• 한국정보처리학회논문지
• /
• 제1권1호
• /
• pp.81-91
• /
• 1994

벡터 양자화기를 이용하여 영상의 부호화기를 설계하는데 있어서 2차원 이산여현 변환(2D-Discrete Cosine Transform)에 근거한 DCT 맵(map)과 새로운 부호책(codebook) 설계로서 알고리듬을 제안한다. 영상을 작은 부블럭으로 나누고 2차원 이산여현변환 으로 대부분의 정보를 포함하는 부분, 즉 부호화하기 어려운 부분과 부호화하기 쉬운 적은 정보를 포함하는 영역으로 나누어 맵을 만들고 이 맵에 따라 영상의 중요한 특징 들을 2차원 이산여현변환으로 추출한다. 부호책은 트리 구조에 근거한 2진 트리로 두 영역을 따로 학습세트로 나눔으로서 만들어 진다. 2진 트리의 중간 노드에서 각 학 습 벡터는 그 노드에서의 문턱 값과 비교하여 두개의 아래 노드중 하나에 속하게 된다. 국제 표준화상인 Lenna와 Boat 영상에 대하여 본 알고리듬으로 영상을 부호화했을 때 PNN과 CVQ 알고리듬에 비하여 수행 시간을 줄이고, PNN 알고리듬보다는 각각 약0.45 dB과 0.33 dB만큼, CVQ 알고리듬보다 각각 약 0.05 dB과 0.1 dB만큼 더 좋은 영상의 화질을 얻을 수 있었다.

• 전기전자학회논문지
• /
• 제22권3호
• /
• pp.558-564
• /
• 2018

In distributed estimation where each node can collect only partial information on the parameter of interest without communication between nodes and quantize it before transmission to a fusion node which conducts estimation of the parameter, we consider a novel quantization technique employed at local nodes. It should be noted that the performance can be greatly improved if each node can transmit its measurement to one designated node (namely, head node) which can quantize its estimate using the total rate available in the system. For this case, the best strategy at the head node would be simply to partition the parameter space using the generalized Lloyd algorithm, producing the global codewords, one of which is closest to the estimate is transmitted to a fusion node. In this paper, we propose an iterative design algorithm that seeks to efficiently assign the codewords into each of quantization partitions at nodes so as to achieve the performance close to that of the system with the head node. We show through extensive experiments that the proposed algorithm offers a performance improvement in rate-distortion perspective as compared with previous novel techniques.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2001년도 하계종합학술대회 논문집(3)
• /
• pp.9-12
• /
• 2001

Learning vector quantization (LVQ) is a supervised learning technique that uses class information to move the vector quantizer slightly, so as to improve the quality of the classifier decision regions. In this paper we propose a selection method of initial codebook vectors for a teaming vector quantization (LVQ3) using support vector machines (SVM). The method is experimented with artificial and real design data sets and compared with conventional methods of the condensed nearest neighbor (CNN) and its modifications (mCNN). From the experiments, it is discovered that the proposed method produces higher performance than the conventional ones and then it could be used efficiently for designing nonparametric classifiers.

• 한국음향학회지
• /
• 제17권7호
• /
• pp.57-60
• /
• 1998

본 논문에서는 MKM(Modified K-Means) 알고리듬의 성능을 개선하기 위해 새로운 학습알고리듬을 제안한다. MKM 알고리듬에서 새로운 코드워드는 이전 코드워드와 새로 얻 은 중심점을 이은 직선 상의 임의적으로 선택된다. 따라서 MKM 알고리듬은 통계적 이완 방법의 코드북 교란 알고리듬으로 이해될 수 있다. MKM 알고리듬을 통계적 이완 알고리듬 과 비교해보면 도입되는 교란의 양이 상대적으로 적고 그 교란 자체도 임의적이지 않다는 걸 알 수 있다. 따라서 MKM 알고리듬에 도입되는 교란의 양을 보다 크고 임의적이게 하면 MKM 알고리듬이 국소 최적화에 빠질 가능성이 줄어들 것이다. 따라서 본 논문에서는 MKM 알고리듬의 코드북 갱신과정을 변화시킨 새로운 알고리듬을 제안하였으며, 화상 데이 터와 음성 데이터를 이용하여 실험한 결과 제안된 알고리듬이 MKM 알고리듬보다 우수한 성능을 보인다는 걸 확인할 수 있다.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2003년도 학술회의 논문집 정보 및 제어부문 A
• /
• pp.75-78
• /
• 2003

