• 정보처리학회논문지B
• /
• 제13B권5호
• /
• pp.515-524
• /
• 2006

배경추출은 비디오 감시 시스템에서 움직이는 물체를 찾는데 중요한 기술이다. 본 논문에서는 벡터 정렬을 이용한 새로운 온라인 컬러 배경 추출 방법을 제안한다. 제안된 방법에서 배경은 물체보다 발생빈도가 높다는 사실을 이용하여, 연속된 프레임의 컬러화소 값들의 벡터 미디언을 그 화소에서의 배경이라 간주한다. 본 알고리즘에서 현재 프레임의 물체는 얻어진 배경과의 거리가 문턱치보다 큰 화소들의 집합으로 구성된다. 알고리즘의 성능을 평가하기 위하여 온라인 가우시안 혼합 모델(Gaussian Mixture Model)을 이용한 다중 배경추출 방법과 비교하였으며, 비교결과 유사 또는 우월한 실험 결과를 확인하였다.

• 한국ITS학회 논문지
• /
• 제7권1호
• /
• pp.32-41
• /
• 2008

비디오를 이용한 비전기반 감시에서 움직이는 객체의 추적은 GMM (Gaussian Mixture Model)을 사용한 배경영상과 현재영상의 차이법을 이용한다. 문턱치를 통해 생성된 이진영상을 이용하여 객체 추적을 할 경우 객체 정보가 아닌 그림자에 의하여 객체가 병합되는 현상이 나타난다. 본 논문에서는 신경망(Backpropagation Neural Network)을 이용하여 그림자를 제거하는 방법을 제안하였다. 10개의 동영상에서 객체영역과 캐스트그림자(Cast-Shadow)영역의 훈련용 이미지에서 특징 값을 추출하여 신경망을 훈련시켰다. 캐스트그림자를 제거하는 방법은 이진영상의 객체로 추정되는 영역에서 그림자를 분리하는 방법을 기초로 하며 기존의 그림자 제거 알고리즘 (SNP, SP, DNM1, DNM2, CNCC)보다 그림자 제거 성능이 (16.2%, 38.2%, 28.1%, 22.3%, 44.4%)로 높게 나타났다.

• 말소리와 음성과학
• /
• 제6권2호
• /
• pp.21-28
• /
• 2014

We propose a novel scheme for digital audio authentication of given audio files which are edited by inserting small audio segments from different environmental sources. The purpose of this research is to detect inserted sections from given audio files. We expect that the proposed method will assist human investigators by notifying suspected audio section which considered to be recorded or transmitted on different environments. GMM-UBM and GSV-SVM are applied for modeling the dominant environment of a given audio file. Four kinds of likelihood ratio based scores and SVM score are used to measure the likelihood for a dominant environment model. We also use an ensemble score which is a combination of the aforementioned five kinds of scores. In the experimental results, the proposed method shows the lowest average equal error rate when we use the ensemble score. Even when dominant environments were unknown, the proposed method gives a similar accuracy.

• 한국정보통신학회논문지
• /
• 제21권12호
• /
• pp.2291-2297
• /
• 2017

본 논문에서는 음성 데이터베이스를 평가하기 위해 여러 가지의 음성 특성 지표 추출 알고리즘을 설명하고 심층 신경망 기반의 새로운 음성 성능 지표 생성 방법을 제안한다. 선행 연구에서는 효과적인 음성 인식 성능 지표를 생성하기 위해 대표적인 음성 인식 성능 지표인 단어 오인식률(Word Error Rate, WER)과 상관도가 높은 여러 가지 음성 특성 지표들을 조합하여 새로운 성능 지표를 생성하였다. 생성된 음성 성능 지표는 다양한 잡음 환경에서 각 음성 특성 지표를 단독으로 사용할 때보다 단어 오인식률과 높은 상관도를 나타내어 음성 인식 성능을 예측하는데 효과적임을 입증 하였다. 본 논문에서는 심층 신경망을 기반으로 한 음성 특성 지표 추출 방법에 대해 설명하며 선행 연구에서 조합에 사용한 GMM(Gaussian Mixture Model) 음향 모델 확률 값을 심층 신경망 학습을 통해 추출한 확률 값으로 대체해 조합함으로써 단어 오인식률과 보다 높은 상관도를 갖는 것을 확인한다.

