• Journal of KIISE:Software and Applications
• /
• v.27 no.5
• /
• pp.583-592
• /
• 2000

This paper proposes feature parameter transformation method using independent component analysis (ICA) for speaker identification. The proposed method assumes that the cepstral vectors from various channel-conditioned speech are constructed by a linear combination of some characteristic functions with random channel noise added, and transforms them into new vectors using ICA. The resultant vector space can give emphasis to the repetitive speaker information and suppress the random channel distortions. Experimental results show that the transformation method is effective for the improvement of speaker identification system.

• Speech Sciences
• /
• v.12 no.3
• /
• pp.79-89
• /
• 2005

Acoustic feedback between microphone and receiver can be effectively cancelled adaptive feedback cancellation algorithm. Although many speech sounds have non-Gaussian distribution, most algorithms were tested with speech like sounds whose distribution were Guassian type. In this paper, we proposed an adaptive feedback cancellation algorithm based on independent component analysis (ICA) for digital hearing aid. The algorithm was tested with not only Gaussian distribution but also Laplacian distribution. We verified that the proposed algorithm has better acoustic feedback cancelling performance than conventional normalized root mean square (NLMS) algorithm, especially speech like sounds with Laplacian distribution.

• Journal of Korea Multimedia Society
• /
• v.8 no.3
• /
• pp.328-335
• /
• 2005

This paper proposes a frequency domain independent component analysis (ICA) algorithm to separate the mixed speech signals using a multiple microphone array By estimating the delay timings using a input cross-correlation, even in the delayed mixture case, we propose a good initial value setting method which leads to optimal convergence. To reduce the calculation, separation process is performed at frequency domain. The results of simulations confirms the better performances of the proposed algorithm.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.9-14
• /
• 2002

We propose a new algorithm for blind source separation (BSS), in which frequency-domain independent component analysis (FDICA) and time-domain ICA (TDICA) are combined to achieve a superior source-separation performance under reverberant conditions. Generally speaking, conventional TDICA fails to separate source signals under heavily reverberant conditions because of the low convergence in the iterative learning of the inverse of the mixing system. On the other hand, the separation performance of conventional FDICA also degrades significantly because the independence assumption of narrow-band signals collapses when the number of subbands increases. In the proposed method, the separated signals of FDICA are regarded as the input signals for TDICA, and we can remove the residual crosstalk components of FDICA by using TDICA. The experimental results obtained under the reverberant condition reveal that the separation performance of the proposed method is superior to that of conventional ICA-based BSS methods.

• Proceedings of the KSPS conference
• /
• 2003.05a
• /
• pp.123-126
• /
• 2003

We discuss how to reduce the number of inverse matrix and its dimensions requested in MLLR framework for speaker adaptation. To find a smaller set of variables with less redundancy, we employ PCA(principal component analysis) and ICA(independent component analysis) that would give as good a representation as possible. The amount of additional computation when PCA or ICA is applied is as small as it can be disregarded. The dimension of HMM parameters is reduced to about 1/3 ~ 2/7 dimensions of SI(speaker independent) model parameter with which speech recognition system represents word recognition rate as much as ordinary MLLR framework. If dimension of SI model parameter is n, the amount of computation of inverse matrix in MLLR is proportioned to O($n^4$). So, compared with ordinary MLLR, the amount of total computation requested in speaker adaptation is reduced to about 1/80~1/150.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.4E
• /
• pp.156-163
• /
• 2002

We apply independent component analysis (ICA) for extracting an optimal basis to the problem of finding efficient features for representing speech signals of a given speaker The speech segments are assumed to be generated by a linear combination of the basis functions, thus the distribution of speech segments of a speaker is modeled by adapting the basis functions so that each source component is statistically independent. The learned basis functions are oriented and localized in both space and frequency, bearing a resemblance to Gabor wavelets. These features are speaker dependent characteristics and to assess their efficiency we performed speaker identification experiments and compared our results with the conventional Fourier-basis. Our results show that the proposed method is more efficient than the conventional Fourier-based features in that they can obtain a higher speaker identification rate.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.4
• /
• pp.156-156
• /
• 2002

We apply independent component analysis (ICA) for extracting an optimal basis to the problem of finding efficient features for representing speech signals of a given speaker The speech segments are assumed to be generated by a linear combination of the basis functions, thus the distribution of speech segments of a speaker is modeled by adapting the basis functions so that each source component is statistically independent. The learned basis functions are oriented and localized in both space and frequency, bearing a resemblance to Gabor wavelets. These features are speaker dependent characteristics and to assess their efficiency we performed speaker identification experiments and compared our results with the conventional Fourier-basis. Our results show that the proposed method is more efficient than the conventional Fourier-based features in that they can obtain a higher speaker identification rate.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2003.06a
• /
• pp.78-81
• /
• 2003

본 논문은 SVM(support vector machine)을 이용한 음성인식기에 대해 효과적인 특징 파라메터를 제안한다. SVM은 특징 공간에서 비선형 경계를 찾아 분류하는 방법으로 적은 학습 데이터에서도 좋은 분류 성능을 나타낸다고 알려져 있으며 최적의 특징 파라메터를 선택하기 위해 본 논문에서는 SVM을 이용한 음성인식기를 사용하여 PCA(principal component analysis), ICA(independent component analysis) 알고리즘을 적용하여 MFCC(met frequency cepstrum coefficient)의 특징 공간을 변화시키면서 각각의 인식 성능을 비교 검토하였다. 실험 결과 ICA에 의한 특징 파라메터가 가장 우수한 성능을 나타내었으며 특징 공간에서 각 클래스의 분포도 또한 ICA가 가장 높은 선형 분별성을 나타내었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04c
• /
• pp.292-294
• /
• 2003

본 논문에서는 독립성분분석(Independent Component Analysis； 이하 ICA)에서의 제안된 그레이 블록 거리 알고리즘(Grey Block Algorithm, 이하 GBD)을 이용한 영상 분류 방법을 제안한다. 이 제안된 방법은 기존의 GBD 알고리듬을 이용한 경우보다 k가 감소할 때 그 편차는 적어 좋은 영상 분류 특징을 보임을 모의 실험을 통하여 확인할 수 있었다.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2674-2676
• /
• 2002

본 논문에서는 독립성분분석(independent component analysis; 이하 ICA)에서의 새로운 그레이 블록 거리(grey block distance; GBD, 이하 GBD)알고리즘을 이용한 영상 분류 방법을 제안한다. 이 제시된 방법은 다중해상도에서 기존의 GBD 알고리즘과 비교하여 이차원 영상간의 상대적 식별을 더 용이하게 하여 영상이 급격히 변화하는 부분의 정보를 잃지 않게 개선할 수 있었다. 모의 실험 결과로부터 기존의 GBD 알고리즘에 비하여 영상간의 상대적 식별이 더 용이하여 빨리 수렴이 되는 것을 모의 실험을 통하여 확인하였다.