• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.35 no.5
• /
• pp.419-425
• /
• 2007

A rapid one-shot alignment method of the Strapdown INS (SDINS) for a vertical launch is proposed. The proposed alignment is performed using the accelerometer output of the Slave INS and the attitude of the Master INS. To improve the accuracy and the speed of the alignment, the equivalent linear transformation and the pre-filtering method are developed. Experiment results show that the proposed method is effective in improving the accuracy and the speed of the alignment.

• 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.

• Speech Sciences
• /
• v.11 no.4
• /
• pp.75-88
• /
• 2004

In this paper, we propose a new approach to sequential linear regression adaptation of continuous density hidden Markov models (CDHMMs) based on transformation space model (TSM). The proposed TSM which characterizes the a priori knowledge of the training speakers associated with maximum likelihood linear regression (MLLR) matrix parameters is effectively described in terms of the latent variable models. The TSM provides various sources of information such as the correlation information, the prior distribution, and the prior knowledge of the regression parameters that are very useful for rapid adaptation. The quasi-Bayes (QB) estimation algorithm is formulated to incrementally update the hyperparameters of the TSM and regression matrices simultaneously. Experimental results showed that the proposed TSM approach is better than that of the conventional quasi-Bayes linear regression (QBLR) algorithm for a small amount of adaptation data.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.4
• /
• pp.843-847
• /
• 2011

In this note, a proof of Utkin's theorem is presented for the SI(Single Input) uncertain integral linear case. The invariance theorem with respect to the two transformation methods so called the two diagonalization methods are proved clearly and comparatively for SI uncertain integral linear systems. With respect to the sliding surface transformation, the equation of the sliding mode, the sliding surface is invariant. Both the applied control inputs have the same gains. By means of the two transformation methods the same results can be obtained. Through an illustrative example and simulation study, the usefulness of the main results is verified.

• ETRI Journal
• /
• v.41 no.1
• /
• pp.109-116
• /
• 2019

This paper explores new techniques that are based on a hidden-layer linear transformation for fast speaker adaptation used in deep neural networks (DNNs). Conventional methods using affine transformations are ineffective because they require a relatively large number of parameters to perform. Meanwhile, methods that employ singular-value decomposition (SVD) are utilized because they are effective at reducing adaptive parameters. However, a matrix decomposition is computationally expensive when using online services. We propose the use of an extended diagonal linear transformation method to minimize adaptation parameters without SVD to increase the performance level for tasks that require smaller degrees of adaptation. In Korean large vocabulary continuous speech recognition (LVCSR) tasks, the proposed method shows significant improvements with error-reduction rates of 8.4% and 17.1% in five and 50 conversational sentence adaptations, respectively. Compared with the adaptation methods using SVD, there is an increased recognition performance with fewer parameters.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.11 no.1
• /
• pp.597-605
• /
• 2019

This paper proposes a novel method for efficient prediction of joint distributions of heights and periods of nonlinear ocean waves. The proposed novel method utilizes a transformed linear simulation which is based on a Hermite transformation model where the transformation is chosen to be a monotonic cubic polynomial, calibrated such that the first four moments of the transformed model match the moments of the true process. This proposed novel method is utilized to predict the joint distributions of wave heights and periods of a sea state with the surface elevation data measured at the Gulfaks C platform in the North Sea, and the novel method's accuracy and efficiency are favorably validated by using comparisons with the results from an empirical joint distribution model, from a linear simulation model and from a second-order nonlinear simulation model.

• Journal of applied mathematics & informatics
• /
• v.20 no.1_2
• /
• pp.535-540
• /
• 2006

In this paper, characterizations of difference quotient transformation in the Krein space which is contained continuously and contractively in the krein space of square summable power series C (z) is obtained from the complementation theory.

• Journal of applied mathematics & informatics
• /
• v.33 no.1_2
• /
• pp.229-234
• /
• 2015

In the setting of semidefinite linear complementarity problems on $S^n$, we focus on the Stein Transformation $S_A(X):=X-AXA^T$ for $A{\in}R^{n{\times}n}$ that is positive semidefinite preserving (i.e., $S_A(S^n_+){\subseteq}S^n_+$) and show that such transformation is strictly monotone if and only if it is nondegenerate. We also show that a positive semidefinite preserving $S_A$ has the Ultra-GUS property if and only if $1{\not\in}{\sigma}(A){\sigma}(A)$.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2009.01a
• /
• pp.334-336
• /
• 2009

Estimation and correction of color temperature of digital images are basis of white balance adjustment after image acquisition stage. White balance is one of the most important image processing techniques for subjective image quality enhancement. Correction of color temperature is applied for white balance adjustment or for changing the mood of a picture. A picture taken under the daylight can be changed to have a mood of sunset or cloudy day, for example. We evaluate color temperature transformation of high dynamic range images in linear and log domain, and we conclude that linear domain transformation shows better results.

• 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.