• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.32B no.11
• /
• pp.153-162
• /
• 1995

When vector quantization is used in low rate image coding (e.g., R<0.5), the primary problem is the tremendous computational complexity which is required to search the whole codebook to find the closest codevector to an input vector. Since the number of code vectors in a vector quantizer is given by an exponential function of the dimension. i.e., L=2$^{nR}$ where Rn. To alleviate this problem, a multiple shell structure of hypercube feature maps (MSSHFM) is proposed. A binary HFM of k-dimension is composed of nodes at hypercube vertices and a multiple shell architecture is constructed by surrounding the k-dimensional hfm with a (k+1)-dimensional HFM. Such a multiple shell construction of nodes inherently has a complete tree structure in it and an efficient partial search scheme can be applied with drastically reduced computational complexity, computer simulations of still image coding were conducted and the validity of the proposed method has been verified.

• Annual Conference of KIPS
• /
• 2018.10a
• /
• pp.805-807
• /
• 2018

본 논문은 인공 대역 확장에서 스펙트럼 포락선 확장 시 발생하는 스펙트럼 왜곡을 줄이는 개선된 조합 매핑(Improved combined mapping) 알고리즘을 제안한다. 벡터양자화를 기반으로 하는 코드북 매핑(Codebook mapping)과 스펙트럼 포락선(Spectrum Envelope)의 선형 의존도를 이용한 선형 매핑(Linear mapping)을 사용하여 각각 확장된 광대역 LSP(Line Spectrum Pair)를 추론하고, 연판정(Soft-decision)을 통해 최적화된 LSP를 추론한다. 제안된 알고리즘으로 합성된 음성신호의 스펙트럼 왜곡(Spectrum Distortion)이 기존 조합매핑으로 얻은 음성 신호의 스펙트럼 왜곡보다 더 적은 왜곡을 갖는 결과를 나타내었다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2012.06c
• /
• pp.448-450
• /
• 2012

본 논문에서는 Kinect 방식의 RGB-D 영상센서를 사용하여, 깊이(Depth) 영상으로부터 3차원 객체의 기하정보를 표현하는 표면 정규 벡터(Surface Normal Vector)를 추출하고, 그 결과를 영상화하는 방법을 제안하며, 제안된 방법으로 생성된 영상으로부터 깊이 영상의 특징점 및 특징 기술자를 추출하여 3차원 객체 인식 성능을 향상시키는 방법을 제안한다. 또한 생성된 RGB-D 특징 기술자들을 객체 단위로 구분 가능한 코드북(CodeBook) 학습을 통한 인식방법을 제안하여 객체의 인식 성능을 높이는 방법을 제안한다. 제안하는 RGB-D 기반의 특징 추출 및 학습 방법은 텍스쳐 유무, 카메라 회전 및 이동 변화 등의 환경변화에 강건함을 실험적으로 증명하였으며, 이 방법은 Kinect 방식의 RGB-D 영상을 사용하는 3차원 객체/공간 인식 및 추적, 혹은 이를 응용하는 증강현실 시스템에 적용하여 사용될 수 있다.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2017.05a
• /
• pp.296-296
• /
• 2017

본 연구에서는 대규모 유역에서 발생하는 침수 현상을 모의하기 위한 강력하고 정확하며 연산효율이 뛰어난 수치해석 모형을 개발하는 데 있다. 개발된 모형은 확산파 모형을 기본으로 하였고 다수의 코어를 동시적으로 해석하는 병렬연산 기법을 부가하였다. 홍수로 인한 대규모 유역에서의 침수해석은 오랜 시간의 연산 비용을 필요로 한다. 특히 수치화된 지형정보의 이용이나 고정밀 사진 측량 등의 방법을 이용하여 정밀하고 넓은 유역의 디지털 지형자료를 이용한 2 차원 침수해석은 연산 연산의 문제를 더욱 어렵게 할 수 있다. 그러므로 본 연구에서는 제내지나 하류 유역에 발생하는 홍수로 발생된 빠른 침수모의를 위해 병렬화된 침수 해석 모형을 이용하여 병렬 해석 모형의 적용성을 검토하고자 하였다. 연구를 위해 MPI 및 OpenMP 기법을 이용하여 2 차원 침수해석 프로그램의 원시코드를 개선하고 실제 제내지 및 실제 댐 하류유역에 적용하였다. 개발된 모형은 실제 제내지에 적용한 결과를 MPI, OpenMP 병렬해석 기법과 기존의 순차적 모형의 결과를 비교하였다. 모형들의 결과를 제내지의 침수양상, 침수 속도벡터의 방향 및 크기 등의 계산 결과 순차적 모형, MPI 및 OpenMP 모형과의 비교하여 연산 시간은 병렬 해석 모형이 우월함을 보였다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.32 no.9C
• /
• pp.895-904
• /
• 2007

