• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.35S no.9
• /
• pp.63-69
• /
• 1998

In this paper, we propose a new Bit-Rate control algorithm based on bit usage matching to substitute encoded GOP(s) for new GOP(s) in MPEG-2 bitstream. It iteratively encodes current picture according to quantization value of previous picture and records bit-usage of each slice until nearly target bits are used. With target bits falling in two output bits, quantization value of slice should be changed to alleviate output bit error. We use recorded bit-usage information to decide which slices should be encoded with one quantization value and others with another. As every macroblock has different activity, we change macroblock quantization value using slice quantization value and activity value. The simulation results demonstrate that the fluctuation of the output bits can be kept within few-several tens of bits while maintaining the quality of the reconstructed pictures at a relatively stable level.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.3
• /
• pp.330-338
• /
• 2002

Spontaneous speech is ungrammatical as well as serious phonological variations, which make recognition extremely difficult, compared with read speech. In this paper, for conversational speech recognition, we analyze the transcriptions of the real conversational speech, and then classify the characteristics of conversational speech in the speech recognition aspect. Reflecting these features, we obtain the baseline system for conversational speech recognition. The classification consists of long duration of silence, disfluencies and phonological variations; each of them is classified with similar features. To deal with these characteristics, first, we update silence model and append a filled pause model, a garbage model; second, we append multiple phonetic transcriptions to lexicon for most frequent phonological variations. In our experiments, our baseline morpheme error rate (WER) is 31.65%; we obtain MER reductions such as 2.08% for silence and garbage model, 0.73% for filled pause model, and 0.73% for phonological variations. Finally, we obtain 27.92% MER for conversational speech recognition, which will be used as a baseline for further study.

• Journal of Broadcast Engineering
• /
• v.4 no.1
• /
• pp.76-85
• /
• 1999

The degradation of image quality by block loss is more serious in fractal coded images with the error propagation due to mapping from the lost blocks than in DCT coded images. Therefore. a new algorithm is presented for recovering the blocks lost in the transmission through the lossy network as A TM network of the images coded by Jacquins fractal coding. Jacquins fractal code is divided into two layers of header code and main code according to its importance. The key technique of the proposed BLRA (block loss recovery algorithm) is a fractal extrapolation that estimates the lost pixels by using the contractive mapping parameters of the neighboring range blocks whose characteristics are similar to a lost block. The proposed BLRA is applied to the lost blocks in the iteration of decoding. Some experimental results show the proposed BLRA yields excellent performance in PSNR as well as subjective quality.

• Journal of Korean Society of Steel Construction
• /
• v.9 no.4 s.33
• /
• pp.467-477
• /
• 1997

A series of experimental tests have been performed on a tube beam in which artificial damage is applied in order to address damage detectability using modal analysis. Modal parameters considered are frequency, displacement mode shape and strain mode shape CoMAC(Coordinate Modal Assurance Criterion) and Modal Vector Error have been adopted for presenting the change of displacement mode shape and strain mode shape. It is revealed strain mode shape is the most sensitive to damage.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.41 no.6
• /
• pp.27-35
• /
• 2004

Finite or Galois fields have been used in numerous applications such as error correcting codes, digital signal processing and cryptography. These applications often require exponetiation on GF(2$^{m}$ ) which is a very computationally intensive operation. Most of the existing methods implemented the exponetiation by iterative methods using repeated multiplications, which leads to much computational load, or needed much hardware cost because of their structural complexity in implementing. In this paper, we present an effective VLSI architecture for exponentiation on GF(2$^{m}$ ). This circuit computes the exponentiation by multiplying product terms, each of which corresponds to an exponent bit. Until now use of this type algorithm has been confined to a primitive element but we generalize it to any elements in GF(2$^{m}$ ).

• Proceedings of the Korea Water Resources Association Conference
• /
• 2020.06a
• /
• pp.219-219
• /
• 2020

하천의 수질을 나타내는 환경지표 중 국가 TMS(Tele Monitoring system)의 수질측정망을 통해 관리되고 있는 지표로는 DO, BOD, COD, SS, TN, TP 등 여러 인자들이 있다. 이러한 수질인자는 하천의 자정작용에 있어 많은 영향을 나타내고 있다. 이를 활용한 경제적이고 합리적인 수질관리를 위해 하천의 자정작용을 활용하는 것이 중요하다. 생물학적 작용을 가장 효과적으로 활용하기 위해서는 수질오염 데이터에 기초한 수질예측을 채택하여 적절한 대책이 필요하다. 이를 위해서는 수질인자의 데이터를 측정하고 축적해 수질오염을 예측하는 것이 필수적인데, 실제적으로 수질인자의 일일 측정은 비용 관점에서 쉽게 접근할 수 없다. 본 연구에서는 시계열 학습으로 알려진 RNN-LSTM(Recurrent Neural Network-Long Term Memory) 알고리즘을 활용하여 기존에 측정된 수질인자의 데이터를 통해 시간당 및 일일 수질인자를 예측하려고 했다. 연구에 앞서, 기존에 시간단위로 측정된 수질인자 데이터의 이상 유무를 확인 후, 에러값은 제거하고 12시간 이하 데이터가 누락되었을 때는 선형 보간하여 데이터를 사용하고, 1일 데이터도 10일 이하 데이터가 누락되었을 때 선형 보간하여 데이터를 활용하여 수질인자를 예측하였다. 수질인자를 예측하기 위해 구글이 개발한 딥러닝 오픈소스 라이브러리인 텐서플로우를 활용하였고, 연구지역으로는 대한민국 부산에 위치한 온천천의 유역을 선정하였다. 수질인자 데이터 수집은 부산광역시에서 운영하는 보건환경정보 공개시스템의 자료를 활용하였다. 모델의 연구를 위해 하천의 수질인자, 기상자료 데이터를 입력자료로 활용하였다. 분석에서는 입력자료와, 반복횟수, 시계열의 길이 등을 조절해 수질 요인을 예측했고, 모델의 정확도도 분석하였다.

