All issues
Predecessor
32.54
Displaying 1-20 of 20 articles from this issue
- |<
- <
- 1
- >
- >|
-
Article type: Cover
Pages Cover1-
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (15K) -
Article type: Index
Pages Toc1-
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (87K) -
Tomonari MASADA, Tsuyoshi HAMADA, Yuichiro SHIBATA, Kiyoshi OGURIArticle type: Article
Session ID: ME2008-174/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn this paper, we propose a GPU acceleration of multi-topic extraction from images by using LDA (latent Dirichlet allocation). LDA is originally proposed as a probabilistic model for documents by Blei et al. In recent days, LDA is applied to multimedia information other than documents. We provide the results of experiments where we apply LDA to Professor Wang's 10,000 test images and extract multiple visional topics. We adpot collapsed variational Bayesian inference method for LDA and accelerate this by using Nvidia CUDA compatible GPU devices.View full abstractDownload PDF (1197K) -
Makoto FUJIMURA, Shinichi MIYATA, Hideo KURODA, Hiroki IMAMURA, Kazuma ...Article type: Article
Session ID: ME2008-175/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSA watermarking technique is a technique to embed copyright information in digital contents. In our study, the watermarking technique is used for compression of images, namely a part of the image data is embedded into own image in the JPEG compression process. In the JPEG compression process, DC terms and AC terms are processed in different way, and are encoded respectively. In the proposed method, at the coder side, the data of DC terms are embedded into AC terms and only the data of AC term transmitted. And at the decoder side, the data of DC terms are extracted firstly from AC terms, then the information of DC terms are reconstructed. Therefore, the proposed method delete the data of DC terms and achieved about 2.48% improvement of compression ratio compared with JPEG method. In this paper, proposed method and experimental results are shown.View full abstractDownload PDF (1123K) -
Masanori Kojima, Kunihiro YamadaArticle type: Article
Session ID: ME2008-176/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSA Super resolution image reconstruction method using logical filter is discussed. Logical filter needs small operation time and small hardware size in integrated-circuit. Then, this method is useful for real time signal processing applications.View full abstractDownload PDF (281K) -
Tsuyoshi Hamada, Tomonari Masada, Yuichiro Shibata, Kiyoshi OguriArticle type: Article
Session ID: ME2008-177/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (715K) -
Hiroaki KATSUMA, Toshihiro NISHIMURAArticle type: Article
Session ID: ME2008-178/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSPreviously, PSNR is widely used as an evaluation method of image quality. PSNR only digitizes differences between two images, so sometimes an acquired quantity does not agree with subjective evaluation and factors of the value are difficult to identify. In this paper, we have discussed the possibility to apply the luminance gradient to the assessment of the image quality in order to develop a new image quality assessment method. For this purpose, we investigate the relationship between luminance gradient and the factors of image quality such as noise and sharpness. As a result, though the possibility to apply the luminance gradient to the index of noise and sharpness was indicated, we conclude that how to combine each parameters is important in order to construct the new index of the image quality.View full abstractDownload PDF (640K) -
Sayaka TAKARA, Nobuhiko MUKAI, Makoto KOSUGIArticle type: Article
Session ID: ME2008-179/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSThis paper describes Ikebana training system by using virtual reality technology. The flower model is constructed with computer graphics modeling software, and the tactile sense that we can obtain at the time during sticking the flower into a pinholder called kenzan, can be realized by a haptic device. The important thing of Ikebana training is the tactile sense so that we sorted the flowers into two types of branch and grass, and measured the solidity by using a durometer. The measured data are used to display the tactile sense. This system also can perform collision detection not only between a flower and the pinholder, but also between flowers by modeling flowers constructed with some parts such as petal, branch, leave and so on. Finally, we could train for Ikebana skill without real flowers by using this system.View full abstractDownload PDF (1613K) -
Deng ZHANG, Jegoon RYU, Toshihiro NishimuraArticle type: Article
Session ID: ME2008-180/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSPrevious papers about Random Telegraph Signal (RTS) noise mainly focus on its characteristics in time domain or its Power Spectra Density, which means almost all the researchers devote to the behavior of a RTS noise in a single pixel. Little attention has been paid to how RTS noise distributes on a CMOS Image Sensor Array (CIS) and how its distribution influences the quality of videos produced by that CIS. In this paper, two main works are introduced. One is that the exposure process of a camera is simulated, in which Gaussian noise and RTS noise on pinned-photodiode CMOS pixels are modeled on spatial domain; the other is that a new video quality evaluation method for RTS noise is proposed. Conclusions on how the spatial distribution of RTS noise affects the quality of motion picture will be drawn finally.View full abstractDownload PDF (355K) -
Yasuhiro Kato, Hidenori Maruta, Fujio KurokawaArticle type: Article
Session ID: ME2008-181/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn Vast facilities like chemical plants and thermal power plants, it is important to detect a fire early and certainly. The purpose of this paper is to present a new smoke detection method, in which the smoke is considered as a significant signal of the fire. At first, moving objects are detected from image sequences, and then the noise is removed with the image binarization and morphological operations. Furthermore, since the smoke pattern must be examined, the smoke feature is extracted with the texture analysis. Lastly, to obtain the final result of the proposed method, we discussed the properties of the proposed features as the time series data.View full abstractDownload PDF (1324K) -
