• 방송공학회논문지
• /
• 제20권1호
• /
• pp.164-170
• /
• 2015

본 논문에서는 잡음 환경에서 취득된 음성 신호에서 잡음을 제거하기 위한 방법으로 사용되는 이진 마스크 분류 모델의 적응과정에 대해 다루고자 한다. 기존 연구결과에 의하면, 잡음 환경 데이터에 이진 마스크 기법을 적용하면 음성 명료도를 향상시킬 수 있다고 알려져 있다. 하지만 이진 마스크 분류 모델 학습 시 테스트 환경 데이터가 포함되어야 한다는 단점을 안고 있다. 본 논문에서는 새로운 잡음 환경에서 이진 마스크 분류 모델을 적응하기 위해, 음성 인식에서 널리 사용되는 화자 적응 기법인 eigenvoice 방법을 적용하고자 한다. 실험결과에서는 모델 적응에 사용되는 데이터량에 따른 성능을 정검출율과 오검출율 관점에서 평가하였고, 그 결과 새로운 잡음 환경에서 데이터량을 증가시켜 모델을 적응함으로써 향상된 성능을 나타냄을 확인할 수 있었다.

• 한국컴퓨터그래픽스학회논문지
• /
• 제27권5호
• /
• pp.45-54
• /
• 2021

최근 노이즈 제거를 위한 심층 학습 모델에 대한 연구가 활발하게 진행되고 있다. 특히 블라인드 노이즈 제거 (blind denoising) 기술이 발전하면서 깨끗한 영상을 얻기가 불가능한 영상의 영역에서 노이즈 영상만으로 심층 학습 기반 노이즈 제거 모델의 학습이 가능해졌다. 우리는 관찰된 노이즈 영상으로부터 깨끗한 영상을 얻기 위해 더는 깨끗한 영상과 노이즈 영상의 짝을 이루는 데이터를 필요하지 않는다. 하지만 노이즈 영상과 깨끗한 영상 간의 차이가 큰 데이터라면 노이즈 영상만으로 학습된 노이즈 제거 모델은 우리가 원하는 품질의 깨끗한 영상을 복원하기 어려울 것이다. 이 문제를 해결하기 위해서 짝지어지지 않는 깨끗한 영상과 노이즈 영상으로 학습한 모델 기반 노이즈 제거 기술은 최근 연구되고 있다. 가장 최신 기술인 ISCL은 깨끗한 영상과 노이즈 영상의 쌍을 기반으로 한 지도학습 기반 모델의 성능과 거의 근접한 성능을 보여 주었다. 우리는 제안된 방법이 ISCL을 포함한 다른 최신 짝을 이루지 않는 영상 기반 노이즈 제거 기술보다 성능이 우수함을 보여준다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2000년도 하계학술대회 논문집 D
• /
• pp.3166-3168
• /
• 2000

This paper proposed variable threshold dual rate ADPCM coding method which is modified from the standard ADPCM of ITU G.726 for speech quality improvement. The speech quality of variable threshold dual rate ADPCM is better than single rate ADPCM at noisy environment without increasing the complexity by using ZCR(Zero Crossing Rate). In this case, ZCR is used to divide input signal samples into two categories(noisy & speech). The samples with higher ZCR is categorized as the noisy region and the samples with lower ZCR is categorized as the speech region. Noisy region uses higher threshold value to be compressed by 16Kbps for reduced bit rates and the speech region uses lower threshold value to be compressed by 40Kbps for improved speech quality. Comparing with the conventional ADPCM, which adapts the fixed coding rate. the proposed variable threshold dual rate ADPCM coding method improves noise character without increasing the bit rate. For real time applications, ZCR calculation was considered as a simple method to obtain the background noise information for preprocess of speech analysis such as FFT and the experiment showed that the simple calculation of ZCR can be used without complexity increase. Dual rate ADPCM can decrease the amount of transferred data efficiently without increasing complexity nor reducing speech quality. Therefore result of this paper can be applied for real-time speech application such as the internet phone or VoIP.

• 한국데이터정보과학회:학술대회논문집
• /
• 한국데이터정보과학회 2004년도 춘계학술대회
• /
• pp.197-201
• /
• 2004

We describe a new edge detector based on the robust rank-order (RRO) test which is a useful alternative to Wilcoxon test, using $r{\times}r$ window for detecting edges of all possible orientations in noisy images. Some experiments of statistical edge detectors based on the Wilcoxon test and T test with our RRO detector are carried out on synthetic and real images corrupted by both Gaussian and impulse noise. We also implement these edge detectors using Java on the Web.

• Smart Structures and Systems
• /
• 제19권4호
• /
• pp.383-391
• /
• 2017

In this paper, we investigate the impacts of network topology on the performance of a distributed estimation algorithm, namely combine-then-adaptive (CTA) diffusion LMS, based on the data with or without the assumptions of temporal and spatial independence with noisy links. The study covers different network models, including the regular, small-world, random and scale-free whose the performance is analyzed according to the mean stability, mean-square errors, communication cost (link density) and robustness. Simulation results show that the noisy links do not cause divergence in the networks. Also, among the networks, the scale free network (heterogeneous) has the best performance in the steady state of the mean square deviation (MSD) while the regular is the worst case. The robustness of the networks against the issues like node failure and noisier node conditions is discussed as well as providing some guidelines on the design of a network in real condition such that the qualities of estimations are optimized.

