• 대한의용생체공학회:학술대회논문집
• /
• 대한의용생체공학회 1994년도 추계학술대회
• /
• pp.39-42
• /
• 1994

An automatic multi-threshold algorithm for segmentation of 2D ultrasound images based on average filtering and the characteristics of speckle noise in 2D ultrasound image is proposed. To do this, we investigate the histogram of difference between $7{\times}7$ averaging histogram and $3{\times}3$ averaging histogram. And, we find zero crossing points in the positive portion of the differenced histogram and select middle points of the zero crossing points. We assign these selected points to characteristic points. The thresholds are the center of two characteristic points. Then we segment 2D ultrasound image by using these thresholds and extract edges from applying edge operator to optimal segmented image. Experimental results show that the segmented regions are devided accurately around the homogeneous region.

• Industrial Engineering and Management Systems
• /
• 제10권4호
• /
• pp.272-278
• /
• 2011

Decision tree as a classification tool is being used successfully in many areas such as medical diagnosis, customer churn prediction, signal detection and so on. The main advantage of decision tree classifiers is their capability to break down a complex structure into a collection of simpler structures, thus providing a solution that is easy to interpret. Since decision tree is a top-down algorithm using a divide and conquer induction process, there is a risk of reaching a local optimal solution. This paper proposes a procedure of optimally determining thresholds of the chosen variables for a decision tree using an adaptive particle swarm optimization (APSO). The proposed algorithm consists of two phases. First, we construct a decision tree and choose the relevant variables. Second, we find the optimum thresholds simultaneously using an APSO for those selected variables. To validate the proposed algorithm, several artificial and real datasets are used. We compare our results with the original CART results and show that the proposed algorithm is promising for improving prediction accuracy.

• 한국통신학회논문지
• /
• 제40권7호
• /
• pp.1217-1233
• /
• 2015

The paper studies characteristics of the minimum mean-square error symmetric scalar quantizers for the generalized gamma, Bucklew-Gallagher and Hui-Neuhoff probability density functions. Toward this goal, asymptotic formulas for the inner- and outermost thresholds, and distortion are derived herein for nonuniform quantizers for the Bucklew-Gallagher and Hui-Neuhoff densities, parallelling the previous studies for the generalized gamma density, and optimal uniform and nonuniform quantizers are designed numerically and their characteristics tabulated for integer rates up to 20 and 16 bits, respectively, except for the Hui-Neuhoff density. The assessed asymptotic formulas are found consistently more accurate as the rate increases, essentially making their asymptotic convergence to true values numerically acceptable at the studied bit range, except for the Hui-Neuhoff density, in which case they are still consistent and suggestive of convergence. Also investigated is the uniqueness problem of the differentiation method for finding optimal step sizes of uniform quantizers: it is observed that, for the commonly studied densities, the distortion has a unique local minimizer, hence showing that the differentiation method yields the optimal step size, but also observed that it leads to multiple solutions to numerous generalized gamma densities.

• 전자공학회논문지SC
• /
• 제39권2호
• /
• pp.176-181
• /
• 2002

이 논문에서는 PDA(Probabilistic Data Association) 필터를 이용한 표적 추적에서 평균자승추정오차의 기대값을 최소화하는 표적 검출 문턱값의 최적제어 방법을 제시하고, 닫힌 형태의 준최적해를 구한다. 최적의 검출 문턱값을 구하기 위한 이전의 연구에서는 그래프를 이용한 부정확한 최적화 방법이나, 매우 시간이 많이 소요하는 수치해석적 최적화 알고리듬을 사용하였다. 하지만, 이 논문에서는 정보감소인자의 수치적 근사화식을 이용하여, 최적제어문제로 정식화하여 닫힌 형태의 준최적 검출 문턱값을 구하였다. 이 결과는 실시간 표적 추적에서 유용하게 사용될 수 있다.

