• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.283-285
• /
• 2001

유전자 프로그래밍은 프로그램 자동생성 도구이다. 문제를 해결하는 프로그램코드를 프로그래머가 직접 구현하는 것이 아니라, 적절한 초기값만을 입력하여 컴퓨터가 스스로 적합한 해를 찾아내도록 하는 방법이다. 유전자 프로그래밍은 생물의 진화개념에서 얻어진 여러 아이디어를 사용하여 최적화된 해를 찾아낸다. 본 논문에서는 세포영상인식 문제를 해결하기 위하여 유전자 프로그래밍을 사용하였다. 실험에 사용된 영상은 자궁경부세포진 영상이다. 여러 가지 종류와 상태의 세포들이 뒤섞여 있어 분석하기에 힘들다는 것이 이 영상의 특징이다. 주어진 문제는 샘플 영상이 암인가 아닌가를 판별하는 것이다. 유전자 프로그래밍을 적용하기 위하여 사용한 특징값들은 영상에서 찾을 수 있는 가장 단순한 대표값들과, 산술 및 논리연산자들이다. 실험결과 실제 인식기 제작에 바로 적용하기엔 무리가 있지만, 80%정도를 제대로 판별해 낼수 있었다. 인식률이 낮은 이유는 사용한 특징들이 영상의 정보를 잘 흡수하지 못했기 때문이라 여겨지고, 앞으로 지나치게 복잡하지 않으면서 여상의 특징을 잘 표현하는 특징값들을 찾는 것이 향후과제이다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.31 no.10
• /
• pp.575-587
• /
• 2004

In this paper. a new Fast Expectation-Maximization algorithm(FEM) is proposed. Firstly the K-means algorithm is modified to reduce the number of iterations for finding the initial values that are used as the initial values in EM process. Conventionally the Initial values in K-means clustering are chosen randomly. which sometimes forces the process of clustering converge to some undesired center points. Uniform partitioning method is added to the conventional K-means to extract the proper initial points for each clusters. Secondly the effect of posterior probability is emphasized such that the application of Maximum Likelihood Posterior(MLP) yields fast convergence. The proposed FEM strengthens the characteristics of conventional EM by reinforcing the speed of convergence. The superiority of FEM is demonstrated in experimental results by presenting the improvement results of EM and accelerating the speed of convergence in parameter estimation procedures.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.9 no.5
• /
• pp.490-495
• /
• 1999

In this paper. a learning method VoD-EBP for neural networks is proposed, which learn patterns by error back propagation. Based on Voronoi diagram, the method initializes the weights of the neural networks systematically, wh~ch results in faster learning speed and alleviated local optimum problem. The method also shows better the reliability of the design of neural network because proper number of hidden nodes are determined from the analysis of Voronoi diagram. For testing the performance, this paper shows the results of solving the XOR problem and the parity problem. The results were showed faster learning speed than ordinary error back propagation algorithm. In solving the problem, local optimum problems have not been observed.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.06a
• /
• pp.1-4
• /
• 2016

효율적으로 깊이지도를 획득하기 위해 다양한 방법의 지역 기반스테레오 매칭 방법이 사용된다. 일반적인 지역기반 스테레오 매칭에 사용되는 비용값 계산 방법을 통해 깊이지도를 생성하게 되면 객체의 경계 영역이 무너지거나, 유사한 텍스쳐 정보가 연속적으로 나타나는 영역에서 부정확한 깊이값을 얻는 문제가 발생한다. 본 논문에서는 깊이지도의 정확성을 높이기 위해 2가지 단계를 거쳐 최종 깊이지도를 생성한다. 처음으로, 일반적으로 사용하는 지역기반 스테레오 매칭 비용 함수와 입력 영상의 기울기를 고려한 초기 비용값을 가이드 필터를 이용하여 최적의 비용값을 찾아 초기 변위지도를 생성한다. 스테레오매칭을 수행할 경우, 시점의 차이로 인해 보이지 않는 영역에서 정확한 변위값을 찾지 못하는 문제가 발생한다. 이러한 문제를 해결하기 위해 좌영상과 우영상을 기반으로 획득한 변위지도를 사용하여 교차검사를 함으로써 폐색영역을 찾아낸다. 폐색 영역을 이웃한 화소의 값을 사용하여 채울 경우 실선과 같은 오류가 결과 영상에 나타나게 된다. 이러한 오류 영역을 제거하기 위해 마지막으로 가중치를 적용한 중간값 필터를 적용한다. 실험 결과 제안한 방법을 사용하여 획득한 깊이지도가 기존의 방법보다 정확한 깊이값을 얻는 것을 확인할 수 있었다.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2007.04a
• /
• pp.386-389
• /
• 2007

