• 제목/요약/키워드: temporal structure of source

검색결과 14건 처리시간 0.028초

유사-가능도 최대화를 통한 가우시안 프로세스 기반 음원분리 (Gaussian Processes for Source Separation: Pseudo-likelihood Maximization)

  • 박선호;최승진
    • 한국정보과학회논문지:소프트웨어및응용
    • /
    • 제35권7호
    • /
    • pp.417-423
    • /
    • 2008
  • 본 논문에서는 각 음원이 시간적 구조를 가졌을 경우 음원들을 분리해내는 확률적 음원분리 방법을 제안한다. 이를 위해 각 음원의 시간적 구조를 가우시안 프로세스(Gaussian process)로 모델링하고 기존의 음원분리 문제를 유사-가능도 최대화 문제(pseudo-likelihood maximization)로 공식화한다. 본 알고리즘을 통해 얻어진 데이타의 유사-가능도는 정규 분포이며 이는 가우시안 프로세스 회귀방법(Gaussian process regression)을 통해 쉽게 계산이 가능하다. 음원분리의 역혼합 행렬은 경도(gradient) 기반최적화 기법을 통해 데이타의 유사-가능도를 최대화하는 해를 찾음으로써 구해진다. 여러 실험을 통하여 제안 알고리듬이 몇 가지 특정 상황에서 기존의 분리 알고리듬들에 비해 우수한 성능을 보임을 확인 할 수 있다.

비디오 기반의 질감 전이 기법 (Texture Transfer Based on Video)

  • 콩푸팔라;이호창;윤경현
    • 한국정보과학회:학술대회논문집
    • /
    • 한국정보과학회 2012년도 한국컴퓨터종합학술대회논문집 Vol.39 No.1(C)
    • /
    • pp.406-407
    • /
    • 2012
  • Texture transfer is a NPR technique for expressing various styles according to source (reference) image. By late 2000s, there are many texture transfer researches. But video base researchers are not active. Moreover, they didn't use important feature like directional information which need to express detail characteristics of target. So, we propose a new method to generate texture transfer animation (using video) with directional effect for maintaining temporal coherence and controlling coherence direction of texture. For maintaining temporal coherence, we use optical flow and confidence map to adapt for occlusion/disocclusion boundaries. And we control direction of texture for taking structure of input. For expressing various texture effects according to different regions, we calculate gradient based on directional weight. With these techniques, our algorithm can make animation result that maintain temporal coherence and express directional texture effect. It is reflect the characteristics of source and target image well. And our result can express various texture directions automatically.

반향환경에서의 가변 적응 상수를 이용한 FDICA와 여러 BSS 알고리즘과의 비교 (Comparison for the variable step-size FDICA with BSS algorithm in reverberant condition)

  • 박근수;박장식;손경식
    • 한국콘텐츠학회:학술대회논문집
    • /
    • 한국콘텐츠학회 2005년도 춘계 종합학술대회 논문집
    • /
    • pp.369-373
    • /
    • 2005
  • 본 논문은 기존의 주파수 영역 독립 성분 분석(FDICA : frequency domain independent component analysis) 방법에 가변 적응 상수를 이용한 알고리즘을 제안한다. 여러 반향 환경에 대한 모의실험으로 기존의 temporal structure(TA)알고리즘, FDICA알고리즘과 비교하여 그 분리 성능을 평가하고 비교한다. 실제 녹음한 혼합 음원에 대한 분리정도를 비교 분석한다.

  • PDF

회전용적형 기어펌프의 유체-구조연동 전산해석 (FLUID-STRUCTURE INTERACTION ANALYSIS OF EXTERNAL GEAR PUMP)

  • 이중호;김태구;이상욱
    • 한국전산유체공학회:학술대회논문집
    • /
    • 한국전산유체공학회 2010년 춘계학술대회논문집
    • /
    • pp.83-85
    • /
    • 2010
  • A hydraulic gear pump is widely used in many industrial applications to provide both high pressure and high flow rate by physical displacement of finite volume of fluid with each revolution. In this study, two dimensional fluid-structure interaction simulation of gear pump flow was carried out to examine detailed complex flow patterns and structural stress distribution on rotors by using a commercial software ADINA. The effect of rotor clearance size on the flow characteristics, specially the temporal variation of velocity and pressure field, which is a main source of flow noise, also was investigated.

