• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.31 no.4
• /
• pp.393-401
• /
• 1995

In this paper, vector analysis on the novel quick torque control of Induction Motors(I.M) based on voltage-controlled type is conducted. It was very difficult to get a step response of torque when the primary voltage was selected as control input of induction motors in conventional quick torque control methods. To solve this problem, the new control method was developed using a new concept of pulse addition which can realize the stepwise torque response of a specified settling time of $\Delta$. The new method was successfully confirmed through DSP(Digital Signal Processor) system-based experiments. However, it was a little difficult to understand the control mechanism intutionally. The purpose of this paper is to provide more understanding about the quick torque control mechanism using the vector analysis.

• The Journal of Korean Association of Computer Education
• /
• v.13 no.1
• /
• pp.65-73
• /
• 2010

This study offered lexical knowledge acquisition model of unsupervised learning method in order to overcome limitation of lexical knowledge hand building manual of supervised learning method for research of natural language processing. The offered model obtains the lexical knowledge from the lexical entry which was given by inputting through the process of vectorization, clustering, lexical knowledge acquisition automatically. In the process of obtaining the lexical knowledge acquisition of model, some parts of lexical knowledge dictionary which changes in the number of lexical knowledge and characteristics of lexical knowledge appeared by parameter changes were shown. The experimental results show that is possibility of automatic building of Machine-readable dictionary, because observed to the number of lexical class information cluster collected constant. also building of lexical ditionary including left-morphosyntactic information and right-morphosyntactic information is reflected korean characteristic.

• The Journal of the Acoustical Society of Korea
• /
• v.6 no.3
• /
• pp.60-67
• /
• 1987

In this paper, a phoneme classification method for Korean speech recognition has been proposed and its performance has been studied. The phoneme classification has been done based on the phonemic features extracted by the prewindowed recursive least-square (PRLS) algorithm that is a kind of adaptive filter algorithms. Applying the PRLS algorithm to input speech signal, precise detection of phoneme boundaries has been made, Reference patterns of Korean phonemes have been generated by the ordinery vector quantization (VQ) of feature vectors obtained manualy from prototype regions of each phoneme. In order to obtain the performance of the proposed phoneme classification method, the method has been tested using spoken names of seven Korean cities which have eleven different consonants and eight different vowels. In the speaker-dependent phoneme classification, the accuracy is about $85\%$ considering simple phonemic rules of Korean language, while the accuracy of the speaker-independent case is far less than that of the speaker-dependent case.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.19 no.12
• /
• pp.2313-2327
• /
• 1994

As the communication environment evolves, there is an increasing need for multiresolution image coding. To meet this need, the entrophy constratined vector quantizer(ECVQ) for coding of image pyramids by spline wavelet transform is introduced in this paper. This paper proposes a new scheme for image compression taking into account psychovisual feature both in the space and frequency domains : this proposed method involves two steps. First we use spline wavelet transform in order to obtain a set of biorthogonal subclasses of images ; the original image is decomposed at different scale using a pyramidal algorithm architecture. The decomposition is along the vertical and horizontal directions and maintains constant the number of pixels required the image. Second, according to Shannon's rate distortion theory, the wavelet coefficients are vectored quantized using a multi-resolution ECVQ(entropy-constrained vector quantizer) codebook. The simulation results showed that the proposed method could achieve higher quality LENA image improved by about 2.0 dB than that of the ECVQ using other wavelet at 0.5 bpp and, by about 0.5 dB at 1.0 bpp, and reduce the block effect and the edge degradation.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.11a
• /
• pp.946-948
• /
• 2005

H.264/AVC는 SoC/IEC MPEG와 ITU-T Video Coding Experts Group에서 함께 발표한 비디오 코딩을 위한 가장 최근의 표준이다. 기존의 표준들 보다 적은 비트로 높은 압축률과 좋은 화질을 제공하고 있다. 그러나 1/4 화소 움직임 예측과 보상의 지원과 7가지의 가변블록에 대한 움직임예측과 블록모드별 RD(Rate-Distortion)를 수행하고 CAVLC등 H.264/AVC 표준에서 채택한 여러 가지 비디오 압축방식으로 인해 그 복잡도가 훨씬 증가하였다. 이 논문에서 H.264/AVC의 복호화기에서 복잡도의 약 $40\%$ 이상을 차지하는 움직임보상 모듈을 효율적으로 수행하고 최적화하기 위한 방법을 제안한다. 예측된 모션벡터에 따라 창조하는 프레임에서 매크로블록을 만들어 내는 움직임 보상 과정을 수행하는 데 있어서 접근 지연시간이 큰 외부 메모리 창조를 선인출 메커니즘을 이용하여 미리 예측하여 수행함으로써 전체 수행시간을 줄이는 기법을 적용하였다. 이를 통하여 가변길이 복호화 모듈과 움직임 보상모듈을 수정하여 반복적으로 읽고, 쓰기를 수행해야 하는 횟수를 줄였다. 실험 결과 본 논문에서 제안하는 방법을 이용하여 복호화 과정을 수행했을 때 PSNR(Peak Signal to Noise Ratio)의 손실은 전혀 없으면서 복호화기의 전체 실행시간을 약 $5\%$ 향상시키고, 핵심 모듈인 움직임 보상과정에서 약 $20\%$ 정도 실행시간을 향상시키는 등 높은 성능 향상을 보였다.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.46 no.3
• /
• pp.1-7
• /
• 2009

