• ETRI Journal
• /
• v.37 no.4
• /
• pp.780-786
• /
• 2015

This paper describes an audio source separation that is based on nonnegative matrix factorization (NMF) and expectation maximization (EM). For stable and highperformance separation, an effective auxiliary source separation that extracts source residuals and reprojects them onto proper sources is proposed by taking into account an ambiguous region among sources and a source's refinement. Specifically, an additional NMF (model) is designed for the ambiguous region - whose elements are not easily represented by any existing or predefined NMFs of the sources. The residual signal can be extracted by inserting the aforementioned model into the NMF-EM-based audio separation. Then, it is refined by the weighted parameters of the separation and reprojected onto the separated sources. Experimental results demonstrate that the proposed scheme (outlined above) is more stable and outperforms existing algorithms by, on average, 4.4 dB in terms of the source distortion ratio.

• Proceedings of the IEEK Conference
• /
• 1999.11a
• /
• pp.569-572
• /
• 1999

In this paper we present a color image segmentation algorithm based on statistical models. A novel deterministic annealing Expectation Maximization(EM) formula is derived to estimate the parameters of the Gaussian Mixture Model(GMM) which represents the multi-colored objects statistically. The experimental results show that the proposed deterministic annealing EM is a global optimal solution for the ML parameter estimation and the image field is segmented efficiently by using the parameter estimates.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04b
• /
• pp.529-531
• /
• 2001

We present an expectation-maximization algorithm for principal component analysis via orthogonalization. The algorithm finds actual principal components, whereas previously proposed EM algorithms can only find principal subspace. New algorithm is simple and more efficient thant probabilistic PCA specially in noiseless cases. Conventional PCA needs computation of inverse of the covariance matrices, which makes the algorithm prohibitively expensive when the dimensions of data space is large. This EM algorithm is very powerful for high dimensional data when only a few principal components are needed.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.103-106
• /
• 2002

본 논문에서는 입력음성신호로부터 음소간의 경계를 찾는 문제를 풀기위해 재귀적인 방식으로 EM 알고리즘을 적용한다. 즉, 예상되는 두 끝점 사이의 부분을 현재의 프레임 n 이라고 하면, 그 전 프레임 n-1 에서 구해진 끝점이 주는 정보와 그 끝점으로부터 이어지는 음성샘플로부터 현재 프레임의 끝점을 구한다. 또한 현재의 프레임 n 에서 끝점을 추정해 내면, 그 추정한 끝점과 그 점 이후에 이어지는 음성샘플값으로부터 다음 프레임 n+1 의 끝점을 구한다. 이러한 방식을 재귀적인 음소분리 방식이라고 한다. 그리고, 각 프레임에서 끝점을 구하기 위해서는 끝점의 좌표를 추정해야 할 파라메터로 하고, 그 주변의 음성샘플 값을 관찰 값으로 하여 EM(Expectation and Maximization) 알고리즘을 이용한다. 이 EM 알고리즘을 이용한 재귀적인 음소분리 방식을 실제 음성 DB 로부터 음소쌍을 추출하여 테스트 했을 때 약 5 회의 EM 반복 후에 경계간으로 수렴함을 볼 수 있었다.

• The KIPS Transactions:PartC
• /
• v.17C no.2
• /
• pp.181-189
• /
• 2010

HAP(High Altitude Platform) is a stationary aerial platform positioned in the stratosphere between 17Km and 22Km height and it could act as an MBS (Mobile Base Station). HAP based Network has advantages of both satellite system and terrestrial communication system. In this paper we study the deploy of multiple HAP MBS that can provides efficient communication for users. For this study, EM(Expectation Maximization) clustering algorithm is used to cluster terrestrial mobile nodes. The object of this paper is improving EM algorithm into the clustering algorithm for efficiency in variety aspects considering distance between mobile terminal units and speed of mobile terminal units, and estimating performance of HAP MBS deploy technique with use of improved EM algorithm using RWP (Random Waypoint) node mobility.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10c
• /
• pp.360-362
• /
• 1998

In the original mixtures of experts framework, the parameters of the network are determined by gradient descent, which is naturally slow. In [2], the Expectation-Maximization(EM) algorithm is used instead, to obtain the network parameters, resulting in substantially reduced training times. This paper presents the new EM algorithm for prediction. We show that an Efficient training algorithm may be derived for the HME network. To verify the utility of the algorithm we look at specific examples in time series prediction. The application of the new EM algorithm to time series prediction has been quiet successful.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.405-406
• /
• 2006

