• Journal of Broadcast Engineering
• /
• v.20 no.1
• /
• pp.153-163
• /
• 2015

This paper proposes a rain detection and removal algorithm that is robust against camera motion in video sequences. In detection part, the proposed algorithm initially detects possible rain streaks by using intensity properties and spatial properties. Then, the rain streak candidates are selected based on Gaussian distribution model. In removal part, a non-rain block matching algorithm is performed between adjacent frames to find similar blocks to the block that has rain pixels. If the similar blocks to the block are obtained, the rain region of the block is reconstructed by non-local means (NLM) filter using the similar neighbors. Experimental results show that the proposed algorithm outperforms the previous works in terms of subjective visual quality of de-rained video sequences.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.47 no.6
• /
• pp.28-38
• /
• 2010

In order to remove the noise contained in recorded analog video, it is important to recognize the real characteristics and strength of the noise. This paper presents an efficient training-based noise reduction method for recorded analog video after analyzing the noise characteristics of analog video captured in a real broadcasting system. First we show that there is non-negligible noise correlation in recorded analog video and describe the limitations of the traditional noise estimation and reduction methods based on additive white Gaussian noise (AWGN) model. In addition, we show that auto-regressive (AR) model considering noise correlation can be successfully utilized to estimate and synthesize the noise contained in the recorded analog video, and the estimated AR parameters are utilized in the training-based noise reduction scheme to reduce the video noise. Experiment results show that the proposed method can be efficiently applied for noise reduction of recorded analog video with non-negligible noise correlation.

• Journal of KIISE:Software and Applications
• /
• v.32 no.9
• /
• pp.901-911
• /
• 2005

With the development of e-commerce and the proliferation of easily accessible information, information filtering has become a popular technique to prune large information spaces so that users are directed toward those items that best meet their needs and preferences. While many collaborative filtering systems have succeeded in capturing the similarities among users or items based on ratings to provide good recommendations, there are still some challenges for them to be more efficient, especially the user bias problem, non-transitive association problem and cold start problem. Those three problems impede us to capture more accurate similarities among users or items. In this paper, we provide probabilistic model approaches for UCHM and ICHM which are suggested to solve the addressed problems in hopes of achieving better performance. In this probabilistic model, objects (users or items) are classified into groups and predictions are made for users considering the Gaussian distribution of user ratings. Experiments on a real-word data set illustrate that our proposed approach is comparable with others.

• Wind and Structures
• /
• v.36 no.6
• /
• pp.393-404
• /
• 2023

Skewness and kurtosis are important higher-order statistics for simulating non-Gaussian wind pressure series on low-rise buildings, but their predictions are less studied in comparison with those of the low order statistics as mean and rms. The distribution gradients of skewness and kurtosis on roofs are evidently higher than those of mean and rms, which increases their prediction difficulty. The conventional artificial neural networks (ANNs) used for predicting mean and rms show unsatisfactory accuracy in predicting skewness and kurtosis owing to the limited capacity of shallow learning of ANNs. In this work, the deep neural networks (DNNs) model with the ability of deep learning is introduced to predict the skewness and kurtosis on a low-rise building. For obtaining the optimal generalization of the DNNs model, the hyper parameters are automatically determined by Bayesian Optimization (BO). Moreover, for providing a benchmark for future studies on predicting higher order statistics, the data sets for training and testing the DNNs model are extracted from the internationally open NIST-UWO database, and the prediction errors of all taps are comprehensively quantified by various error metrices. The results show that the prediction accuracy in this study is apparently better than that in the literature, since the correlation coefficient between the predicted and experimental results is 0.99 and 0.75 in this paper and the literature respectively. In the untrained cornering wind direction, the distributions of skewness and kurtosis are well captured by DNNs on the whole building including the roof corner with strong non-normality, and the correlation coefficients between the predicted and experimental results are 0.99 and 0.95 for skewness and kurtosis respectively.

• Tribology and Lubricants
• /
• v.17 no.3
• /
• pp.236-243
• /
• 2001

In this study, flow factors are evaluated in terms of kurtosis using random rough surface generated numerically. As h/$\sigma$become large ø$\sub$x/, ø$\sub$y/, ø$\sub$fp/, approach to 1 and ø$\sub$s/, ø$\sub$fs/ to 0 asymptotically regardless of kurtosis. ø$\sub$x/, ø$\sub$y/, ø$\sub$fp/ increase with increasing kurtosis in the mixed lubrication regime. ø$\sub$s/, ø$\sub$fs/ is associated with an additional flow transport due to the combined effect of sliding and roughness. As h/$\sigma$ decreases ø$\sub$s/, ø$\sub$fs/ increase up to a certain point, and then decrease toward zero. This behavior can be attributed to the increasing number of contacts in the mixed lubrication regime. ø$\sub$x/ in the presence of elastic deformation on the surface is larger than ø$\sub$x/ in the absence of it because local film thickness(h$\sub$T/) increases by elastic deformation.

