propose particle filters for tracking soccer players and estimate camera parameters in soccer videos. The proposed method consists of the following two filters; one is a filter tracking position of soccer players, the other is estimates a camera parameter which provide focusing area of the target soccer video. The structure of the proposed particle filter makes use of two parallel filters whose results are simultaneously input the observation model of the other one to track the soccer players and estimate the field area.
View full abstract
-
Mineki TAKAYA, Hiroki TAKIGUCHI, Junji MAEDA
Article type: Article
Session ID: HI2009-44/ME2009-44/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In the early process of the human vision, it is known that people pay preferential attention to salient objects. Many researches have been made to achieve this process on a computer. We propose the method to automatically identify perceptually salient objects in a static image by using the visual attention model based on the early process of human vision. We also propose to apply the method to the detection of human beings by optimizing the feature extraction of the visual attention model and by using nonlinear support vector machines. This research is applicable to detect suspicious individuals in a surveillance image and prevent car accidents.
View full abstract
-
Shougo YAMAUCHI, Tatsuya HIGASHIKAWA, Takehiro HARADA, Junji MAEDA
Article type: Article
Session ID: HI2009-45/ME2009-45/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
The purpose of this paper is perceptual segmentation of natural color images and its application to the segmentation of dermoscopy images, which are closeup images of pigmented skin lesions. We propose the fuzzy split-and-merge algorithm that integrates the L^*a^*b^* components as color features and the statistical geometrical features as texture features by using fuzzy inference. The proposed method is composed of four stages: simple splitting, local merging, global merging and boundary refinement. We perform the comparative experiments of the proposed method in comparison with the human segmented results to dermostrate the effectiveness of the proposed method.
View full abstract
-
Shota IZUMI, Akihiro TABATA, Junji MAEDA
Article type: Article
Session ID: HI2009-46/ME2009-46/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper presents the color quantization of an image using a self-organizing map (SOM) and image segmentation. We try to implement perceptual segmentation of color images using only color quantization without using texture features. We propose an improved SOM that is a modified version of the Ong's method. We also propose to introduce the pre-processing including the correction of achromatic regions based on fracral dimension and the edge-preserving smoothing by using anisotropic diffusion, that is suitable for reducing small regions and realizing better segmentation than the conventional methods.
View full abstract
-
Ikuko UWANO, Masashi KAMEDA, Takashi INOUE, Hideaki NISHIMOTO, Kuniaki ...
Article type: Article
Session ID: HI2009-47/ME2009-47/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
The sulci on the cerebral cortex are important anatomical landmarks and are frequently used for determining the surgical trajectory. We have previously developed a novel computer-based approach to assist in identifying the central sulcus from MR data of patients with various brain tumors based on only the anatomical features of the central sulcus which appeared to be quite efffective. Thus, our method provides accurate guidance for identifying the central sulcus. In our previous method, we extracted the sulci from Curved Planer Reformation (CPR) image, which is two-dimensional image reconstructed by unfolding the cortical surface and calculating whether their parameters matched the parameters of central sulcus. The sulcal extraction process is based on determining the minimum intensity regions, because sulci on T1-weighted MR images have the lowest signal intensity. However, when the sulci include vessel areas, the previous method is problematic: It cannot detect the minimum intensity area, because vessel areas have higher intensity than sulci areas, so the sulci are extracted as discontinuous lines. Therefore, we have improved the process for extracting sulci, taking the approach based on intensity of vessel areas among the sulci.
View full abstract
-
Nobuhiro Fujii, Takashi Komatsu, Takahiro Saito
Article type: Article
Session ID: HI2009-48/ME2009-48/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We have proposed the L1-L2 and L2-L2 color-shrinkage schemes, and have shown that the application of the color-shrinkage schemes to color-image denoising in a redundant wavelet transform domain achieves excellent denoising performance. This paper presents how to derive the color-shrinkage schemes, their iterative update algorithms and their convergence characteristics, the improvement of denoising performance by utilizing the ML estimates of signal variance. Moreover, this paper shows the result of denoising simulations conducted on noisy color-images contaminated with signal-dependent noise equivalent to ISO1600 and/or ISO6400, and demonstrates the superiority of our proposed denoising methods over the existing methods.
View full abstract
-
Tomoki HIRAMATSU, Takahiro OGAWA, Miki HASEYAMA
Article type: Article
Session ID: HI2009-49/ME2009-49/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, an ER (Error-Reduction) algorithm-based method for resolution enhancement of still images is proposed. Since Fourier-domain and object-domain constraints are needed for any ER algorithm-based method, the proposed method introduces the following two constraints for resolution enhancement. The first one is the Fourier-domain constraint that utilizes the Fourier transform magnitude estimated from an autocorrelation function of the target low-resolution (LR) image as that of a high-resolution (HR) image. The second constraint is the object-domain constraint that utilizes the frequency components of the LR image as the known low-frequency components of the HR image. By using the ER algorithm which is introduced the above two constraints, an accurate HR image can be obtained.
View full abstract
-
Makoto TAKIZAWA, Miki HASEYAMA
Article type: Article
Session ID: HI2009-50/ME2009-50/
Published: February 04, 2009
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes an accurate restoration and resolution enhancement method of a sequence deteriorated by camera shake based on super-resolution reconstruction. The proposed method estimates the PSF (Point Spread Function) which denotes the deterioration function of a camera shake. Then, in order to calculate the correspondence between frames in the sequence accurately, we utilize restored frames as substitutes to the original ones. Furthermore, by introducing a new weighting factor, which prevents oversmoothing in the camera shake direction, to the super-resolution reconstructon, an accurate restoration and resolution enhancement can be realized. Simulation results are shown to verify the performance of the proposed method.
View full abstract