In this paper, a second-order 3-bit DSM using DWA(Data Weighted Averaging) algorithm is designed for bluetooth Zero-IF Receiver. The designed circuit has two integrators using a designed OTA, nonoverlapping two-phase clerk generator, 3-bit A/D converter, DWA algorithm and 3-bit D/A converter An ideal model of second-order lowpass DSM with a 3-bit quantizer was configured by using MATLAB, and each coefficients and design specification of each blocks were determined to have 10-bit resolution in 1MHz channel bandwidth. The designed second-order 3-blt lowpass DSM has maximum SNR of 74dB and power consumption is 50mW at 3.3V.

• 대한전자공학회논문지
• /
• 제27권5호
• /
• pp.808-815
• /
• 1990

This paper treats with the design method for the single loop oversampled Sigma-Delta A/D converter with one delay and the digital integrator. Such an integrator was kgenerated by means of the bilinear transform of the analog integrator. The frequency spectrums of the quantizer and the decimator output signal are evaluated by FFT respectively. With the performance evaluation system, the values of SNR are obtained versus the input sinusoidal signal amplitude, frequency, the oversampling ratio, the DC-input level, the loop gain and the limitting value of the integrator. As compared with existing results, values of SNR versus the input signal amplitude and the oversampling ratio for the suggested system are about 6dB higher then previously reported results respectively. Furthermore, this approach achieves an about 60dB input dynamic range.

• ETRI Journal
• /
• 제38권3호
• /
• pp.463-468
• /
• 2016

With the constant increase in network traffic, wireless operators are finding it more challenging to keep network hardware costs to a minimum. At the same time, the energy cost associated with operating a network has increased proportionally. Therefore, the search for higher network capacity is simultaneously accompanied by the search for a cost-efficient network deployment. In this paper, we show that a saving in transmitted signal energy can be achieved at the signal design level by deploying very specific signal processing techniques. Using an orthogonal frequency-division multiplexing signal for Long-Term Evolution networks as an example, we utilize a novel non-uniform companding quantizer to save a transmitted signal energy. Our results show that by using non-uniform quantization it is possible to further optimize 4G wireless networks.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 Ⅳ
• /
• pp.1767-1770
• /
• 2003

This paper proposes an algorithm that can reduce the estimation error of surface spectral-reflectance(SR) when using a conventional 3-band RGB camera. In the proposed method, the estimation error is reduced by using adaptive principle components (PCs) for each color region. To build an adaptive set of PCs, n SR populations are organized for n PC sets using the Lloyd quantizer design algorithm. The Macbeth Color Checker is utilized for the initial representative SR values for 1485 Munsell color chips as the total color population, then the Munsell chips arc divided into subsets with a set of corresponding adaptive PCs organized for each subset.

• 한국통신학회논문지
• /
• 제8권4호
• /
• pp.156-163
• /
• 1983

旣存의 44,736Mbps 傳送路에 適用하여 컬러TV信號를 傳送할수 있는 合成信號形DPCM시스템이 設計, 構成되었다. 使用된 豫測方法은 4개의 畵素를 利用하는 旣存의 2次元豫測方法이며 5비트에 해당하는 非線形量子器가 使用되었다. 實時間處理를 위해 주로 ECL素子가 많이 使用되었다. 實驗結果를 통하여 本DPCM시스템은 旣存의 傳送路를 使用하여 TV放送信號를 CATV品質이상으로 傳送할 수 있는 能力이 있음을 確認할 수 있었다.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2007년도 하계종합학술대회 논문집
• /
• pp.345-346
• /
• 2007

In this paper, we propose a Pyramid VQ(PVQ) to quantize the transform coefficients of TCX module for the music improvement of AMR-WB+ codec. The proposed PVQ is compared to the $RE_8$ Lattice VQ used in the AHR-WB+ standard codec, demonstrating improvement 4% and 5.7%, respectively, in Mean Squared Error(MSE) and 3.3% and 4.7%, respectively, in Perceptual Evaluation of Audio Quality(PEAQ) by 8-dimensional and 16-dimensional Pyramid VQ.