This paper proposes the new algorithm for the extraction of the corresponding points. Our algorithm is based on RANSAC(Random Sample Consensus) with EM(Expectation-Maximization). In the procedure of RANSAC, N-points are selected by the result of EM instead of the random selection. EM+SAC algorithm is applied to the correspondence for the mosaicing.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06b
• /
• pp.222-225
• /
• 2011

이전에 수행된 연구에서 인간의 이동 패턴은 Levy flight 행동을 보인다고 알려져있다. 그러나 우리의 경험적 지식을 바탕으로 생각해 볼 때 인간의 이동 패턴을 Levy flight 행동만 가지고 나타내기에는 한계가 있어 보인다. 인간의 이동 패턴은 주위환경, 시간, 개인의 습관, 그리고 사회적 지위 등에 따라 서로 다른 모양을 보인다. 즉, 인간 이동의 형태를 파악하기 위해서는 좀 더 다양한 정보가 있어야만 인간 이동의 패턴을 사실적으로 모델링 할 수 있다. 인간의 이동 패턴을 사실적으로 모델링하기에 필요한 정보를 얻기 위해서 상향식 방법(Bottom up)으로 우선 실제 이동 패턴을 분석하여 모델링에 필요한 정보를 추출하고 다시 그 정보를 검증하는 과정으로 모델링에 필요한 정보가 구체적으로 나타나게 될 것이다. 이에 실제 인간의 이동 패턴을 분석하기 위해 아무런 매개변수 없이 개인의 GPS 데이터를 바탕으로 위치정보만을 가지고 군집화(Clustering)를 하게 되면 특정 위치에 대한 군집이 생성된다. 이러한 군집이 나타내는 것은 자주 머무는 지역, 이동 경로 등이 될 것이다. 본 논문에서는 인간의 이동 정보인 GPS 데이터를 가지고 EM 군집화를 통하여 생성된 군집을 통해 인간의 이동 패턴을 분석할 것이다.

• IEIE Transactions on Smart Processing and Computing
• /
• v.4 no.4
• /
• pp.202-208
• /
• 2015

In this study, we propose a new inference algorithm for a multiclass Gaussian process classification model using a variational EM framework and the Laplace approximation (LA) technique. This is performed in two steps, called expectation and maximization. First, in the expectation step (E-step), using Bayes' theorem and the LA technique, we derive the approximate posterior distribution of the latent function, indicating the possibility that each observation belongs to a certain class in the Gaussian process classification model. In the maximization step, we compute the maximum likelihood estimators for hyper-parameters of a covariance matrix necessary to define the prior distribution of the latent function by using the posterior distribution derived in the E-step. These steps iteratively repeat until a convergence condition is satisfied. Moreover, we conducted the experiments by using synthetic data and Iris data in order to verify the performance of the proposed algorithm. Experimental results reveal that the proposed algorithm shows good performance on these datasets.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.5
• /
• pp.2299-2318
• /
• 2019

Location-aware networks are of great importance for both civil lives and military applications. Methods based on line-of-sight (LOS) measurements suffer sever performance loss in harsh environments such as indoor scenarios, where sensors can receive both LOS and non-line-of-sight (NLOS) measurements. In this paper, we propose a data association (DA) process based on the expectation maximization (EM) algorithm, which enables us to exploit multipath components (MPCs). By setting the mapping relationship between the measurements and scatters as a latent variable, coefficients of the Gaussian mixture model are estimated. Moreover, considering the misalignment of sensor position, we propose a space-alternating generalized expectation maximization (SAGE)-based algorithms to jointly update the target localization and sensor position information. A two dimensional (2-D) circularly symmetric Gaussian distribution is employed to approximate the probability density function of the sensor's position uncertainty via the minimization of the Kullback-Leibler divergence (KLD), which enables us to calculate the expectation step with low computational complexity. Moreover, a distributed implementation is derived based on the average consensus method to improve the scalability of the proposed algorithm. Simulation results demonstrate that the proposed centralized and distributed algorithms can perform close to the Monte Carlo-based method with much lower communication overhead and computational complexity.