• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.7
• /
• pp.2785-2799
• /
• 2020

We propose a novel speaker adaptation method using acoustic model clustering. The similarity of different speakers is defined by the cosine distance between their i-vectors (intermediate vectors), and various efficient clustering algorithms are applied to obtain a number of speaker subsets with different characteristics. The speaker-independent model is then retrained with the training data of the individual speaker subsets grouped by the clustering results, and an unknown speech is recognized by the retrained model of the closest cluster. The proposed method is applied to a large-scale speech recognition system implemented by a hybrid hidden Markov model and deep neural network framework. An experiment was conducted to evaluate the word error rates using Resource Management database. When the proposed speaker adaptation method using i-vector based clustering was applied, the performance, as compared to that of the conventional speaker-independent speech recognition model, was improved relatively by as much as 12.2% for the conventional fully neural network, and by as much as 10.5% for the bidirectional long short-term memory.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.1
• /
• pp.47-50
• /
• 2019

This paper proposes an acoustic model adaptation method for effective speech recognition in noisy environments. In the proposed algorithm, the noise corruption function is estimated employing DNN (Deep Neural Network), and the function is applied to the model parameter estimation. The experimental results using the Aurora 2.0 framework and database demonstrate that the proposed model adaptation method shows more effective in known and unknown noisy environments compared to the conventional methods. In particular, the experiments of the unknown environments show 15.87 % of relative improvement in the average of WER (Word Error Rate).

• Asian-Australasian Journal of Animal Sciences
• /
• v.17 no.8
• /
• pp.1039-1046
• /
• 2004

A model Indigenous Thai beef cattle breeding structure consisting of nucleus, multiplier and commercial units was used to evaluate the effect of changes in heritabilities of and genetic correlations between adaptation traits on genetic gain and profitability. A breeding objective that incorporated adaptation was considered. Two scenarios for improving both the production and the adaptation of animals where also compared in terms of their genetic and economic efficiency. A base scenario was modelled where selection is for production traits and adaptation is assumed to be under the forces of natural selection. The second scenario (+Adaptation) included all the information available for base scenario with the addition of indirect measures of adaptation. These measures included tick count (TICK), faecal egg count (FEC) and rectal temperature (RECT). Therefore, the main difference between these scenarios was seen in the records available for use as selection criteria and hence the level of investments. Additional genetic gain and profitability was generated through incorporating indirect measures of adaptation as criteria measured in the breeding program. Unsurprisingly, the results were sensitive to the changes in heritabilities and genetic correlations between adaptation traits. However, there were more changes in the genetic gain and profitability of the breeding program when the genetic correlations of adaptation and its indirect measures were varied than when the correlations between these measures were. The changes in the magnitudes of the genetic gain and profit per cow stresses the importance of using reliable estimates of these traits in any breeding program.

• International Journal of Human Ecology
• /
• v.16 no.1
• /
• pp.11-23
• /
• 2015

Although Korean international students/scholars are among the largest groups of international students/scholars on most campuses in the United States, little is known about what types of demands their families face and how they adapt successfully in the face of demands. The purpose of this study was to explore family resilience, which consists of family demands, capabilities, and adaptation, perceived by Korean international student/scholar families, being theoretically guided by the Family Adjustment and Adaptation Response (FAAR) model. Data were collected through face-to-face interviews with couple informants. Following procedures of theory-based content analysis, data were analyzed using key FAAR concepts. Findings showed that most informants reported normative types of family demands such as hardships due to childcare; primary family capabilities were "maintaining social integration," "affective and instrumental communication," and "family cohesiveness," and "nurturance, education, and socialization" was the primary family adaptation mode. New categories under family capabilities, "religious commitment" and "transnational family support" were developed. The results suggest that there is a unique set of family capabilities that contribute to the successful adaptation of Korean international student/scholar families. Implications and limitations are discussed.

• Journal of IKEEE
• /
• v.24 no.4
• /
• pp.998-1003
• /
• 2020

In this paper, an emotion training method using emotional adaptation is proposed to improve the performance of the existing emotion recognition system. For emotion adaptation, an emotion speech model was created from a speech model without emotion using a small number of training emotion voices and emotion adaptation methods. This method showed superior performance even when using a smaller number of emotional voices than the existing method. Since it is not easy to obtain enough emotional voices for training, it is very practical to use a small number of emotional voices in real situations. In the experimental results using a Korean database containing four emotions, the proposed method using emotional adaptation showed better performance than the existing method.

