• Progress in Medical Physics
• /
• v.9 no.1
• /
• pp.47-53
• /
• 1998

Nuclear Medicine Images have comparatively poor spatial resolution, making it difficult to relate the functional information which they contain to precise anatomical structures. Anatomical structures useful in the interpretation of SPECT /PET Images were radiolabelled. PET/SPECT Images Provide functional information, whereas MRI mainly demonstrate morphology and anatomical. Fusion or Image Registration improves the information obtained by correlating images from various modalities. Brain Scan were studied on one or more occations using MRI and SPECT. The data were aligned using a point pair methods and surface matching. SPECT and MR Images was tested using a three dimensional water fillable Hoffman Brain Phantom with small marker and PET and MR Image was tested using a patient data. Registration of SPECT and MR Images is feasible and allows more accurate anatomic assessment of sites of abnormal uptake in radiolabeled studies. Point based registration was accurate and easily implemented three dimensional registration of multimodality data set for fusion of clinical anatomic and functional imaging modalities. Accuracy of a surface matching algorithm and homologous feature pair matching for three dimensional image registration of Single Photon Emission Computed Tomography Emission Computed Tomography (SPECT), Positron Emission Tomography (PET) and Magnetic Resonance Images(MRD was tested using a three dimensional water fill able brain phantom and Patients data. Transformation parameter for translation and scaling were determined by homologous feature point pair to match each SPECT and PET scan with MR images.

• Proceedings of the Korean Society of Computational and Applied Mathematics Conference
• /
• 2003.09a
• /
• pp.11-11
• /
• 2003

오늘날 단일 슈퍼컴퓨터로는 처리가 불가능한 거대한 문제들의 해법이 시도되고 있는데, 이들은 지리적으로 분산된 슈퍼컴퓨터, 데이터베이스, 과학장비 및 디스플레이 장치 등을 초고속 통신망으로 연결한 GRID 환경에서 효과적으로 실행시킬 수 있다. GRID는 1990년대 중반 과학 및 공학용 분산 컴퓨팅의 연구 과정에서 등장한 것으로, 점차 응용분야가 넓어지고 있다. 그러나 GRID 같은 분산 환경은 기존의 단일 병렬 시스템과는 많은 점에서 다르며 이전의 기술들을 그대로 적용하기에는 무리가 있다. 기존 병렬 시스템에서는 주로 동기 알고리즘(synchronous algorithm)이 사용되는데, 직렬 연산과 같은 결과를 얻기 위해 동기화(synchronization)가 필요하며, 부하 균형이 필수적이다. 그러나 부하 균형은 이질 클러스터(heterogeneous cluster)처럼 프로세서들의 성능이 서로 다르거나, 지리적으로 분산된 계산자원을 사용하는 GRID 환경에서는 이기종의 문제뿐 아니라 네트워크를 통한 메시지의 전송 지연 등으로 유휴시간이 길어질 수밖에 없다. 이처럼 동기화의 필요성에 의한 연산의 지연을 해결하는 하나의 방안으로 비동기 반복법(asynchronous iteration)이 나왔으며, 지금도 활발히 연구되고 있다. 이는 알고리즘의 동기점을 가능한 한 제거함으로써 빠른 프로세서의 유휴 시간을 줄이는 것이 목적이다. 즉 비동기 알고리즘에서는, 각 프로세서는 다른 프로세서로부터 갱신된 데이터가 올 때까지 기다리지 않고 계속 다음 작업을 수행해 나간다. 따라서 동시에 갱신된 데이터를 교환한 후 다음 단계로 진행하는 동기 알고리즘에 비해, 미처 갱신되지 않은 데이터를 사용하는 경우가 많으므로 전체적으로는 연산량 대비의 수렴 속도는 느릴 수 있다 그러나 각 프로세서는 거의 유휴 시간이 없이 연산을 수행하므로 wall clock time은 동기 알고리즘보다 적게 걸리며, 때로는 50%까지 빠른 결과도 보고되고 있다 그러나 현재까지의 연구는 모두 어떤 수렴조건을 만족하는 선형 시스템의 해법에 국한되어 있으며 비교적 구현하기 쉬운 공유 메모리 시스템에서의 연구만 보고되어 있다. 본 연구에서는 행렬의 주요 고유쌍을 구하는 데 있어 비동기 반복법의 적용 가능성을 타진하기 위해 우선 이론적으로 단순한 멱승법을 사용하여 실험하였고 그 결과 순수한 비동기 반복법은 수렴하기 어렵다는 결론을 얻었다 그리하여 동기 알고리즘에 비동기적 요소를 추가한 혼합 병렬 알고리즘을 제안하고, MPI(Message Passing Interface)를 사용하여 수원대학교의 Hydra cluster에서 구현하였다. 그 결과 특정 노드의 성능이 다른 것에 비해 현저하게 떨어질 때 전체적인 알고리즘의 수렴 속도가 떨어지는 것을 상당히 완화할 수 있음이 밝혀졌다.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.39 no.5
• /
• pp.1-7
• /
• 2002