• 한국정보과학회논문지:데이타베이스
• /
• 제28권3호
• /
• pp.324-333
• /
• 2001

본 논문에서는 가우시안 혼합모형을 이용한 새로운 칼라 영상의 분할 알고리즘을 제안한다. 기존의 EM 알고리즘의 문제점인 국부적 최대값의 문제를 해결하기 위하여 최대 엔트로피의 원리를 이용하는 결정적 어닐링 EM 알고리즘을 소개하였고, 여러 색상들로 구성된 영상에 대하여 가우시안 혼합모형을 가정하였으며, 결정적 어닐링 EM 알고리즘을 사용하여 이들의 모수를 추정하는 방법을 알아보았다. 또한 혼합모형에 성분의 수를 자동으로 결정할 수 있는 방법을 제시하였으며 선택된 최적의 혼합모형을 사용하여 각 화소에 대한 사후확률을 계산하고 이들의 최대값을 이용하여 영상분할을 실시하였다. 결정적 어닐링 EM 알고리즘이 기존의 EM 알고리즘보다 혼합모형의 모수를 더 정확하게 추정한다는 것과 혼합모형의 성분의 수를 결정하는 제안된 방법의 성능을 실험결과를 통하여 고찰하였고, 또한 두 가지 실제 영상을 통하여 제안된 알고리즘이 기존의 알고리즘 보다 영상을 더 효율적으로 분할 할 수 있음을 보였다.

• Journal of Korea Trade
• /
• 제27권5호
• /
• pp.153-166
• /
• 2023

Purpose - This research empirically analyzes the influence of economic policy uncertainty and free trade agreements (FTAs) on bilateral trade volumes between Korea and its trading partners. The study investigates whether fluctuations in the Economic Policy Uncertainty Index (EPUI) for both Korea and its trading partners significantly impact trade volumes and whether the implementation of FTAs mitigates these effects. Design/methodology - The study employs dynamic panel data analysis using the system generalized method of moments (system GMM) estimation method to achieve its research objectives. It utilizes country-month-level panel data, including the EPUI, trade volume between Korea and its trading partner countries, and other pertinent variables. The use of system GMM allows for the control of potential endogeneity issues and the incorporation of country-specific and time-specific effects. Findings - The analysis yields significant results regarding the impact of economic policy uncertainty on Korea's exports and imports, particularly before the implementation of FTAs. An increase in the EPUI of trading partners leads to a notable increase in Korea's exports to them. Conversely, an increase in Korea's EPUI negatively affects its imports from trading partners. However, post-FTA implementation, the influence of each country's EPUI on trade volume is neutralized, with no significant difference observed. Originality/value - This research contributes to the existing literature by providing empirical evidence on the interaction effects between economic policy uncertainty and FTAs on bilateral trade volumes. The study's uniqueness lies in its examination of how FTAs mitigate the impact of economic uncertainty on trade relations between countries. The findings underscore the importance of trade agreements as mechanisms to address economic risks and promote international trade relations. In a world where global market uncertainties persist, these insights can aid policymakers in Korea and other countries in enhancing their trade cooperation strategies and navigating challenges posed by evolving economic landscapes.

