• 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.

• Journal of Multimedia Information System
• /
• v.4 no.4
• /
• pp.317-320
• /
• 2017

This paper examines approaches to the development of algorithms of parametric identification of models of complex systems from discrete observations. A modification of a known algorithm Kaczmarz which is designed for closed systems with perturbations, based on the methods of random search and investigates their statistical properties.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.4 no.1
• /
• pp.43-49
• /
• 1999

We propose a nonlinear feedback controller that can control the induction motors with high dynamic performance and high power efficiency by means of decoupling of motor speed and rotor flux. The nonlinear feedback controller needs the information on some motor parameters. New recursive adaptation algorithms for rotor resistance and mutual inductance which can be applied to our nonlinear feedback controller are also presented in this paper. The recursive adaptation algorithms make the estimated values of rotor resistance and mutual inductance track their real values. Some simulation and experimental results show that the adaptation algorithms are robust against the variation of stator resistance and stator inductance.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.2
• /
• pp.190-198
• /
• 2002

In general, we have used the Viterbi algorithm of Speech recognition for decoding. But a decoder in speaker verification has to recognize same word of every speaker differently. In this paper, we propose a noble decoding algorithm that could replace the typical Viterbi algorithm for the speaker verification system. We utilize for the proposed algorithm the speaker adaptation algorithms that transform feature vectors into the region of the client' characteristics in the speech recognition. There are many adaptation algorithms, but we take MLLR (Maximum Likelihood Linear Regression) and MAP (Maximum A-Posterior) adaptation algorithms for proposed algorithm. We could achieve improvement of performance about 30% of EER (Equal Error Rate) using proposed algorithm instead of the typical Viterbi algorithm.

• Journal of Ship and Ocean Technology
• /
• v.5 no.3
• /
• pp.27-35
• /
• 2001

In this research, hybrid method with case-based reasoning and rule-based reasoning is applied. Using case-based reasoning, design experts'experience and know-how are effectively represented in order to obtain a proper configuration of midship section in the initial ship design stage. Since there is not sufficient domain knowledge available to us, traditional case-adaptation algorithms cannot be applied to our problem, i.e., creating the configuration of midship section. Thus, new case-adaptation algorithms not requiring any domain knowledge are developed antral applied to our problem. Using the knowledge representation of DnV rules, rule-based reasoning can perform deductive inference in order to obtain the scantling of midship section efficiently. The results from the case-based reasoning and the rule-based reasoning are examined by comparing the results with various conventional methods. And the reasonability of our results is verified by comparing the results wish actual values from parent ship.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.36 no.10A
• /
• pp.839-846
• /
• 2011

Link Adaptation should select the best modulation and coding scheme (MCS) which gives the highest throughput as channel conditions vary. Several link adaptation algorithms for wireless local area network (WLAN) have been proposed but for the future WLAN systems such as 802.11n system, these algorithms do not guarantee the best performance. In this paper, we propose a new link adaptation algorithm in which an MCS level is chosen by the received SNR plus the offset value obtained from the transmission results. The performance of proposed algorithm is simulated by an IEEE 802.11n system. From the analysis, we conclude the proposed algorithm performs better than the well-known link adaptation algorithms such as auto rate fallback and general SNR-based techniques. Particularly, the proposed algorithm improves throughput when the packet error ratio (PER) is constrained for fast fading channels.

• Proceedings of the Korea Multimedia Society Conference
• /
• 2004.05a
• /
• pp.326-329
• /
• 2004

Content adaptation is an effective solution to support the quality of service over multimedia services over heterogeneous environments. This paper deals with the accuracy and the real-time requirement, two important issues in making decision on content adaptation. From our previous problem formulation, we propose an optimal algorithm and a fast approximation based on the Viterbi algorithm of dynamic programming. Through extensive experiments, we show that the proposed algorithms can enable accurate adaptation decisions, and especially they can support the real-time processing.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.24 no.4
• /
• pp.382-391
• /
• 2021

Unsupervised domain adaptation often gives impressive solutions to handle domain shift of data. Most of current approaches assume that unlabeled target data to train is abundant. This assumption is not always true in practices. To tackle this issue, we propose a general solution to solve the domain gap minimization problem without any target data. Our method consists of two regularization steps. The first step is a pixel regularization by arbitrary style transfer. Recently, some methods bring style transfer algorithms to domain adaptation and domain generalization process. They use style transfer algorithms to remove texture bias in source domain data. We also use style transfer algorithms for removing texture bias, but our method depends on neither domain adaptation nor domain generalization paradigm. The second regularization step is a feature regularization by feature alignment. Adding a feature alignment loss term to the model loss, the model learns domain invariant representation more efficiently. We evaluate our regularization methods from several experiments both on small dataset and large dataset. From the experiments, we show that our model can learn domain invariant representation as much as unsupervised domain adaptation methods.

• Journal of electromagnetic engineering and science
• /
• v.4 no.2
• /
• pp.93-96
• /
• 2004

In this paper we introduce a new escalator(ESC) structure-based adaptation algorithm. The proposed algorithm is independent of eigenvalues spread ratio(ESR) of channel and has faster convergence speed than that of the conventional ESC algorithms. This algorithm combines the fast adaptation ability of least square methods and the orthogonalization property of the ESC structure. From the simulation results the proposed algorithm shows superior convergence speed and no slowing down of convergence speed when we increase the ESR of the channel.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.8
• /
• pp.3856-3872
• /
• 2018

HTTP-based adaptive streaming (HAS) has recently been widely deployed on the Internet. In the HAS system, a video content is encoded at multiple bitrates and the encoded video content is segmented into small parts of fixed durations. The HAS client requests a video segment and stores it in the playout buffer. The rate adaptation algorithm employed in HAS clients dynamically determines the video bitrate depending on the time-varying bandwidth. Many studies have shown that an efficient rate adaptation algorithm is critical to ensuring quality-of-experience in HAS systems. However, existing algorithms have problems estimating the network bandwidth because bandwidth estimation is performed on the client-side application stack. Without the help of transport layer protocols, it is difficult to achieve accurate bandwidth estimation due to the inherent segment-based transmission of the HAS. In this paper, we propose an alternative approach that utilizes the playout buffer occupancy rather than using bandwidth estimates obtained from the application layer. We start with a queueing analysis of the playout buffer. Then, we present a buffer-aware rate adaptation algorithm that is solely based on the mean buffer occupancy. Our simulation results show that compared to conventional algorithms, the proposed algorithm achieves very smooth video quality while delivering a similar average video bitrate.