  • PDF

미세먼지 농도의 공간적 현황 및 잠재영향인자를 고려한 환경계획적 대응 방향 (Environmental Planning Contermeasures Considering Spatial Distribution and Potential Factors of Particulate Matters Concentration)

  • 성선용
    • 한국환경복원기술학회지
    • /
    • 제23권1호
    • /
    • pp.89-96
    • /
    • 2020
  • Adverse impact of Particulate Matters(PM10, PM2.5; PMs) significantly affects daily lives. Major countermeasures for reducing concentration of PMs were focused on emission source without considering spatial difference of PMs concentration. Thus, this study analyzed spatial·temporal distribution of PMs with observation data as well as potential contributing factors on PMs concentration. The annual average concentration of PMs have been decreased while the particulate matter warnings and alerts were significantly increased in 2018. The average concentration of PMs in spring and winter was higher than the other seasons. Also, the spatial distribution of PMs were also showed seasonality while concentration of PMs were higher in Seoul-metropolitan areas in all seasons. Climate variables, emission source, spatial structure and potential PM sinks were selected major factors which could affects on ambient concentrations of PMs. This paper suggest that countermeasures for mitigating PM concentration should consider characteristics of area. Climatic variables(temperature, pressure, wind speed etc.) affects concentrations of PMs. The effects of spatial structure of cities(terrain, ventilation corridor) and biological sinks(green infrastructure, urban forests) on concentration of PMs should be analyzed in further studies. Also, seasonality of PMs concentration should be considered for establishing effective countermeasures to reduce ambient PMs concentration.

Joint Source/Channel Coding Based on Two-Dimensional Optimization for Scalable H.264/AVC Video

  • Li, Xiao-Feng;Zhou, Ning;Liu, Hong-Sheng
    • ETRI Journal
    • /
    • 제33권2호
    • /
    • pp.155-162
    • /
    • 2011
  • The scalable extension of the H.264/AVC video coding standard (SVC) demonstrates superb adaptability in video communications. Joint source and channel coding (JSCC) has been shown to be very effective for such scalable video consisting of parts of different significance. In this paper, a new JSCC scheme for SVC transmission over packet loss channels is proposed which performs two-dimensional optimization on the quality layers of each frame in a rate-distortion (R-D) sense as well as on the temporal hierarchical structure of frames under dependency constraints. To compute the end-to-end R-D points of a frame, a novel reduced trellis algorithm is developed with a significant reduction of complexity from the existing Viterbi-based algorithm. The R-D points of frames are sorted under the hierarchical dependency constraints and optimal JSCC solution is obtained in terms of the best R-D performance. Experimental results show that our scheme outperforms the existing scheme of [13] with average quality gains of 0.26 dB and 0.22 dB for progressive and non-progressive modes respectively.

Empirical Study of Dynamic Chinese Corporate Governance Based on Chinese-listed Firms with A Panel VAR Approach

  • Shao, Lin;Zhang, Li;Yu, Xiaohong
    • 산경연구논집
    • /
    • 제8권1호
    • /
    • pp.5-13
    • /
    • 2017
  • Purpose - In this article, a dynamic model like a VAR is an appropriate choice for estimating the possible interrelationship between ownership structure and firm performance as a dynamic process. Research design, data, and methodology - Data of this work are collected from Chinese stock exchange including 350 Chinese-listed firms during the period of 1999-2012. We hypothesize that this interrelationship dynamically exists between ownership structure and firm performance. To examine the correlation, a panel Vector Auto-regression (PVAR) approach generated by GMM method is utilized to test the possible dynamic relation embedded in corporate governance. Another two dynamic analysis solutions such as orthogonalized impulse-response function and variance decomposition are also used simultaneously. Results - Findings of this study indicate the evidence that dynamically endogenous relationship exists between ownership structure and firm performance. Further, there is a dynamical correlation between investment and performance. Impulse response and variance decomposition illustrate that impact of a shock to variables themselves is the main source for their variability. Conclusions - The conclusion in this study is that there is a bidirectional and inter-temporal effect between proportion of ownership and corporate performance for a long run in accordance with impulse response function. Overall, our results suggest that corporate governance in China is more market oriented.