Takao JINNO, Masahiro OKUDAArticle type: Article
Session ID: ME2008-182/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSWe propose a method for pseudo high dynamic range image acquisition. Usually the pseudo high dynamic range image is created by fusing multiple images with different exposures, while our method uses only one photograph. By using a low exposure image and effective noise reduction, we successfully create a high contrast image with few over- and underexposures.View full abstractDownload PDF (1123K) -
Eiji WATANABE, Katsumi MORIArticle type: Article
Session ID: ME2008-183/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSImage segmentation plays an important role in image understanding. It is necessary for such a method to extract each region consisting of images with enough accuracy. This report proposes an image segmentation method using multiple multi-layered neural Networks (NNs). As well-known, we can obtain compressed represen-tations in the hidden layer, when input patterns are identical with output patterns. However, when images have complex features, single neural network model can not compress them with enough accuracy. In this report, multiple NNs multi-layered neural networks is newly introduced and an image segmentation method is proposed based on the competitive learning among the compression abilities in multiple NNs. Finally, the segmentation results by the proposed method are concretely shown by some ex periments.View full abstractDownload PDF (1438K) -
Sayaka OKAMOTO, Hiroshi HANAIZUMI, Rica HAGIWARAArticle type: Article
Session ID: ME2008-184/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSA multi-camera system was proposed for tracking persons using PC controlled cameras connected with network. In the system, observed area was freely expanded by connecting PC controlled cameras whose FOVs were overlapped with those of the neighboring. The overlapping area was used for relaying the person information among neighboring cameras. The person information was collected by using a server-client protocol. Each client PC extracts walking persons by using Background Difference. Validity of the proposed system was confirmed by numerical simulation.View full abstractDownload PDF (1044K) -
Hiroshi INABA, Sei-ichiro KAMATAArticle type: Article
Session ID: ME2008-185/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn this paper, we propose a method for extracting raindrops based on analysis on motion vectors. The proposed method is applied to extract raindrops on a water repellent windshield from in-vehicle camera images. As a result of experiments, apporoximately 40 percent of raindrops in an image are extracted by our method.View full abstractDownload PDF (1256K) -
Lossless Image Compression Using Predictor Selection Based on Difference to Average Prediction ValueKenjiro SUGIMOTO, Yoshimitsu KUROKI, Sei-ichiro KAMATAArticle type: Article
Session ID: ME2008-186/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSIn predictive coding, prediction methods that aim to more efficient prediction by switching some predictors based on local index have been proposed in the past. Difference between adjacent pixels, i.e. gradients, is popularly used in such methods. However, relationship between the gradients and predictors is discussed only qualitatively. In this paper, we propose a prediction method using difference to average prediction value as local index. In the experiments, the proposed prediction scheme shows its superiority to LOCO-I and CALIC in terms entropy of prediction error and code length.View full abstractDownload PDF (596K) -
Ken-ichi TANAKA, Yasuhito FUJITAArticle type: Article
Session ID: ME2008-187/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSA computer-generated pseudo hologram was synthesized using the complex Hadamard transform instead of the Fourier transform used to synthesize a computer-generated hologram. As a result, an image was reconsurcted through ae simulation, and a result that was similar result that from to the Fourier transform was obtained for the binary phase quantization, although a physical image (optical reconstructed image) was obtained. In addition, we propose the application of computer-generated pseudo holograms, which are used to hide the confidential information in a dither image. The strength of a StirmMark attack on a dithered image embedded with computer-generated pseudo holograms is weaker than the results of that using the Fourier transform hologram. It was previously reported that there was the possibility of the use of the computer generated holograms for information security.View full abstractDownload PDF (1327K) -
Yosuke TAKADA, Kiichi URAHAMAArticle type: Article
Session ID: ME2008-188/BCT2008-1
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSSome fast approximate methods have been developed for bilateral niters. One of them decomposes filter coefficients spatially and another one reduces computational costs by downsampling of voxel data in coordinate-graylevel product spaces. In this paper, we combine these two techniques to reduce computational times of bilateral filtering by spatially decomposing filter coefficients in linear Gaussian filters in the coordinate-graylevel product space and reducing the voxel sites by their downsampling. Though our method is slightly slower than the previous method of directly decomposing filter coefficients in image planes, approximation errors are smaller in our method than the previous technique.View full abstractDownload PDF (772K) -
Article type: Appendix
Pages App1-
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (79K) -
Article type: Appendix
Pages App2-
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (79K) -
Article type: Appendix
Pages App3-
Published: November 28, 2008
Released on J-STAGE: September 20, 2017
CONFERENCE PROCEEDINGS FREE ACCESSDownload PDF (79K)
- |<
- <
- 1
- >
- >|