• Journal of the Institute of Convergence Signal Processing
• /
• v.13 no.3
• /
• pp.130-135
• /
• 2012

In this paper, we propose a technique to detect the size and location of the small target in images by using Gaussian kernel repeatedly. In order to detect the size and location of the small target, we find the local maximum value by applying the facet model and then use the $3{\times}3$ Gaussian kernel repeatedly. we determine the size of small target by comparing the local maximum value $D_2$ according to the number of iteration. To reduce the computational complexity, we use the Gaussian pyramid when using the kernel repeatedly. Through the experiment, we verified that the size and location of the small target is detected by the number of iterations and results show improvements from conventional methods.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2015.10a
• /
• pp.1523-1526
• /
• 2015

본 논문에서는 커널 기반 객체 추적 방식을 이용하여 실시간으로 객체를 추적하는 기술에서 객체의 크기 추정에 대한 기존 연구를 비교 분석한다. 커널 기반 객체 추적은 추적하고자 하는 객체를 초기 프레임에서 타켓으로 설정한 후, 각 프레임마다 타켓 후보들을 제시하고 그 중 가장 객체와 유사한 후보에 초점을 맞춰가며 객체를 추적한다. 이때, 목표 객체와 객체 후보간 유사성을 기반으로 정의된 배경 영사 영상(back-projection image)을 이용하여 객체의 크기를 추정하는 방법들이 제시되고 있다. 안정적인 객체 크기 추정 방법의 설계를 위한 사전 연구로서 대표적인 객체 크기 추정 기존 연구를 비교, 분석하고자 한다.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.1
• /
• pp.111-120
• /
• 2016

In usual, the discontinuous variance function was estimated nonparametrically using a kernel type estimator with data sets split by an estimated location of the change point. Kang et al. (2000) proposed the Gasser-$M{\ddot{u}}ller$ type kernel estimator of the discontinuous regression function using the adjusted observations of response variable by the estimated jump size of the change point in $M{\ddot{u}}ller$ (1992). The adjusted observations might be a random sample coming from a continuous regression function. In this paper, we estimate the variance function using the Nadaraya-Watson kernel type estimator using the adjusted squared residuals by the estimated location of the change point in the discontinuous variance function like Kang et al. (2000) did. The rate of convergence of integrated squared error of the proposed variance estimator is derived and numerical work demonstrates the improved performance of the method over the exist one with simulated examples.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06a
• /
• pp.179-181
• /
• 2012

본 연구는 NAND 플래시 기반의 임베디드 시스템에 warm-start 개념을 도입하여 재부팅 시간을 단축시키는 방법에 관한 연구이다. NAND 플래시 기반의 임베디드 시스템에서는 부팅 시 커널 이미지를 NAND 플래시에서 DRAM으로 로드하는 과정이 필요하다. 최근 SoC (System-On-Chip)는 "software reset" 기능을 지원하며 DRAM의 내용을 훼손하지 않고 시스템을 리셋하는 것이 가능하다. 우리는 이 기능을 이용하여 DRAM 상에 이미 로드되어 있는 커널 요소를 재부팅 시에도 다시 사용하도록 구현하고, 그에 따라 중복된 커널 요소의 로드를 제거했다. 이 기법을 우리는 KWS(Kernel-Warm-Start) 라는 이름으로 정의했다. KWS 에서는 부팅이 재부팅인 경우 필요한 커널 요소만을 로드한다. 결과적으로 재부팅 시 커널 요소의 로드 크기를 89.29% 감소시킬 수 있었고, 재부팅 시간을 첫 번째 부팅에서 75.47% 단축할 수 있었다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.29 no.3
• /
• pp.143-152
• /
• 2002

As the multimedia application becomes ubiquitous, the size of message used for Inter Process Communication (IPC) grows up to cope with the requirements of multimedia applications. And the rapid development of new hardware platforms makes the portability of operating system more important. But the traditional micro-kernel operating system is Implemented platform dependently for better performance, and especially focused on handling short message. In this paper, we present the design and implementation of IPC mechanism in M3K (MultiMedia Micro-Kernel) to address the above problems. Our IPC mechanism provides enhanced performance and efficiently handles large message without performance degrading.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.40 no.6
• /
• pp.365-372
• /
• 2016

Using the computational fluid dynamics (CFD) technique, we simulated the fluid flow in a Taylor reactor considering the aggregation and breakage of particles. We calculated the population balance equation (PBE) to determine the particle-size distribution by implementing the quadrature method-of-moment (QMOM). It was used that six moments for an initial moments, the sum of Brownian kernel and turbulent kernel for aggregation kernel, and power-law kernel for breakage kernel. We predicted the final mean particle size when the particle had various initial volume fraction values. The result showed that the mean particle size and initial growth rate increased as the initial volume fraction of the particle increased.