• Journal of the Korean Statistical Society
• /
• v.26 no.2
• /
• pp.231-243
• /
• 1997

In this paper we investigate weak convergence of the intergral processes whose index set is the non-compact infinite time interval. Our first goal is to develop the empirical central limit theorem as random elements of [0, .infty.) for an integral process which is constructed from iid variables. In developing the weak convergence as random elements of D[0, .infty.), we will use a result of Ossiander(4) whose proof heavily depends on the total boundedness of the index set. Our next goal is to establish the empirical central limit theorem for the Kaplan-Meier integral process as random elements of D[0, .infty.). In achieving the the goal, we will use the above iid result, a representation of State(6) on the Kaplan-Meier integral, and a lemma on the uniform order of convergence. The first result, in some sense, generalizes the result of empirical central limit therem of Pollard(5) where the process is regarded as random elements of D[-.infty., .infty.] and the sample paths of limiting Gaussian process may jump. The second result generalizes the first result to random censorship model. The later also generalizes one dimensional central limit theorem of Stute(6) to a process version. These results may be used in the nonparametric statistical inference.

• Proceedings of the KIEE Conference
• /
• 1993.07b
• /
• pp.1247-1249
• /
• 1993

In this paper, NUDFET(NonUniformly Doped Field Effect Transistor) is presented as an alternative which offers the possibility of reducing the power necessary to operate switching circuits without a substantial loss in speed. The purpose of this NUDFET is to modify the electric field profile in order to cause carrier velocity saturation to occur at a lower voltage than it would occur in the uniformly doped device of the same channel length. The more MESFET and NUDFET circuits are realized, the more accurate model ins the performance of these devices become required. Analytic model ins was replaced by numerical analysis because of the complexity of device configuration. In this paper, FEM is selected because of simpler local mesh refinement and smaller computer memory than FDM. For accurate analysis, this paper has applied the Scharfetter-Gummel(S-G) Scheme and seven-point Gaussian Quadrature rule to assembly of the finite-element stiffness matrices and right-hand side vector of the semiconductor equations.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.41 no.8
• /
• pp.851-858
• /
• 2016

This paper presents a new statistical voice activity detection (VAD) based on the probabilistic interpretation of nonnegative matrix factorization (NMF). The objective function of the NMF using Kullback-Leibler divergence coincides with the negative log likelihood function of the data if the distribution of the data given the basis and encoding matrices is modeled as Poisson distributions. Based on this probabilistic NMF, the VAD is constructed using the likelihood ratio test assuming that speech and noise follow Poisson distributions. Experimental results show that the proposed approach outperformed the conventional Gaussian model-based and NMF-based methods at 0-15 dB signal-to-noise ratio simulation conditions.

• The Bulletin of The Korean Astronomical Society
• /
• v.46 no.2
• /
• pp.70.1-70.1
• /
• 2021

We examine the HI gas kinematics and distributions of galaxy pairs in group or cluster environments from high-resolution Australian Square Kilometer Array Pathfinder (ASKAP) WALLABY pilot observations. We use 32 well-resolved close pair galaxies from the Hydra, Norma, and NGC 4636, two clusters and a group of which are identified by their spectroscopy information and additional visual inspection. We perform profile decomposition of HI velocity profiles of the galaxies using a new tool, BAYGAUD which allows us to separate a line-of-sight velocity profile into an optimal number of Gaussian components based on Bayesian MCMC techniques. Then, we construct super profiles via stacking of individual HI velocity profiles after aligning their central velocities. We fit a model which consists of double Gaussian components to the super profiles, and classify them as kinematically cold and warm HI gas components with respect to their velocity dispersions, narrower or wider 𝜎, respectively. The kinematically cold HI gas reservoir (M_cold/M_HI) of the paired galaxies is found to be relatively higher than that of unpaired control samples in the clusters and the group, showing a positive correlation with the HI mass in general. Additionally, we quantify the gravitational instability of the HI gas disk of the sample galaxies using their Toomre Q parameters and HI morphological disturbances. While no significant difference is found for the Q parameter values between the paired and unpaired galaxies, the paired galaxies tend to have larger HI asymmetry values which are derived using their moment0 map compared to those of the non-paired control sample galaxies in the distribution.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.6 no.2
• /
• pp.288-295
• /
• 2011

The paper proposes a localization method of an underwater robot using Monte Carlo Localization(MCL) approach. Localization is one of the fundamental basics for autonomous navigation of an underwater robot. The proposed method resolves the problem of accumulation of position error which is fatal to dead reckoning method. It deals with uncertainty of the robot motion and uncertainty of sensor data in probabilistic approach. Especially, it can model the nonlinear motion transition and non Gaussian probabilistic sensor characteristics. In the paper, motion model is described using Euler angles to utilize the MCL algorithm for position estimation of an underwater robot. Motion model and sensor model are implemented and the performance of the proposed method is verified through simulation.