In this paper, an extended type of a direct learning control(DLC) method is proposed for the effective control of systems which perform a given task repetitively. DLC methods have been suggested to overcome the defects of iterative learning control, the learning process should be resumed from the beginning even if a slight change occurs in the desired output pattern. If a given desired output trajectory is "proportional" to the output trajectories which are learned previously, we can obtain the desired control input directly without the iterative learning process by using the DLC. First, most existing DLC methods are shown to be applicable only to single-input single-output systems with the relative degree one and then, an extended type of DLC is proposed for a class of nonlinear systems having the relative degree more than or equal to one by using the known relative degree of a nonlinear system. By the simulation results for the arbitrary nonlinear system with the relative degree more than one, the validity and the performance of the proposed DLC method are examined.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.23 no.8
• /
• pp.1048-1062
• /
• 1999

A numerical method and experiments for the aerodynamic design of high performance two-stage axial flow fans was carried out. A vortex ring element method used for the aerodynamic analysis of the propellers was extended to the fan-duct system. Fan Performance and velocity profiles at the fan inlet and outlet are compared with experimental data for the validations of numerical method. Performance test was done based on KS B 6311(testing methods for turbo-fans and blowers). The velocity profile was obtained using a 5-hole pitot tube by the non-nulling method. The two stage axial flow fan configurations for the optimal operation conditions were set by using the experimental results for the single rotating axial flow fan and the single stage axial flow fan. The single rotating axial flow fan showed relatively low efficiency due to the swirl velocities behind rotor exit which produced pressure losses. In contrast, the single stage and the two-stage axial flow fans showed performance improvements due to the swirl velocity reduction by the stator. The peak efficiency of the two stage axial flow fan was improved by 21% and 6%, compared to the single rotating axial flow fan and the single stage axial flow fan, respectively.

• Journal of Broadcast Engineering
• /
• v.25 no.2
• /
• pp.242-251
• /
• 2020

Recently, many deep convolutional neural networks for image super-resolution have been studied. Existing deep learning-based super-resolution algorithms are architecture that up-samples the resolution at the end of the network. The post-upsampling architecture has an inefficient structure at large scaling factor result of predicting a lot of information for mapping from low-resolution to high-resolution at once. In this paper, we propose a single image super-resolution using Channel Attention Residual Dense Block based on an iterative up-down sampling architecture. The proposed algorithm efficiently predicts the mapping relationship between low-resolution and high-resolution, and shows up to 0.14dB performance improvement and enhanced subjective image quality compared to the existing algorithm at large scaling factor result.

• Journal of Korean Society of Steel Construction
• /
• v.9 no.2 s.31
• /
• pp.193-204
• /
• 1997

In order to quantify the relationships of the important physical factors relating failure to strong earthquake loading, the plastic fatigue problems for structural components under repeated loading were reviewed first. A new concept of very low cycle fatigue failure for structural components under severe cyclic excitations as in strong earthquakes was represented. Also, an experimental study was made of the very low cycle fatigue failure of structural steel elements. It was attempted to realize the ultimate failure in the course of loading repetitions of the order of several to twenty. The test specimen had a form of rectangular plate, representing a thin-plated element in a steel member as wide-flange cross section. It was subjected to uniaxial loading repeatedly, until complete failure takes place after undergoing inelastic buckling, plastic elongation and/or their combination. It was seen as a result that the state of the ultimate failure is closely related to the maximum strain at the extreme fiber in the cross section.