• Journal of Electrical Engineering and Technology
• /
• 제9권6호
• /
• pp.2118-2125
• /
• 2014

Gait Recognition is a new technique to identify the people by the way they walk. Human gait is a spatio-temporal phenomenon that typifies the motion characteristics of an individual. The proposed method makes a simple but efficient attempt to gait recognition. For each video file, spatial silhouettes of a walker are extracted by an improved background subtraction procedure using Gaussian Mixture Model (GMM). Here GMM is used as a parametric probability density function represented as a weighted sum of Gaussian component densities. Then, the relevant features are extracted from the silhouette tracked from the given video file using the Principal Component Analysis (PCA) method. The Fisher Linear Discriminant Analysis (FLDA) classifier is used in the classification of dimensional reduced image derived by the PCA method for gait recognition. Although gait images can be easily acquired, the gait recognition is affected by clothes, shoes, carrying status and specific physical condition of an individual. To overcome this problem, it is combined with footprint as a multimodal biometric system. The minutiae is extracted from the footprint and then fused with silhouette image using the Discrete Stationary Wavelet Transform (DSWT). The experimental result shows that the efficiency of proposed fusion algorithm works well and attains better result while comparing with other fusion schemes.

• Asian Journal of Business Environment
• /
• 제10권2호
• /
• pp.23-31
• /
• 2020

Purpose: The purpose of this paper is to find out the impact of financial leverage on firm's profitability in the listed textile sector of Bangladesh. Research design, data and methodology: A sample of 22 DSE listed textile firms has been used to conduct the study. In this study, firm profitability is measured by Return on Equity (ROE) and both short term debt and long term debt are used as the as proxies of financial leverage. Pooled Ordinary Least Squares (OLS), Fixed Effect (FE), and Generalized Method of Moments (GMM) models have been used to test the relationship between financial leverage and profitability of firms. Result: This study finds a significant negative relationship between leverage and firm's profitability using the Pooled OLS method. The result is also consistent with the fixed effect and GMM method. This result implies that firm's profitability is negatively affected by the firm's capital structure. Conclusion: The study concludes that maximum textile firms use external debt as a source of finance as they don't have sufficient internally generated funds. This study recommends that firm should give more emphasize on generating fund internally to meet up their financing needs.

• 전기학회논문지
• /
• 제61권12호
• /
• pp.1920-1925
• /
• 2012

In this paper, we propose a novel method to improve the performance of the improved minima controlled recursive averaging (IMCRA) in non-stationary noisy environment. The conventional IMCRA algorithm efficiently estimate the noise power by averaging past spectral power values based on a smoothing parameter that is adjusted by the signal presence probability in frequency subbands. Since the minimum of smoothing parameter is defined as 0.85, it is difficult to obtain the robust estimates of the noise power in non-stationary noisy environments that is rapidly changed the spectral characteristics such as babble noise. For this reason, we proposed the modified IMCRA, which adaptively estimate and updata the noise power according to the noise type classified by the Gaussian mixture model (GMM). The performances of the proposed method are evaluated by perceptual evaluation of speech quality (PESQ) and composite measure under various environments and better results compared with the conventional method are obtained.

• 대한원격탐사학회지
• /
• 제35권1호
• /
• pp.93-115
• /
• 2019

The availability of high-resolution laser scanning data and advanced machine learning algorithms has enabled an accurate potential rockfall source identification. However, the presence of other mass movements, such as landslides within the same region of interest, poses additional challenges to this task. Thus, this research presents a method based on an integration of Gaussian mixture model (GMM) and ensemble artificial neural network (bagging ANN [BANN]) for automatic detection of potential rockfall sources at Kinta Valley area, Malaysia. The GMM was utilised to determine slope angle thresholds of various geomorphological units. Different algorithms(ANN, support vector machine [SVM] and k nearest neighbour [kNN]) were individually tested with various ensemble models (bagging, voting and boosting). Grid search method was adopted to optimise the hyperparameters of the investigated base models. The proposed model achieves excellent results with success and prediction accuracies at 95% and 94%, respectively. In addition, this technique has achieved excellent accuracies (ROC = 95%) over other methods used. Moreover, the proposed model has achieved the optimal prediction accuracies (92%) on the basis of testing data, thereby indicating that the model can be generalised and replicated in different regions, and the proposed method can be applied to various landslide studies.