• Journal of the Korean Data and Information Science Society
• /
• 제25권6호
• /
• pp.1521-1530
• /
• 2014

The ROC analysis is considered for multiple class diagnosis. There exist many criteria to find optimal thresholds and measure the accuracy of diagnostic tests for k dimensional ROC analysis. In this paper, we proposed a diagnostic accuracy measure called the correct classification simple rate, which is defined as the summation of true rates for each classification distribution and expressed as a function of summation of sequential true rates for two consecutive distributions. This measure does not weight accuracy across categories by the category prevalence and is comparable across populations for multiple class diagnosis. It is found that this accuracy measure does not only have a relationship with Kolmogorov - Smirnov statistics, but also can be represented as a linear function of some optimal threshold criteria. With these facts, the suggested measure could be applied to test for comparing multiple distributions.

• 지질공학
• /
• 제20권1호
• /
• pp.89-100
• /
• 2010

암석의 손상상태를 평가하기위한 여러 방법들이 제안되어 있으나, 일부의 방법은 명확한 손상기준을 제시하기도 하지만 일부의 방법은 매우 모호하여 분석자의 주관에 따라 값이 달라지기도 한다. 그러므로 이 연구에서는 황등화강암을 대상으로 현재까지 제안된 모든 손상기준 결정방법을 적용하여, 각 방법의 적용성, 오차 및 최적의 손상기준결정 방법 등을 연구하였다. 또한 암석의 균열발달 및 파괴특성의 규명에 가장 중요한 손상기준인 균열개시응력과 균열손상응력을 FSR 및 장기 정하중 시험을 이용하여 검정하였다. 황등화강암의 균열닫힘응력과 균열개시응력은 각각 57.5 MPa, 77.6 MPa이며 균열체적변형률에서 측정하는 것이 가장 정확한 것으로 판단된다. 2차 균열개시응력은 90.6 MPa로 측정되었으며, 미소파괴음 계수 및 계수율이 균열개시응력의 측정에 가장 효과적인 것으로 판단된다. 균열결합응력 측정은 체적강성곡선, 미소파괴음 계수 및 미소파괴음 계수율이 가장 효과적인 방법으로 판단되며, 균열결합응력은 110.3 MPa이다. 균열손상응력은 체적강성곡선 및 미소파괴음 계수율에서 가장 명확히 측정되며, 약 127.5 MPa이다. 일축압축강도에 대한 비로서 나타낸 균열개시응력은 0.47로 FSR 값 0.46과 매우 유사하며, 균열손상응력은 0.77로 장기 정하중 시험을 통하여 측정된 장기 강도비 0.75~0.8과 거의 일치하여 균열개시응력 및 균열손상응력 값이 정확함을 검정하였다.

• 정보처리학회논문지C
• /
• 제18C권3호
• /
• pp.127-134
• /
• 2011

악성 봇넷은 DDoS(Distributed Denial of Service) 공격이나 각종 스팸 메시지 발송, 개인 정보 탈취, 클릭 사기 등 많은 악성 행위에 이용되고 있다. 이를 방지하기 위해 많은 연구가 선행되었지만 악성 봇넷 또한 진화하여 탐지 시스템을 회피하고 있다. 특히 최근에는 어플리케이션 계층의 취약성을 공략한 HTTP GET 공격이 주로 사용되고 있다. 한국전자통신연구원에서 개발한 ALADDIN 시스템의 ALAB(Application Layer Attack detection Block)는 서비스 거부 공격 HTTP GET, Incomplete GET Request flooding 공격을 탐지하는 알고리즘이 적용된 탐지 시스템이다. 본 논문에서는 ALAB 탐지 알고리즘의 Incomplete GET 탐지 알고리즘을 확장하고 장기간 조사한 정상적인 패킷 및 공격 패킷들의 분석을 통해 최적 threshold를 도출하여 ALAB 알고리즘의 유효성을 검증한다.