• Journal of the Korean Society of Propulsion Engineers
• /
• v.19 no.1
• /
• pp.9-17
• /
• 2015

The basic flow configuration is composed of a plane, double shear layer where relatively thin mid gas layer is sandwiched between air and fuel stream. The present study describes numerical investigations concerning the combustion enhancement according to a variation of mid layer thickness. In this case, the effect of heat release in turbulent mixing layers is important. For the numerical solution, a fully conservative unsteady $2^{nd}$ order time accurate sub-iteration method and $2^{nd}$ order TVD scheme are used with the finite volume method including k-${\omega}$ SST model. The results consists of three categories; single shear layer consists of fuel and air, inert gas sandwiched between fuel and air, cold fuel gas sandwiched between fuel and air. The numerical calculations has been carried out in case of 1, 2, 4 mm of mid layer thickness. The height of total gas stream is 4 cm. The combustion region is broadened in case of inert gas layer of 2, 4 mm thickness and cold fuel layer of 4 mm thickness compared with single shear layer.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38C no.9
• /
• pp.746-756
• /
• 2013

The emitter geolocation method using the time difference of arrival (TDOA) and the frequency difference of arrival (FDOA) has more accurate performance comparing to the single TDOA or FDOA based method. The estimation performance varies with the sensor paring strategies, the deployment and velocities of the sensors. Therefore, to establish effective strategy on the electronic warfare system, it is required to analyze the relation between the estimation accuracy and the operational condition of sensors. However, in the conventional non-iterative method, the restriction of the deployment of sensors and the reference sensor exists. Therefore, we derive the emitter geolocation method based on a Gauss-Newton method which is available to apply to any various sensor pairs and the deployment and velocities of the sensors. In addition, simulation results are included to compare the performance of geolocation method according to the used measurements: the combined TDOA/FDOA, TDOA, and FDOA. Also, we present that the combined TDOA/FDOA method outperforms over single TDOA or FDOA on the estimation accuracy with the CEP plane.

• Nuclear Medicine and Molecular Imaging
• /
• v.43 no.5
• /
• pp.459-467
• /
• 2009

Purpose: The maximum likelihood-expectation maximization (ML-EM) is the statistical reconstruction algorithm derived from probabilistic model of the emission and detection processes. Although the ML-EM has many advantages in accuracy and utility, the use of the ML-EM is limited due to the computational burden of iterating processing on a CPU (central processing unit). In this study, we developed a parallel computing technique on GPU (graphic processing unit) for ML-EM algorithm. Materials and Methods: Using Geforce 9800 GTX+ graphic card and CUDA (compute unified device architecture) the projection and backprojection in ML-EM algorithm were parallelized by NVIDIA's technology. The time delay on computations for projection, errors between measured and estimated data and backprojection in an iteration were measured. Total time included the latency in data transmission between RAM and GPU memory. Results: The total computation time of the CPU- and GPU-based ML-EM with 32 iterations were 3.83 and 0.26 see, respectively. In this case, the computing speed was improved about 15 times on GPU. When the number of iterations increased into 1024, the CPU- and GPU-based computing took totally 18 min and 8 see, respectively. The improvement was about 135 times and was caused by delay on CPU-based computing after certain iterations. On the other hand, the GPU-based computation provided very small variation on time delay per iteration due to use of shared memory. Conclusion: The GPU-based parallel computation for ML-EM improved significantly the computing speed and stability. The developed GPU-based ML-EM algorithm could be easily modified for some other imaging geometries.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.12 no.3
• /
• pp.371-383
• /
• 1999

An efficient and numerically stable eigensolution method for structures with multiple natural frequencies is presented. The proposed method is developed by improving the well-known subspace iteration method with shift. A major difficulty of the subspace iteration method with shift is that because of singularity problem, a shift close to an eigenvalue can not be used, resulting in slower convergence. In this paper, the above singularity problem has been solved by introducing side conditions without sacrifice of convergence. The proposed method is always nonsingular even if a shift is on a distinct eigenvalue or multiple ones. This is one of the significant characteristics of the proposed method. The nonsingularity is proved analytically. The convergence of the proposed method is at least equal to that of the subspace iteration method with shift, and the operation counts of above two methods are almost the same when a large number of eigenpairs are required. To show the effectiveness of the proposed method, two numerical examples are considered.