• MALSORI
• /
• no.48
• /
• pp.93-105
• /
• 2003

Since telephone channel has bandlimited frequency characteristics, speech signal over the telephone channel shows degraded speech quality. In this paper, we propose an algorithm using neural network to reconstruct wideband speech from its narrowband version. Although single neural network is a good tool for direct mapping, it has difficulty in training for vast and complicated data. To alleviate this problem, we modularize the neural networks based on appropriate clustering of the acoustic space. We also introduce fuzzy computing to compensate for probable misclassification at the cluster boundaries. According to our simulation, the proposed algorithm showed improved performance over the single neural network and conventional codebook mapping method in both objective and subjective evaluations.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.11
• /
• pp.2251-2259
• /
• 2009

We propose an algorithm and system which generates 3D stereo image by composition of 2D image from 4 multiple clusters which 1 cluster was composed of 4 multiple cameras based on network. Proposed Schemes have a network-based client-server architecture for load balancing of system caused to process a large amounts of data with real-time as well as multiple cluster environments. In addition, we make use of JPEG compression and RAM disk method for better performance. Our scheme first converts input images from 4 channel, 16 cameras to binary image. And then we generate 3D stereo images after applying edge detection algorithm such as Sobel algorithm and Prewiit algorithm used to get disparities from images of 16 multiple cameras. With respect of performance results, the proposed scheme takes about 0.05 sec. to transfer image from client to server as well as 0.84 to generate 3D stereo images after composing 2D images from 16 multiple cameras. We finally confirm that our scheme is efficient to generate 3D stereo images in multiple view and multiple clusters environments with real-time.

• Journal of KIISE:Software and Applications
• /
• v.27 no.4
• /
• pp.422-431
• /
• 2000

This paper proposes a new codebook generation method, called a PCA-Based VQ, that incorporates the PCA (Principal Component Analysis) technique into VQ (Vector Quantization) codebook design. The PCA technique reduces the data dimensions by transforming input image vectors into the feature vectors. The cluster of feature vectors in the transformed domain is bisectioned into two subclusters by an optimally chosen partitioning hyperplane. We expedite the searching of the optimal partitioning hyperplane that is the most time consuming process by considering that (1) the optimal partitioning hyperplane is perpendicular to the first principal axis of the feature vectors, (2) it is located on the equilibrium point of the left and right cluster's distortions, and (3) the left and right cluster's distortions can be adjusted incrementally. This principal axis bisectioning is successively performed on the cluster whose difference of distortion between before and after bisection is the maximum among the existing clusters until the total distortion of clusters becomes as small as the desired level. Simulation results show that the proposed PCA-based VQ method is promising because its reconstruction performance is as good as that of the SOFM (Self-Organizing Feature Maps) method and its codebook generation is as fast as that of the K-means method.

• Journal of KIISE:Computing Practices and Letters
• /
• v.10 no.1
• /
• pp.25-36
• /
• 2004

This paper proposes and measures the performance of a distributed EIT (Electrical Impedance Tomography) image reconstruction algorithm which has a master-slave structure. The image construction is a computation based application of which the execute time is proportional to the cube of the unknowns. After receiving a specific frame from the master, each computing node extracts the basic elements by executing the first iteration of Kalman Filter in parallel. Then the master merges the basic element lists into one group and then performs the sequential iterations with the reduced number of unknowns. Every computing node has MATLAB functions as well as extended library implemented for the exchange of MATLAB data structure. The master implements another libraries such as threaded multiplication, partitioned inverse, and fast Jacobian to improve the speed of the serial execution part. The parallel library reduces the reconstruction time of image visualization about by half, while the distributed grouping scheme further reduces by about 12 times for the given target object when there are 4 computing nodes.

• Journal of Information Processing Systems
• /
• v.6 no.1
• /
• pp.67-78
• /
• 2010

In this study, a clustering algorithm that uses DCT transformed data is presented. The algorithm is a grid density-based clustering algorithm that can identify clusters of arbitrary shape. Streaming data are transformed and reconstructed as needed for clustering. Experimental results show that DCT is able to approximate a data distribution efficiently using only a small number of coefficients and preserve the clusters well. The grid based clustering algorithm works well with DCT transformed data, demonstrating the viability of DCT for data stream clustering applications.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.2
• /
• pp.78-87
• /
• 2006

In this paper. we propose an effective mask estimation scheme for missing-feature reconstruction in order to achieve robust speech recognition under unknown noise environments. In the previous work. colored noise is used for training the mask classifer, which is generated from the entire frequency Partitioned signals. However it gives a limited performance under the restricted number of training database. To reflect the spectral events of more various background noise and improve the performance simultaneously. a new Bayesian classifier for mask estimation is proposed, which works independent of other frequency bands. In the proposed method, we employ the colored noise which is obtained by combining colored noises generated from each frequency band in order to reflect more various noise environments and mitigate the 'sparse' database problem. Combined with the cluster-based missing-feature reconstruction. the performance of the proposed method is evaluated on a task of noisy speech recognition. The results show that the proposed method has improved performance compared to the Previous method under white noise. car noise and background music conditions.