• Journal of Mechanical Science and Technology
• /
• 제15권3호
• /
• pp.300-308
• /
• 2001

Tracking of moving objects within video streams is a complex and time-consuming process. Large number of moving objects increases the time for computation of tracking the moving objects. Because of large computations, there are real-time processing problems in tracking of moving objects. Also, the change of environment causes errors in estimation of tracking information. In this paper, we present a new method for tracking of moving objects using optical flow motion analysis. Optical flow represents an important family of visual information processing techniques in computer vision. Segmenting an optical flow field into coherent motion groups and estimating each underlying motion are very challenging tasks when the optical flow field is projected from a scene of several moving objects independently. The problem is further complicated if the optical flow data are noisy and partially incorrect. Optical flow estimation based on regulation method is an iterative method, which is very sensitive to the noisy data. So we used the Combinatorial Hough Transform (CHT) and Voting Accumulation for finding the optimal constraint lines. To decrease the operation time, we used logical operations. Optical flow vectors of moving objects are extracted, and the moving information of objects is computed from the extracted optical flow vectors. The simulation results on the noisy test images show that the proposed method finds better flow vectors and more correctly estimates the moving information of objects in the real time video streams.

• The Journal of the Acoustical Society of Korea
• /
• 제15권3E호
• /
• pp.4-12
• /
• 1996

In this study, we examine the applicability of an artificial neural network(ANN) for filtering underwater random noise and for identifying underlying signals taken from noisy environment. The approach is to find a way of compressing the input data and then decompressing it using an ANN as in image compressing process. It is well known that random signal is hard to compress while ordered information is not. The use of a limited number of processing elements(PEs) in the hidden layer of an Ann ensures that some of the noise would be removed in the reconstruction process. Two types of the signals, synthesized and measured, are used to examine the effectiveness of the ANN-based filter. After training process is completed, the ANN successfully extracts the underlying signals form the synthesized or measured noisy signals. In particular, compared with the results form without filtering or moving averaged, the ANN-based filter gives much better spectrograms to identify underlying signals from the measured noisy data. This filtering process is achieved without using and kind of highly accurate signal processing technique. More experimentation needs to be followed to develop the ANN-based filtering technique to the level of complete understanding.

• 한국음향학회지
• /
• 제29권2호
• /
• pp.97-101
• /
• 2010

일반적으로 시스템 인식 방법은 입출력에 잡음이 없거나, 출력에만 잡음이 있는 경우를 주 대상으로 한다. 본 논문은 입력 및 출력이 모두 잡음으로 오염되었을 뿐만 아니라 입력에 비해서 출력에 같거나 더 많은 양의 잡음이 개입된 환경에 노출된 Finite Impulse Response 형태의 시스템을 인식하는 새로운 방법을 제안한다. 이를 위해서 입출력의 잡음 수준이 같을 때 최적인 완전최소자승 기법과 출력에만 잡음이 있을 때 최적인 최소자승 기법을 서로 볼록 결합 (convex combination)하여 앞에서 언급한 것과 같은 좀 더 일반화된 잡음 환경에서도 향상된 결과가 나오도록 하였다. 또 제안한 방법이 다양한 잡음 환경에서 응용 가능함을 모의 실험을 통해서 확인하였다.

• Earthquakes and Structures
• /
• 제15권3호
• /
• pp.263-273
• /
• 2018

Operational modal analysis is being widely used in aerospace, mechanical and civil engineering. Common research fields include optimal design and rehabilitation under dynamic loads, structural health monitoring, modification and control of dynamic response and analytical model updating. In many practical cases, influence of noise contamination in the recorded data makes it difficult to identify the modal parameters accurately. In this paper, an improved frequency domain method called Enhanced Least Square Complex Frequency (eLSCF) is developed to extract modal parameters from noisy recorded data. The proposed method makes the use of pre-defined approximate mode shape vectors to refine the cross-power spectral density matrix and extract fundamental frequency for the mode of interest. The efficiency of the proposed method is illustrated using an example five story shear frame loaded by random excitation and different noise signals.

• 대한전기학회논문지
• /
• 제39권8호
• /
• pp.868-874
• /
• 1990

A singular value approach is presented in order to determine the number of multiple sinusoids from the finite noisy data. Simulations are conducted for Akaike's information criterion (AIC), Rissanen's shortest data description (MDL) and a singular value approach, for various examples with different SNR's and methods of estimating frequencies. And then the performances are compared. Simulation results that the singular value approach is superior to AIC and MDL for FBLP, HOYW and covariance matrix based methods are investigated. The approach with contribute to the frequency estimation of multiple sinusoids from the finite noisy data. Furthermore, this will be applied to the DSPs of communication and bio-medical engineering.