EBGM은 자세와 포즈, 조명 변화에 강인한 얼굴 인식 기법중의 하나이다. 하지만 EBGM을 통한 얼굴 인식 시스템은 얼굴의 특징점을 추출하기 위해 주어지는 초기값에 상당한 영향을 받는다. 이러한 문제를 해결하기 위해서 얼굴의 윤곽 추출에 우수한 성능을 보이는 AAM을 통하여 EBGM의 초기값을 주고 EBGM을 통하여 개선하는 방법을 제안하였었다. 본 논문에서는 등록자마다 다른 경계값을 갖는 방법을 제안한다. 기존의 경계값에 비해 성능의 향상이 어느 정도 이뤄지는가에 대해 다룰 것이다.

• Convergence Security Journal
• /
• v.11 no.3
• /
• pp.25-30
• /
• 2011

WEP is proposed networks dominate the market in the future wireless LAN encryption and authentication features to provide a secure protocol. However, WEP does not suggest a specific measures when generating the initial values used for the creation cipher text, the initial value problem because tile size and no-encryption if you have been raised about the safety issue. In this paper pointed out the vulnerabilities of WEP and the proposed improvement plan for this improvement was proposed based on the initial value to avoid re-creating the initial value of the system and using a block cipher in CBC mode for confidentiality and to provide mutual authentication New WLAN security model was proposed.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.1
• /
• pp.338-348
• /
• 2020

This study examined the longitudinal relationship between paternal depressive symptoms and maternal depressive symptoms simultaneously. This study also identified the interplay of paternal and maternal depressive symptoms for predicting elementary children's language development, their problem behaviors and their school adjustment. Using the data from the Panel Study on Korean Children for the 4th-8th years (2011~2015 year), this study used the Latent Growth Curve Model, which is helpful for examining longitudinal relationship differences among variables. The sample subjects were 1,754 parents and children. The results are as follows. The initial level of paternal depressive symptoms had a positive impact on the rate of change in maternal depressive symptoms. The initial level of maternal depressive symptoms had a negative impact on the rate of change in maternal depressive symptoms. Mothers' depressive symptoms showed not only the mothers' own depression problem. but also the self-effect and counterpart effect of depression on the fathers' depression problem. The rate of change in maternal depressive symptom mediated the relation between the initial level of parental depressive symptoms and children's receptive language, internalizing/externalizing problems, and school adjustment. Therefore, depression prevention and intervention programs for both fathers and mothers are needed for the healthy development and school adaptation of school-age children.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.21 no.2
• /
• pp.199-208
• /
• 2008

In this paper, to analyze the initial valued non-homogeneous elastic half space by the scaled boundary analysis, the infinite element approach was introduced. The free surface of the initial valued non-homogeneous elastic half space was modeled as a circumferential direction of boundary scaled boundary coordinate. The infinite element was used to represent the infinite length of the free surface. The initial value of material property(elastic modulus) was considered by the combination of the position of the scaling center and the power function of the radial direction. By use of the mapping type infinite element, the consistent elements formulation could be available. The performance and the feasibility of proposed approach are examined by two numerical examples.

• The Korean Journal of Applied Statistics
• /
• v.32 no.5
• /
• pp.693-702
• /
• 2019

Deep learning has gained popularity for the classification and prediction task. Neural network layers become deeper as more data becomes available. Saturation is the phenomenon that the gradient of an activation function gets closer to 0 and can happen when the value of weight is too big. Increased importance has been placed on the issue of saturation which limits the ability of weight to learn. To resolve this problem, Glorot and Bengio (Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, 249-256, 2010) claimed that efficient neural network training is possible when data flows variously between layers. They argued that variance over the output of each layer and variance over input of each layer are equal. They proposed a method of initialization that the variance of the output of each layer and the variance of the input should be the same. In this paper, we propose a new method of establishing initialization by adopting truncated normal distribution and truncated cauchy distribution. We decide where to truncate the distribution while adapting the initialization method by Glorot and Bengio (2010). Variances are made over output and input equal that are then accomplished by setting variances equal to the variance of truncated distribution. It manipulates the distribution so that the initial values of weights would not grow so large and with values that simultaneously get close to zero. To compare the performance of our proposed method with existing methods, we conducted experiments on MNIST and CIFAR-10 data using DNN and CNN. Our proposed method outperformed existing methods in terms of accuracy.

• Journal for History of Mathematics
• /
• v.23 no.3
• /
• pp.107-119
• /
• 2010

In this paper we study the expectation from the past which would be based in the initial Probability Theory. These study can show the value of expectation in the initial concept of Probability and illuminate the concept being taught.