• The Transactions of the Korea Information Processing Society
• /
• v.3 no.3
• /
• pp.439-448
• /
• 1996

Extraction and selection of the informative features play a central role in pattern recognition. This paper describes a modified back-propagation algorithm that performs selection of the informative features and trains a neural network simultaneously. The algorithm is mainly composed of three repetitive steps : training, connection pruning, and input unit elimination. Afer initial training, the connections that have small magnitude are first pruned. Any unit that has a small number of connections to the hidden units is deleted,which is equivalent to excluding the feature corresponding to that unit.If the error increases,the network is retraned,again followed by connection pruning and input unit elimination.As a result,the algorithm selects the most im-portant features in the measurement space without a transformation to another space.Also,the selected features are the most-informative ones for the classification,because feature selection is tightly coupled with the classifi-cation performance.This algorithm helps avoid measurement of redundant or less informative features,which may be expensive.Furthermore,the final network does not include redundant parameters,i.e.,weights and biases,that may cause degradation of classification performance.In applications,the algorithm preserves the most informative features and significantly reduces the dimension of the feature vectors whiout performance degradation.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.42 no.5 s.305
• /
• pp.51-58
• /
• 2005

This paper proposes an efficient algorithm for extracting a tumor region from an breast ultrasound volume image by using rate-distortion (R-D) based seeded region growing. In the proposed algorithm the rate and the distortion represent the roughness of the contour and the dissimilarity of pixels in a region, respectively. Staring from an initial seed region set in each cutting plane of a volume, a pair of the seed region and one of adjacent regions whose R-D cost is minimal is searched and then they are merged into a new updated seed region. This procedure is recursively performed until the averaged R-D cost values per the number of contour pixels in the seed region becomes maxim. As a result, the final seed region has good pixel homogeneity and a much smooth contour. Finally, the tumor volume is extracted using the contours of the final seed regions in all the cutting planes. Experimental results show that the averaged error rate of the proposed method is shown to be below 4%.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.7C
• /
• pp.691-699
• /
• 2006

Despite of the in-depth effort to cantrol the variability in IP networks, quality of service (QoS) is still not guaranteed in the IP networks. Thus, it is necessary to deal with the audible artifacts caused by packet lasses. To overcame the packet loss problem, most speech coding standard have their own embedded packet loss concealment (PLC) algorithms which adapt extrapolation methods utilizing the dependency on adjacent frames. Since many low bit rate CELP coders use predictive schemes for increasing coding efficiency, however, error propagation occurs even if single packet is lost. In this paper, we propose an efficient PLC algorithm with consideration about the speech characteristics of lost frames. To design an efficient PLC algorithm, we perform several experiments on investigating the error propagation effect of lost frames of a predictive coder. And then, we summarize the impact of packet loss to the speech characteristics and analyze the importance of the encoded parameters depending on each speech classes. From the result of the experiments, we propose a new PLC algorithm that mainly focuses on reducing the error propagation time. Experimental results show that the performance is much higher than conventional extrapolation methods over various frame erasure rate (FER) conditions. Especially the difference is remarkable in high FER condition.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.10C
• /
• pp.1017-1026
• /
• 2005

In burst OFDM system, the frame synchronization should be performed first for the acquisition of received frame and the estimation of the correct FFT-window position. The conventional frame synchronization algorithms using design features of the preamble symbol, the repetition pattern of the OFDM symbol by pilot sub-carrier allocation rule and Cyclic Prefix(CP), has difficulty in the detection of precise frame timing because its correlation characteristics would increase and decrease gradually. Also, the algorithm based on the correlation between the reference signal and the received signal has performance degradation due to frequency offset. Therefore, we adopt a differential correlation method that is robust to frequency offset and has the clear peak value at the correct frame timing for frame synchronization. However, performance improvement is essential for differential correlation methods, since it usually shows multiple peak values due to the repetition pattern. In this paper, we propose an enhanced frame synchronization algorithm based on the differential correlation method that shows a clear single peak value by using differential correlation between samples of identical repeating pattern. We also introduce a normalization scheme which normalizes the result of differential correlation with signal power to reduce the frame timing error in the high speed mobile channel environments.