• The Journal of the Korea Contents Association
• /
• v.11 no.11
• /
• pp.10-18
• /
• 2011

Medical asset protection is important in each medical institution especially because of the law on private medical record protection and face recognition for this protection is one of the most interesting and challenging problems. In recognizing human faces, the distortion of face images can be caused by the change of pose, illumination, expressions and scale. It is difficult to recognize faces due to the locations of lights and the directions of lights. In order to overcome those problems, this paper presents an analysis of coefficients of Gabor wavelets, kernel decision, feature point, size of kernel, for face recognition in CCTV surveillance. The proposed method consists of analyses. The first analysis is to select of the kernel from images, the second is an coefficient analysis for kernel sizes and the last is the measure of changes in garbo kernel sizes according to the change of image sizes. Face recognitions are processed using the coefficients of experiment results and success rate is 97.3%. Ultimately, this paper suggests empirical application to verify the adequacy and the validity with the proposed method. Accordingly, the satisfaction and the quality of services will be improved in the face recognition area.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.5C
• /
• pp.542-549
• /
• 2007

Recently, many research activities in the image processing area are concentrated on developing new algorithms by finding the solution of the 'diffusion equation'. The diffusion algorithms are expected to be utilized in numerous applications including noise removal and image restoration, edge detection, segmentation, etc. In this paper, at first, it will be shown that the anisotropic diffusion algorithms have the similar structure with the adaptive FIR filters with cross-shaped 5-tap kernel, and this relatively small-sized kernel causes many iterating procedure for satisfactory filtering effects. Moreover, it will also be shown that lots of modifications which are adopted to the conventional Gaussian diffusion method in order to weaken the edge blurring nature of the linear filtering process increases another computational burden. We propose a new Median diffusion scheme by replacing the adaptive linear filters in the diffusion process with the AWM (Adaptive Weighted Median) filters. A diffusion-equation-based adaptation scheme is also proposed. With the proposed scheme, the size of the diffusion kernel can be increased, and thus diffusion speed greatly increases. Simulation results shows that the proposed Median diffusion scheme outperforms in noise removal (especially impulsive noise), and edge preservation.

• Journal of KIISE:Computing Practices and Letters
• /
• v.6 no.2
• /
• pp.173-186
• /
• 2000

Real-time embedded systems should accommodate many kinds of hardware platforms and resource management policies that vary depending on their operating environments and purposes. It is not an easy job to adapt a multitasking kernel to new system services and hardware platforms, as the kernel must strictly satisfy constraints on its size and performance. This paper describes the design and implementation of an object-oriented multitasking framework that can be reused for implementing microprocessor-based real-time embedded multitasking kernels, In this kernel framework, those parts that can vary depending on hardware platforms and system resource management policies are separated into the hot spots and encapsulated by abstract classes. Our framework thus can be effectively used to implement microprocessor-based real-time embedded kernels that demand high portability and adaptability.

• The KIPS Transactions:PartB
• /
• v.10B no.3
• /
• pp.243-248
• /
• 2003

An incremental kernel principal component analysis (IKPCA) is proposed for the nonlinear feature extraction from the data. The problem of batch kernel principal component analysis (KPCA) is that the computation becomes prohibitive when the data set is large. Another problem is that, in order to update the eigenvectors with another data, the whole eigenspace should be recomputed. IKPCA overcomes these problems by incrementally computing eigenspace model and empirical kernel map The IKPCA is more efficient in memory requirement than a batch KPCA and can be easily improved by re-learning the data. In our experiments we show that IKPCA is comparable in performance to a batch KPCA for the feature extraction and classification problem on nonlinear data set.