-
Article type: Cover
Pages
Cover1-
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Article type: Index
Pages
Toc1-
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Xiaohua Zhang, Yuelan Xin, Ning Xie, Wenbo Jiang, Yue Zhao
Article type: Article
Session ID: ME2013-119
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Binarization is regarded as a compulsory step to separate text and background. When document images are acquired under poor illumination, the images are soiled with uneven shading. This paper presents an approach for removing the uneven shading from document image. Piecewise polynomials are adopted in computing the background surface with complicated shading to be removed. The document image is divided into several blocks overlapped with each other to reduce the discontinuity between polynomials. The fitting is again adopted in the normalized image to estimate the coefficients of piecewise polynomials and the foreground text pixels are excluded as could as possible. Our experiments show that the proposed method can remove complicated shading well.
View full abstract
-
Kenji HONDA, Makoto SATO
Article type: Article
Session ID: ME2013-120
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this article, we discuss the boundary detection by the zero-crossing. Let gradient and Hessian be considered as ∇f and H. Then, the zero-crossing representing the geometric boundary are the trH zero-crossing and the (∇f,H∇f) zero-crossing. In this article, especially, we discuss the (∇f,H∇f) zero-crossing. The (∇f,H∇f) zero-crossing means the inflection point of f(x, y) or the inflection point of the line which emanates from the maximum and reaches the minimum. In this article, we return to the geometric meaning of the (∇f,H∇f) zero-crossing and define the line which emanate from the maximum and reach the minimum. And we propose the technique for detecting the zero-crossing by analyzing the inflection point along the line and confirm if there is a difference in result between the traditional local difference operator and the proposed technique.
View full abstract
-
Isao MIYAGAWA, Hiroyuki ARAI, Yukinobu TANIGUCHI
Article type: Article
Session ID: ME2013-121
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We describe geometrical correspondences between projector pixels and camera pixels for uncalibrated projector-camera systems by introducing light transports, which have direct reflection, indirect reflection, and multipath reflection from unknown spatial structure. Since brute-force scan illuminates the objects using the light yielded from each projector pixel in order, the number of images captured by the camera grows exponentially with the number of pixels used for projector illumination in order to measure the full light transports. Accordingly, we encode projector pixels by coded illumination with multiple layers and emit coded patterns on each layer for unknown 3D objects. We obtain approximately light transports that have direct and indirect reflections by decoding the observed images. In experiments, the proposed method shows its effectiveness of acquiring light transports, and it yields excellent projection mapping on unknown 3D objects.
View full abstract
-
Masaki ICHIKAWA, Toshiyuki GOTOH, Seiichiro KAGEI, Tae IWASAWA
Article type: Article
Session ID: ME2013-122
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes a new pulmonary blood flow analysis with a two input model which evaluates aorta and pulmonary artery contributions. It is assumed that each response from inputs of pulmonary artery and recirculated aorta routes is assumed to be a gamma function, and that the sum of the responses is observed. Based on this model, a method to estimate the blood flow, blood volume and mean transit time is proposed. The experimental results using contrast enhanced MR images with lung tumor are shown, to evaluate the proposed method.
View full abstract
-
Nihad Karim Chowdhury, Atsushi Tatsuma, Masaki Aono
Article type: Article
Session ID: ME2013-123
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
We propose a new shape descriptor based on local and global features for 3D shape retrieval. Local features include "slope density" and "chain code", while global features include "distance histogram between a randomly chosen point and the shortest distance polygon from the point". We demonstrate our proposed shape descriptor and compare it with several previous methods by using Engineering Shape Benchmark (ESB). From our experiments, our method outperforms previous methods in terms of Precision@1 (P@1).
View full abstract
-
Ryuichi Kosaka, Atsushi Tatsuma, Masaki Aono
Article type: Article
Session ID: ME2013-124
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Research of 3D shape retrieval has been popular recently. However, only small amount of research has focused on partial 3D shape retrieval. In this paper, we propose new features KAZE+VLAD for partial 3D shape retrieval. In KAZE+VLAD, firstly, extracts the KAZE as local features from Depth Buffer images generated from multiple view points of the 3D Object. Secondly, integrates local features into features of 3D object by encoding method. We used VLAD for encoding method. We used a modified Princeton Segmentation Benchmark, and conducted comparative experiments between our proposed and previous features.
View full abstract
-
Takahiro OKUBO, Kenji HONDA, Katsuhito AKAHANE, Makoto SATO
Article type: Article
Session ID: ME2013-125
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this research, we propose a haptization system for animal images by using a 3D model of the animal. The haptization methods which were proposed before, there has been some problems, for example, insufficiencies of a sense of presence in some methods, requirement of the depth information in others, so they are not fit to apply to animal images. Therefore, the purpose of this research is to enhance the three-dimensional impression and the sense of presence by placing a 3D model of an animal on the image and performing the haptic feedback. But generating the 3D model of the animal from an image automatically and to estimate the pose is difficult. So we are going to make a system which have an existing 3D models and pose information. In this paper, we verify the effect of the haptization with the 3D model.
View full abstract
-
Yoshinobu TONOMURA
Article type: Article
Session ID: ME2013-126
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this talk, the concept of interactive ambience is introduced for the upcoming information environment in which we are surrounded by various kinds of information device. Our recent approaches towards the interactive ambience to find out its features, possibility, and technical issues are also covered.