• Proceedings of the IEEK Conference
• /
• 2003.07d
• /
• pp.1491-1494
• /
• 2003

Video adaptation in response to dynamic resource conditions and user preferences is required as a key technology to enable universal multimedia access (UMA) through heterogeneous networks by a multitude of devices in a seamless way. Although many adaptation techniques exist, selections of appropriate adaptations among multiple choices are often ad hoc. To provide a systematic solution, we present a general conceptual framework to model video entity, adaptation, resource, utility, and relations among them. It allows for formulation of various adaptation problems as resource-constrained utility maximization. We apply the framework to a practical case of dynamic bit rate adaptation. Furthermore, we present a description tool, which has been accepted as a part of the MPEG-21 Digital Item Adaptation (DIA), along with a brief overview of the .elated descriptors to support terminal and network quality of service (QoS).

• Journal of information and communication convergence engineering
• /
• v.21 no.2
• /
• pp.139-144
• /
• 2023

Engineers prefer deep neural networks (DNNs) for solving computer vision problems. However, DNNs pose two major problems. First, neural networks require large amounts of well-labeled data for training. Second, the covariate shift problem is common in computer vision problems. Domain adaptation has been proposed to mitigate this problem. Recent work on adversarial-learning-based unsupervised domain adaptation (UDA) has explained transferability and enabled the model to learn robust features. Despite this advantage, current methods do not guarantee the distinguishability of the latent space unless they consider class-aware information of the target domain. Furthermore, source and target examples alone cannot efficiently extract domain-invariant features from the encoded spaces. To alleviate the problems of existing UDA methods, we propose the mixup regularization in adversarial discriminative domain adaptation (ADDA) method. We validated the effectiveness and generality of the proposed method by performing experiments under three adaptation scenarios: MNIST to USPS, SVHN to MNIST, and MNIST to MNIST-M.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.40 no.4
• /
• pp.351-359
• /
• 1991

This paper describes performances of speaker adaptation using a probabilistic spectral mapping matrix in hidden-Markov model(HMM) -based Korean isolated word recognition. Speaker adaptation based on probabilistic spectral mapping uses a well-trained prototype HMM's and is carried out by Viterbi, dynamic time warping, and forward-backward algorithms. Among these algorithms, the best performance is obtained by using the Viterbi approach together with codebook adaptation whose improvement for isolated word recognition accuracy is 42.6-68.8 %. Also, the selection of the initial values of the matrix and the normalization in computing the matrix affects the recognition accuracy.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2007.05a
• /
• pp.285-288
• /
• 2007

There is substantial evidence from earlier researches that older adults have difficult seeing under low illumination and at night, even in the absence of ocular diseases. During human aging, there is a rampant decrease in rod/cone-meditated adaptation which is caused by delayed rhodopsin regeneration and pigment depletion. This calls for a need to develop appropriate visual gadgets to effectively aid the aging generation. Our research culminates its approach from Pattanaik's model by making extensions to temporal visual filtering, thereby simulating a reduction of visual response which comes with age. Our filtering model paves way and lays a foundation for future research to develop a more effective adaptation model that may be further used in developing visual content adaptation aids and guidelines in MPEG 21 environment. We demonstrate our visual model using a High Dynamic Range image and the experiment results are in conversant with the psychophysical data from previous vision researches.

• The Journal of the Acoustical Society of Korea
• /
• v.25 no.4
• /
• pp.159-163
• /
• 2006

In this paper a data-driven method to improve the performance of the Jacobian adaptation (JA) for the noisy speech recognition is proposed. In stead of constructing the reference HMM by using the model composition method like the parallel model combination (PMC), we propose to train the reference HMM directly with the noisy speech. This was motivated from the idea that the directly trained reference HMM will model the acoustical variations due to the noise better than the composite HMM. For the estimation of the Jacobian matrices, the Baum-Welch algorithm is employed during the training. The recognition experiments have been done to show the improved performance of the proposed method over the Jacobian adaptation as well as other model compensation methods.