• 대기
• /
• 제21권4호
• /
• pp.497-506
• /
• 2011

Korean mid- and upper-level aviation turbulence guidance (KTG) system is developed using the unified model (UM)-based regional data assimilation and prediction system (RDAPS) of the Korea Meteorological Administration. The KTG system includes three steps. First, the KTG system calculates a suite of diagnostics in the UM-RDAPS domain. Second, component diagnostics that have different units and numerical magnitudes are normalized into the values between 0 and 1, according to their own thresholds in the KTG system. Finally, normalized diagnostics are combined into one KTG predictor by measuring the weighting scores based on the probability of detection, which is calculated using the observed pilot reports (PIREPs) exclusively of moderate-or-greater (MOG) and null (NIL) intensities. To investigate the optimal performance of the KTG system, two types (RD-KTG and UM-KTG) of the KTG systems are developed and evaluated using the PIREPs over Korea and East Asia. Component diagnostics and their thresholds in the RD-KTG are founded on the 8-yrs (2002.12-2010.11) MM5-based RDAPS (previous version of the RDAPS; ${\Delta}x$ = 30 km) and PIREPs data, while those in the UM-KTG are based on the 6 months (2010.12-2011.5) UM-based RDAPS (${\Delta}x$ = 12 km) and PIREPs data. In comparison between the RD-KTG and UM-KTG, overall performance of the UM-KTG (0.815) is better than that of the RD-KTG (0.79) during the recent 6 months, because forecasting skill for the upper-level wind is higher in the UM-RDAPS than in the MM5-RDAPS. It is also found that the UM-KTG is more efficient than the RD-KTG according to the statistical evaluations and sensitivity tests to the number of component diagnostics.

• 한국환경복원기술학회지
• /
• 제16권1호
• /
• pp.159-170
• /
• 2013

The effect of climate change on urban woody plants remains difficult to predict in urban areas. Depending on its tolerances, a plant species may stay and survive or stay with slowly declining remnant populations under a changing climate. To predict those vulnerabilities on urban woody plants, this study suggests a basic bioclimatic envelop model of heat requirements, cold tolerance, chilling requirements and moisture requirements that are well documented as the 'climatic niche'. Each component of the 'climatic niche' is measured by the warmth index, the absolute minimum temperature, the number of chilling weeks and the water balance. Regarding the utility of the developed model, the selected urban plant's present probabilities are suggested in the future climate of Seoul. Both Korea and Japan's thermal thresholds are considered for a plant's optimal climatic niche. By considering the thermal thresholds of these two regions for the same species, the different responses observed will reflect the plant's 'hardening' process in a rising climate. The model illustrated that the subpolar plants Taxus cuspidata and Ulmus davidiana var. japonica are predicted to have low suitability in Seoul. The temperate plants Zelkova serrata and Pinus densiflora, which have a broad climatic niche, exhibited the highest present probability in the future. The subtropical plants Camellia japonica and Castanopsis cuspidata var. sieboldii may exhibit a modest growth pattern in the late 21C's future climatic period when an appropriate frost management scheme is offered. The model can be used to hypothesize how urban ecosystems could change over time. Moreover, the developed model can be used to establish selection guidelines for urban plants with high levels of climatic adaptability.

• 한국위성정보통신학회논문지
• /
• 제8권4호
• /
• pp.100-104
• /
• 2013

본 논문은 열영상 기반의 화염 검출을 위한 기존의 문턱치 설정 기법들을 분석하고 최적 문턱치 설정 방안을 제시한다. 기존의 열영상 기반의 화염검출 알고리즘들은 보통 고정 문턱치를 이용하여 화염 후보영역을 추출하고 후처리를 통해 화염 검출을 최종 판정하므로 화염 후보영역의 결정 과정은 최종 화재 검출 결과에 많은 영향을 준다. 따라서 카메라의 종류나 운영 환경에 따라 입력 영상의 대비와 밝기의 변화가 발생하기 때문에 화염 검출 문턱치는 입력영상의 특성에 연동하여 설정되어져야 한다. 따라서 최적 문턱치 설정 방안을 제시하기 위해 고정 명암도, 평균값, 표준편차 및 최대값을 이용한 문턱치 설정 기법들을 비교 분석하였다. 결론적으로 최적 문턱치는 평균과 표준편차의 합보다 크며 최대값 보다는 작은 값으로 설정 한다면 화염 검출 정확도가 기존 고정 문턱치 방식에 비해 크게 개선될 것으로 기대된다.