View full abstract
-
Makoto SATO
Article type: Article
Session ID: ME2013-127
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Dongliang Qu, Kiyosi Nisikawa
Article type: Article
Session ID: ME2013-128
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this paper, we propose an method for improving the noise judging accuracy of NL-means noise rejection framework. The NL-means noise rejection framework was proposed as the technique of removing random noise effectively by Xiong Bo et al. The proposal method improves the characteristic of a NL-means noise rejection framework by detecting the noise using robust statistics and the directivity based on the regional dispersion of a pixel value.
View full abstract
-
Toshiyuki DOBASHI, Tatsuya MUROFUSHI, Masahiro IWAHASHI, Hitoshi KIYA
Article type: Article
Session ID: ME2013-129
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes a tone mapping operation (TMO) with a low memory cost and a low computational cost. A TMO generates a low dynamic range (LDR) image from a high dynamic range (HDR) image by compressing its dynamic range. Due to an HDR image is generally expressed in floating-point data, a TMO is also executed with floating-point arithmetic. The proposed method treats a floating-point number as two 8-bit integer numbers which correspond to a exponent part and a mantissa part, and applies tone mapping to these integer numbers separately. Moreover, the method conducts calculations in the tone mapping with fixed-point arithmetic. As a result, the method reduces a memory cost and a computational cost. The experimental and evaluation results show that the method has high accuracy of a TMO, though reducing memory and computational cost.
View full abstract
-
Shoi HAYAMA, Kiyoshi NISHIKAWA, Hitoshi KIYA
Article type: Article
Session ID: ME2013-130
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
A novel configuration for the non-local (NL) means algorithm is proposed for denoising the low dynamic range (LDR) images that are generated by tone-mapping high dynamic range (HDR) images. The proposed method improves the denoising performance by coordinately using the information of both the LDR and its corresponding HDR images. Through the results of computer simulation, it is shown that the proposed method could achieve better performance than the conventional NL-means algorithm.
View full abstract
-
Natsuki TAKAYAMA, Hiroki TAKAHASHI
Article type: Article
Session ID: ME2013-131
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
Choshi is a new carving method which can represent uneven 3D structures with unique colored papers. The color must be selected from a number of existing colors of papers. This paper describes halftone image representation for Choshi that uses circle cutting pattern and changes size and density of pattern under the restriction of creation.
View full abstract
-
Tokinobu MITASAKI, Naoki ONO, Atsushi SHIMIZU
Article type: Article
Session ID: ME2013-132
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
HEVC (High Efficiency Video Coding) is the video encoding standard with high efficiency that uses four CUs (Coding Unit) adaptively. But this adaptive selection of CUs greatly increases the processing complexity. In this report, we propose the fast CU decision method on HEVC. We calculate RD cost for two CU layers and decide the optimum CU size based on the relation between two RD costs. We are able to reduce CU the complexity by 40% with high efficiency and is independent of the video-bitrate.
View full abstract
-
Hideaki OKADA, Shinji FUKUMA, Shin-ichiro MORI
Article type: Article
Session ID: ME2013-133
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
This paper proposes an optimization method for lossless transform for lossless coding. Orthonormal transform can be realized as a lossless transform by doing lifting decomposition to a transformation matrix. It can used pivoting. But, compression ratio changes under the influence of the rounding error because lifting coefficient is different each pivoting. And it is necessary that we compare all compression ratio each picture in order to optimization. But, optimization is difficult because combination number is the factorial matrix size. Therefore, I propose a method to narrow down the candidates by deriving the rounding errors by using the additive noise model.
View full abstract
-
Masakazu KOGA, Yuji MARU, Kiyoshi NISHIKAWA
Article type: Article
Session ID: ME2013-134
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
A kernel adaptive algorithm enables to learn a nonlinear system by applying kernel method. However, increasing the number of training vectors while estimation processes causes a huge computational cost. This paper proposes a sparsification of proportionate-type of KNLMS, which updates filter coefficients with the individual step size, in the case that the size of the dictionary is limited.
View full abstract
-
Makoto TADENUMA, Toshiya MORITA
Article type: Article
Session ID: ME2013-135
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
The authors have developed systems that individually estimate the unpleasantness degree that is evoked when viewers watch images including one of shakiness, flicker or stripes. The systems, however, were not constructed to estimate the unpleasantness degree that is evoked when viewers watch images complexly including these unpleasantness factors. We derived a series of experiments in order to investigate the relationship between individual unpleasantness degree of images including each unpleasantness factor and total unpleasantness degree of images including plural unpleasantness factors, then, we found out the adequate nonlinear summation model to estimate the total unpleasantness degree.
View full abstract
-
Eiji WATANABE, Takashi OZEKI, Takeshi KOHAMA
Article type: Article
Session ID: ME2013-136
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
In this report, we consider a learning environment, which students solve the problem while reading contents. Here, we analyze the relationship among the behaviors (reading and writing), the frequency for reading each page and the understanding for contents. Concretely, we analyze the influence the frequency for the objective page and the behaviors for solving on the score for given problems.
View full abstract
-
Article type: Appendix
Pages
App1-
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Article type: Appendix
Pages
App2-
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS
-
Article type: Appendix
Pages
App3-
Published: December 02, 2013
Released on J-STAGE: September 21, 2017
CONFERENCE PROCEEDINGS
FREE ACCESS