An approach for the classification of music genres using a Fuzzy c-Means(FcM) with divergence-based kernel is proposed and presented in this paper. The proposed model utilizes the mean and covariance information of feature vectors extracted from music data and modelled by Gaussian Probability Density Function (GPDF). Furthermore, since the classifier utilizes a kernel method that can convert a complicated nonlinear classification boundary to a simpler linear one, he classifier can improve its classification accuracy over conventional algorithms. Experiments and results on collected music data sets demonstrate hat the proposed classification scheme outperforms conventional algorithms including FcM and SOM 17.73%-21.84% on average in terms of classification accuracy.

• Journal of KIISE:Computer Systems and Theory
• /
• v.33 no.6
• /
• pp.344-353
• /
• 2006

In this paper, we present a new algorithm which imitate real pencil strokes. The purpose of research on NPR(Non-Photorealistic Rendering) is simulating automatically manmade artistic expressions such as pen-and-ink illustrations, watercolor paintings, pencil sketches and pastel drawings with computers. Recently, there has been a great deal of research works on NPR. One of them is researching in pencil illustration methods for NPR, and a lot of researchers have investigated into the LIC(Linear Integral Convolution) techniques which would change the initial images into the output images by directional vector field images for generating effects of pencil. However, the LIC techniques can not be applied to real-time drawing tools because they are post processing techniques. This paper presents a real-time pencil strokes algorithm which is based on an observation of how pencils(from 6B to 6H) draw lines. Although this algorithm using some pencil variables and noise generation is simple, it is fast and also can draw real-time pencil strokes similar to real manmade pencil strokes in a GUI drawing tool.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.40 no.5
• /
• pp.303-311
• /
• 2003

Since the MPEG-4 visual standard enables content-based functionalities, it is necessary to extract video object from natural video sequences. Segmentation algorithms can largely be classified into automatic segmentation and user-assisted segmentation. In this paper, we propose a user-assisted VOP generation method based on the geometric active contour. Since the geometric active contour, unlike the parametric active contour, employs the level set method to evolve the curve, we can draw the initial curve independent of the shape of the object. In order to generate the edge function from a smoothed image, we propose a vector-valued diffusion process in the LUV color space. We also present a discrete 3-D diffusion model for easy implementation. By combining the curve shrinkage in the vector field space with the curve expansion in the empty vector space, we can make accurate extraction of visual objects from video sequences.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.2C
• /
• pp.109-116
• /
• 2006

Many application scenarios of ultra-wideband(UWB) radio assume non-line-of-sigit(non-LoS) signal propagations. Through-material propagation of UWB signal introduces a distortion of the waveform as well as attenuation, which will introduce a decrease of the correlation coefficient between the correlator template and the received signal. A hypothesis test approach to selection of the template waveform for UWB rake receivers is posed. Linde-Buzo-Gray(LBG) algorithm is used to select the candidate waveforms which are used to setup the hypothesis test. The performance of the algorithm is tested using a set of indoor non-LoS propagation measurement data.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.939-942
• /
• 2001

본 논문에서는 ITU-T G.723.1, G.729 부호화기와 MLT(Modulated Lapped Transform) 방법을 이용한 광대역 음성 부호화방법을 제안한다. 제안된 광대역 음성부호화 방법은 16 kHz로 샘플링된 입력신호를 QMF(Quadrature Mirror Filter)사용하여 저대역과 고대역으로 나누며, 각 대역은 8 kHz의 샘플링을 갖는 협대역 음성 신호로 변환된다. 고대역은 MLT변환 후 벡터 양자화하며 또한 MLT를 사용한 ATC(Adaptive Transform Coding)방법을 적용하여 표현하며 저대역은 G.723.1과 G.729 부호화기를 사용한다. 설계된 광대역 음성부호화기의 성능을 평가하기 위하여 MOS (Mean Opinion score)실험을 수행하였다. MOS 실험을 통해 16 kbps G.729-MLT VQ방식이 G.722 56kbps 와 비슷한 음질을 나타내었다.