• Journal of the Institute of Convergence Signal Processing
• /
• v.23 no.1
• /
• pp.15-20
• /
• 2022

An improvement of computational resources with a large scale cluster service is available to the individual person, which has been limited to the original industry and research institute. Therefore, the application of powerful computational resources to the engineering design has been increased fast. In naval and marine industry, the application of Computational Fluid Dynamics, which requires a huge computational effort, to a design of ship and offshore structure has been increased. Floating bodies such as the ship or offshore structure is exposed to ocean waves, current and wind in the ocean, therefore the precise modelling of those environmental disturbances is important in Computational Fluid Dynamics. Especially, ocean waves has to be nonlinear rather than the linear model based on the superposition due to a nonlinear characteristics of Computational Fluid Dynamics. In the present study, a fast reconstruction technique is suggested and it is validated from a series of simulations by using the Computational Fluid Dynamics.

• Journal of the Korea Computer Graphics Society
• /
• v.23 no.4
• /
• pp.31-40
• /
• 2017

As 3D cell culture has recently become possible, it has been able to observe a 3D shape of cell and volume. Generally, 3D information of a cell should be observed with a special microscope such as a confocal microscope or an electron microscope. However, a confocal microscope is more expensive than a conventional microscope and takes longer time to capture images. Therefore, there is a need for a method that can reconstruct the 3D shape of cells using a common microscope. In this paper, we propose a method of reconstructing 3D cells using the focus estimator value from multi-focal fluorescence images of cells. Initially, 3D cultured cells are captured with an optical microscope by changing the focus. Then the approximate position of the cells is assigned as ROI (Region Of Interest) using the circular Hough transform in the images. The MSBF (Modified Sliding Band Filter) is applied to the obtained ROI to extract the outlines of the cell clusters, and the focus estimator values are computed based on the extracted outlines. Using the computed focus estimator values and the numerical aperture (NA) of the microscope, we extract the outline of the cell cluster considering the depth and reconstruct the cells into 3D based on the extracted outline. The reconstruction results are examined by comparing with the combined in-focus portions of the cell images.

• Korean Journal of Veterinary Research
• /
• v.53 no.3
• /
• pp.155-158
• /
• 2013

Paedomorphosis is the retention of juvenile morphology at maturity and is important in generating evolutionary change in domestic species and species in the wild. This preliminary study compared morphological characteristics of seven domestic cat pure breeds and felid species from around the world. An original procedure based on elliptical Fourier (EF) methods was applied to head lateral views of specimens and were assessed in order to analyse head profile. For comparative purposes three domestic kittens of different ages and three species of genus Lynx were also used. EF descriptors, allowed for the quantification of the head profile. Using the Fourier transform, reconstruction of the mean head profile revealed that there was a general shape difference between wild cats, domestic cats and kittens. Results suggested that variability in head profile differentiate quite well between adult cats and kittens, but domestic and wild cats appeared grouped into a similar cluster. The similarity between breeds can thus be attributed more to the general head profile than to flatness, i.e. to the general conformation rather than facial profile. Therefore, no effect of paedomorphism on the studied breeds can be undertaken. The present approach opens interesting ethnological perspectives for the aloidic characterisation for domestic breeds.

• Journal of the Korean Association of Oral and Maxillofacial Surgeons
• /
• v.43 no.1
• /
• pp.29-36
• /
• 2017

Objectives: Acquired immunodeficiency syndrome (AIDS) is a disease of the human immune system caused by the human immunodeficiency virus (HIV). People with AIDS are much more vulnerable to infections, including opportunistic infections and tumors, than people with a healthy immune system. The objective of this study was to correlate oral lesions associated with HIV/AIDS and immunosuppression levels by measuring clusters of differentiation 4 (CD4) cell counts among patients living in the middle western regions of Ghana. Materials and Methods: A total of 120 patients who visited the HIV clinic at the Komfo Anokye Teaching Hospital and the Regional Hospital Sunyani of Ghana were consecutively enrolled in this prospective and cross-sectional study. Referred patients' baseline CD4 counts were obtained from medical records and each patient received an initial physician assessment. Intraoral diagnoses were based on the classification and diagnostic criteria of the EEC Clearinghouse, 1993. After the initial assessment, extra- and intraoral tissues from each enrolled patient were examined. Data analyses were carried out using simple proportions, frequencies and chi-square tests of significance. Results: Our study included 120 patients, and was comprised of 42 (35.0%) males and 78 (65.0%) females, ranging in age from 21 to 67 years with sex-specific mean ages of 39.31 years (males) and 39.28 years (females). Patient CD4 count values ranged from 3 to 985 cells/mL with a mean baseline CD4 count of 291.29 cells/mL for males and 325.92 cells/mL for females. The mean baseline CD4 count for the entire sample was 313.80 cells/mL. Of the 120 patients we examined, 99 (82.5%) were observed to have at least one HIV-associated intraoral lesion while 21 (17.5%) had no intraoral lesions. Oral candidiasis, periodontitis, melanotic hyperpigmentation, gingivitis and xerostomia were the most common oral lesions. Conclusion: From a total of nine oral lesions, six lesions that included oral candidiasis, periodontitis, melanotic hyperpigmentation, gingivitis, xerostomia and oral hairy leukoplakia were significantly correlated with declining CD4 counts.