• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.5 no.3
• /
• pp.443-449
• /
• 2001

This paper is simulation research to improve speech recognition rates under the noisy environment. We examines recognition ratio based on frequency-weighted HMM together with spectral subtraction. As results, frequency-weighted HMM with scaling coefficients is trained as a minimum error classification criterion, and is presents a higher recognition rates in noisy condition than a conventional method. Furthermore, spectral subtraction method gives 11 to 28% improvements for this frequency-weighted HMM in low SNR, and gives recognition rates of 81.7% at 6dB SNR of noisy speech.

• ETRI Journal
• /
• v.31 no.4
• /
• pp.454-456
• /
• 2009

The maximum likelihood linear spectral transformation (ML-LST) using a numerical iteration method has been previously proposed for robust speech recognition. The numerical iteration method is not appropriate for real-time applications due to its computational complexity. In order to reduce the computational cost, the objective function of the ML-LST is approximated and a closed-form solution is proposed in this paper. It is shown experimentally that the proposed closed-form solution for the ML-LST can provide rapid speaker and environment adaptation for robust speech recognition.

• Proceedings of the KSRS Conference
• /
• 2008.10a
• /
• pp.111-114
• /
• 2008

The existence of noise bands may deform the typical shape of the spectrum, making the accuracy of clustering degraded. This paper proposes a statistical approach to remove noise bands in hyperspectral data using the correlation coefficient of bands as an indicator. Considering each band as a random variable, two adjacent signal bands in hyperspectral data are highly correlative. On the contrary, existence of a noise band will produce a low correlation. For clustering, the unsupervised ${\kappa}$-nearest neighbor clustering method is implemented in accordance with three well-accepted spectral matching measures, namely ED, SAM and SID. Furthermore, this paper proposes a hierarchical scheme of combining those measures. Finally, a separability assessment based on the between-class and the within-class scatter matrices is followed to evaluate the applicability of the proposed noise band removal method. Also, the paper brings out a comparison for spectral matching measures.

• The Journal of the Acoustical Society of Korea
• /
• v.17 no.4E
• /
• pp.35-40
• /
• 1998

In this paper, an efficient quantization method of line spectrum pairs(LSP) with cascaded structure of vector quantizer and scalar quantizer is proposed. First, input LSP parameters is vector-quantized using a codebook a with a moderate number of entries. In the second stage of quantization, the components of residual vector are individually quantized by the scalar quantizer. The utilization of ordering property of LSP parameters and the inclusion of interframe prediction improve the quantizer performance and remove the stability check routine after quantization procedure. The new vector-scalar hybrid quantizer using 26 bits/frame shows a transparent quality of speech that an average spectral distortion is 1 dB and the frame proportion with above 2 dB spectral distortion is less than 2%. The performances of proposed quantization method is evaluated in the transmission errors.

• Journal of The Korean Astronomical Society
• /
• v.29 no.2
• /
• pp.157-170
• /
• 1996

In order to determine the metallicity of a globuar cluster, M3,by using the spectral indices, a kind of index grid has been establshed by stars in globular clusters, M3, M15, M71 and old open cluster, NGC 188. The indices were measured from the medium resolution spectra of about $2{\AA}$. The summed indices were used to determine metallicity in order to increase signals. It is found that the core depth index is measured more accurately and leads result more accurate than the pseudo-equivalent width index. This method can be further improved by including many more calibration globular clusters of various metallicity to make finer grids. By this method, the metallicity of M3 is determined as $[Fe/H] = -1.46\pm0.15$.

• Journal of the Optical Society of Korea
• /
• v.15 no.2
• /
• pp.111-117
• /
• 2011

Full-range spectral domain optical coherence tomography (SD-OCT) with a 1-${\mu}m$ band light source is shown here. The phase of the reference beam is continuously stepped while the probing beam scans the sample laterally (B-scan). The two dimensional spectral interferogram obtained is processed by a Fourier transform method to obtain a complex spectrum leading to a full-range OCT image. A detailed mathematical explanation of the complex conjugate resolving method utilized is provided. The system's measurement speed was 7.96 kHz, the measured axial resolution was $9.6{\mu}m$ in air and the maximum sensitivity 99.4 dB. To demonstrate the effect of mirror image elimination, In vivo human eye pathology was measured.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.48 no.11
• /
• pp.83-88
• /
• 2011

The spectral field data on the observation planes inside a mode stirred chamber have been computed by using the FDTD method. They can describe the directional dependence of all supported modes in the reverberation chamber (RC). Correlation coefficients have been found to evaluate the stirring effects. The correlation coefficients were obtained and compared for three different sizes of stirrers and statistical uniformity was investigated.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.25-28
• /
• 2002

Voice activity detection is important Problem in the speech recognition and communication. This paper introduces feature parameter which is reconstructed by the spectral entropy of information theory for the robust voice activity detection in the noise environment, analyzes and compares it with the energy method of voice activity detection and performance. In experiment, we confirmed that the spectral entropy is more feature parameter than the energy method for the robust voice activity detection in the various noise environment.

• Journal of the Korean Graphic Arts Communication Society
• /
• v.18 no.2
• /
• pp.1-12
• /
• 2000

This paper proposes a new algorithm classifing same hues in order toe estimate the spectral reflectance of object from 3 band color image information. To estimate the spectral reflectance of object, the conventional estimation methods are required of 5 or 9 band digital color values. The 5 or 9 band image acquisition systems are required of 5 or 3 times same work for color image acquisition process. To solve the above problems, we propose a new method that can be estimated spectra reflectance estimation of object. The proposed method is to classify same hues corresponding a color stimulus, by using hue angle and chroma vector of a color stimulus. The classified same hues are used as the population corresponding a color stimulus. The range of same hue is estimated by the cumulative proportional ration according to the number of basis function.

• Proceedings of the KAIS Fall Conference
• /
• 2010.05a
• /
• pp.265-268
• /
• 2010

There have been proposed two types of low bit rate vocoder upto now : One is MBE type using the spectrum modeling and another is CELP type using the hybrid coding method. CELP type vocoder has mainly studied between them. Specially, much of intensity is concentrated in CELP vocoder due to the emergence of Internet Phone and PCS in a domestic. In order to improve the speech quality in CELP vocoder, in this paper, we proposed a new spectrum analysis algorithm with variable window. In CELP vocoder, the spectrum of the synthesised speech signal is distorted because the fixed size windows is used for spectrum analysis. So we have measured the spectral leakage and in order to minimize the spectral leakage have adjusted the window size. Applying this method G.723.1 ACELP, we can get SD(Spectral Distortion) reduction 0.084(dB), residual energy reduction 6.3% and MOS(Mean Opinion Score) improvement 0.1.