복합형 유역모델 STREAM의 개발(I): 모델 구조 및 이론 (Development of a Hybrid Watershed Model STREAM: Model Structures and Theories)

  • 조홍래;정의상;구본경
    • 한국물환경학회지
    • /
    • 제31권5호
    • /
    • pp.491-506
    • /
    • 2015
  • Distributed models represent watersheds using a network of numerous, uniform calculation units to provide spatially detailed and consistent evaluations across the watershed. However, these models have a disadvantage in general requiring a high computing cost. Semi-distributed models, on the other hand, delineate watersheds using a simplified network of non-uniform calculation units requiring a much lower computing cost than distributed models. Employing a simplified network of non-uniform units, however, semi-distributed models cannot but have limitations in spatially-consistent simulations of hydrogeochemical processes and are often not favoured for such a task as identifying critical source areas within a watershed. Aiming to overcome these shortcomings of both groups of models, a hybrid watershed model STREAM (Spatio-Temporal River-basin Ecohydrology Analysis Model) was developed in this study. Like a distributed model, STREAM divides a watershed into square grid cells of a same size each of which may have a different set of hydrogeochemical parameters reflecting the spatial heterogeneity. Like many semi-distributed models, STREAM groups individual cells of similar hydrogeochemical properties into representative cells for which real computations of the model are carried out. With this hybrid structure, STREAM requires a relatively small computational cost although it still keeps the critical advantage of distributed models.

단일센서와 시간역전법을 이용한 판에서의 충격위치 결정에 관한 연구 (Determination of Impact Source Location Using a Single Transducer and Time Reversal Technique)

  • 정현조;조성종
    • 비파괴검사학회지
    • /
    • 제32권1호
    • /
    • pp.47-55
    • /
    • 2012
  • 본 논문에서는 단일 센서와 시간역전 원리를 이용하여 간단한 판에서 충격 위치를 결정할 수 있는 탐상법을 다루었다. 수치적인 시뮬레이션과 실험을 통하여 시간역전에 의한 충격위치와 그 주변에서 신호의 집속효과를 관찰하고, 충격위치 결정에 영향을 미칠 수 있는 인자들(가진력의 크기, 신호의 기록시간)에 대해 살펴보았다. 이러한 결과를 기초로 두 가지 다른 충격위치에 대한 실험을 수행하고 그 결과를 가시화하였으며, 실제 충격위치를 정확하게 결정할 수 있음을 확인하였다. 여기서 제안한 방법의 특징은 단일센서를 사용하는 것과 시험체의 형상과 물성을 몰라도 된다는 점이다. 또한 판에서와 같이 분산성의 다중모드 파동이 발생하는 경우에도 특정 모드나 주파수에 의존할 필요가 없다. 따라서 기존의 충격위치 결정법에 비해 많은 장점을 지니고 있다.

MPEG-2의 GOP 구조를 이용한 I 프레임의 시공간적 오류 은닉 (Spatio-Temporal Error Concealment of I-frame using GOP structure of MPEG-2)

  • 강민정;류철
    • 한국통신학회논문지
    • /
    • 제29권1C호
    • /
    • pp.72-82
    • /
    • 2004
  • 본 논문은 MPEG-2 인트라 (I) 프레임의 더욱 강화된 오류 은닉 기술을 제안한다. MPEG-2 소스 부호화 알고리즘은 가변장 부호화를 사용하기 때문에 채널 오류에 매우 민감하다. 채널에서 발생된 전송 오류는 오류 정정 기술을 사용하여 보정되지만 제대로 보정되지 않은 오류는 디코더에서 오류 은닉 기술을 사용함으로써 시각적 왜곡이 최소화될 수 있다. 또한 GOP의 시작 프레임인 I 프레임의 오류는 움직임 보상 예측 부호화 방식으로 인하여 다른 인터 프레임에 전달되어 더욱 심각한 화질 저하를 가져온다. 본 논문에서 제안하는 오류 은닉 방식은 I 프레임에 발생된 연속적인 슬라이스 오류를 인접한 인터 프레임의 시공간적 정보를 이용하여 기존 방식보다 효과적으로 처리하는 방식이다. 이 방식은 기존의 방식들이 갖는 시간적 그리고 공간적에서 발생하는 단점을 효율적으로 개선할 수 있으며, 전송 오류가 심한 망에서 발생되는 극심한 슬라이스 오류에서도 좋은 시각적 효과를 얻을수 있다. 알고리즘은 MPEG-2 비디오 코덱 범위에서 수행되었으며 모의실험을 통하여 제안된 방식이 다른 방식들보다 높은 PSNR과 개선된 시각적 효과를 확인할 수 있었다.