• Bulletin of the Korean Mathematical Society
• /
• v.51 no.2
• /
• pp.595-612
• /
• 2014

The spectral collocation method for a second order elliptic boundary value problem on a domain ${\Omega}$ with curved boundaries is studied using the Gordon and Hall transformation which enables us to have a transformed elliptic problem and a square domain S = [0, h] ${\times}$ [0, h], h > 0. The preconditioned system of the spectral collocation approximation based on Legendre-Gauss-Lobatto points by the matrix based on piecewise bilinear finite element discretizations is shown to have the high order accuracy of convergence and the efficiency of the finite element preconditioner.

• Proceedings of the KSPS conference
• /
• 2005.04a
• /
• pp.53-56
• /
• 2005

In this paper, we propose a transformation based robust adaptation technique that uses the maximum mutual information(MMI) estimation for the objective function and the linear spectral transformation(LST) for adaptation. LST is an adaptation method that deals with environmental noises in the linear spectral domain, so that a small number of parameters can be used for fast adaptation. The proposed technique is called MMI-LST, and evaluated on TIMIT and FFMTIMIT corpora to show that it is advantageous when only a small amount of adaptation speech is used.

• Phonetics and Speech Sciences
• /
• v.9 no.1
• /
• pp.41-47
• /
• 2017

The main advantage of the statistical parametric speech synthesis is its flexibility in changing voice characteristics. A personalized text-to-speech(TTS) system can be implemented by combining a speech synthesis system and a voice transformation system, and it is widely used in many application areas. It is known that the fundamental frequency and the spectral envelope of speech signal can be independently modified to convert the voice characteristics. Also it is important to maintain naturalness of the transformed speech. In this paper, a speech synthesis system based on Hidden Markov Model(HMM-based speech synthesis, HTS) using the STRAIGHT vocoder is constructed and voice transformation is conducted by modifying the fundamental frequency and spectral envelope. The fundamental frequency is transformed in a scaling method, and the spectral envelope is transformed through frequency warping method to control the speaker's vocal tract length. In particular, this study proposes a voice transformation method using the correlation between fundamental frequency and vocal tract length. Subjective evaluations were conducted to assess preference and mean opinion scores(MOS) for naturalness of synthetic speech. Experimental results showed that the proposed voice transformation method achieved higher preference than baseline systems while maintaining the naturalness of the speech quality.

• Proceedings of the KSPS conference
• /
• 2006.05a
• /
• pp.131-134
• /
• 2006

The narrowband speech over the telephone network is lacking in the information from low-band (0-300 Hz) and high-band (3400-8000 Hz) that are found in wideband speech (0-8000 Hz). As a result, narrowband speech is characterized by the reduced intelligibility and muffled quality, and degraded speaker identification. Spectral folding is the easiest way to reconstruct the missing high-band; however, the reconstructed speech still brings the sense of band-limited characteristic because of the absence of low-band and mid-band frequency components. To compensate for the lack of the extended speech, we propose to combine the spectral folding method and GMM transformation method, which is a statistical method to reconstruct wideband speech. The reconstructed wideband speech showed that the absent frequency components was filled up with relatively low spectral mismatch. According to the subjective speech quality evaluations, the proposed method was preferred to other methods.

• 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.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.17 no.3
• /
• pp.311-320
• /
• 1999

This study conducts to analyze comparative accuracy when both Image Transformation Methods and Supervised Classifications on multi-spectral TM using a case of Lower Tumen River Area. In terms of overall classification accuracy, maximum likelihood method turns out higher than other one, but in a case of vegetation only, MNF and TC image transformation methods produce a better quality of the result. Especially, seven dimensional images including MNF, TC, and NDVI create better image than three dimensional one. Among these transformation methods, maximum likelihood method results out the best one. Multi-spectral image could be useful as an important basic material for site selection of industrial allocation as well as Tumen River Area Economic Development Plan.

• Korean Journal of Optics and Photonics
• /
• v.14 no.3
• /
• pp.266-271
• /
• 2003

The reflectance spectrum of optical films thicker than a few microns shows an intensity oscillation due to interference. Since the spectral period of the oscillation is inversely related to film thickness, the thickness of an optical film can be determined from the spectral frequency of the oscillation. For rapid data processing, the spectral frequency is obtained by use of a Fast Fourier Transformation technique. The conventional method of applying a Fast Fourier Transformation to the reflectance spectrum versus photon energy is modified so as to clear the ambiguity in choosing the proper effective refractive index value and to prevent the broadening of the Fourier transformed peak due to the refractive index dispersion. This technique of modified Fast Fourier Transformation is suggested by the authors for the first time to their knowledge. From the analysis of the calculated reflectance spectrum of a 30-${\mu}{\textrm}{m}$-thick dielectric film. it is shown to improve the accuracy in determining film thickness by a great amount. The improved accuracy of the modified Fast Fourier Transformation is also confirmed from the analysis of the reflectance spectra of a sample with 80-${\mu}{\textrm}{m}$-thick cover layer and 13-${\mu}{\textrm}{m}$-thick spacer layer on a PC substrate.

• Proceedings of the KSPS conference
• /
• 2005.11a
• /
• pp.67-70
• /
• 2005

Voice conversion (VC) is a technique for modifying the speech signal of a source speaker so that it sounds as if it is spoken by a target speaker. Most previous VC approaches used a linear transformation function based on GMM to convert the source spectral envelope to the target spectral envelope. In this paper, we propose several nonlinear GMM-based transformation functions in an attempt to deal with the over-smoothing effect of linear transformation. In order to obtain high-quality modifications of speech signals our VC system is implemented using the Harmonic plus Noise Model (HNM)analysis/synthesis framework. Experimental results are reported on the English corpus, MOCHA-TlMlT.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.776-782
• /
• 2003

This paper develops an approach to fuse 1-meter resolution spatial panchromatic and 4-meter resolution multi-spectral IKONOS images. The approach is based on the characteristics of four-band wavelet transformation. The experiment shows that the fused images based on four-band wavelet method contain with not only high spatial resolution but also rich spectral characteristic.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.322-324
• /
• 2003

Leaf C:N ratio is a new factor in the field of biochemical inversion with hyperspectral data. Effects of common-used spectral transformations including log(R), log(1/R), 1/R, etc. from 400nm to 2490nm on its inversion are compared. Results show that their effects on statistical modeling are not apparent. Continuum removal is used on original reflectance in the range of 2030nm to 2220nm, in which exists an apparent absorption peak due to cellulose, lignin, protein, etc. The effect is distinctive and tends to improve the precision of C:N ratio inversion. Further, it is a robust and physically based transformation.