This paper proposes a robust sub-optimal smart antenna in Code Division Multiple Access (CDMA) basestation. It makes use of the property of the Least Mean Square (LMS) algorithm and the Conjugate Gradient Method (CGM) algorithm for beamforming processes. The weight update takes place at symbol level which follows the PN correlators of receiver module under the assumption that the post correlation desired signal power is far larger than the power of each of the interfering signals. The proposed algorithm is simple and has as low computational load as five times of the number of antenna elements(O(5N)) as a whole per each snapshot. The output Signal to Interference plus Noise Ratio (SINR) of the proposed smart antenna system when the weight vector reaches the steady state has been examined. It has been observed in computer simulations that proposed beamforming algorithm improves the SINR significantly compared to the single antenna case. The convergence property of the weight vector has also been investigated to show that the proposed hybrid algorithm performs better than CGM and LMS during the initial stage of the weight update iteration. The Bit Error Rate (BER) characteristics of the proposed array has also been shown as the processor input Signal to Noise Ratio (SNR) varies.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.10
• /
• pp.2615-2628
• /
• 1997

In this paper, a new codebook design algorithm is proposed. It uses a DCT map based on two-dimensional discrete cosine of transform (2D DCT) and finite state vector quantizer (FSVQ) when the vector quantizer is designed for image transmission. We make the map by dividing input image according to edge quantity, then by the map, the significant features of training image are extracted by using the 2D DCT. A master codebook of FSVQ is generated by partitioning the training set using binary tree based on tree-structure. The state codebook is constructed from the master codebook, and then the index of input image is searched at not master codebook but state codebook. And, because the coding of index is important part for high speed digital transmission, it converts fixed length codes to variable length codes in terms of entropy coding rule. The huffman coding assigns transmission codes to codes of codebook. This paper proposes single-side growing huffman tree to speed up huffman code generation process of huffman tree. Compared with the pairwise nearest neighbor (PNN) and classified VQ (CVQ) algorithm, about Einstein and Bridge image, the new algorithm shows better picture quality with 2.04 dB and 2.48 dB differences as to PNN, 1.75 dB and 0.99 dB differences as to CVQ respectively.

• The Journal of the Acoustical Society of Korea
• /
• v.33 no.4
• /
• pp.248-254
• /
• 2014

Among acoustic signal analysis tasks, acoustic context awareness is one of the most formidable tasks in terms of complexity since it requires sophisticated understanding of individual acoustic events. In conventional context awareness methods, individual acoustic event detection or recognition is employed to generate a relevant decision on the impending context. However this approach may produce poorly performing decision results in practical situations due to the possibility of events occurring simultaneously or the acoustically similar events that are difficult to distinguish with each other. Particularly, the babble noise acoustic event occurring at a bus or subway environment may create confusion to context awareness task since babbling is similar in any environment. Therefore in this paper, a frequency-cepstral feature vector is proposed to mitigate the confusion problem during the situation awareness task of binary decisions: bus or metro. By employing the Support Vector Machine (SVM) as the classifier, the proposed feature vector scheme is shown to produce better performance than the conventional scheme.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.38A no.1
• /
• pp.1-9
• /
• 2013

This paper propose a human action recognition method that uses bag-of-features (BoF) based on CS-LBP (center-symmetric local binary pattern) and a spatial pyramid in addition to the random forest classifier. To construct the BoF, an image divided into dense regular grids and extract from each patch. A code word which is a visual vocabulary, is formed by k-means clustering of a random subset of patches. For enhanced action discrimination, local BoF histogram from three subdivided levels of a spatial pyramid is estimated, and a weighted BoF histogram is generated by concatenating the local histograms. For action classification, a random forest, which is an ensemble of decision trees, is built to model the distribution of each action class. The random forest combined with the weighted BoF histogram is successfully applied to Standford Action 40 including various human action images, and its classification performance is better than that of other methods. Furthermore, the proposed method allows action recognition to be performed in near real-time.

• Journal of Korea Multimedia Society
• /
• v.5 no.2
• /
• pp.158-166
• /
• 2002

In this paper, we proposed a contour shape description method which use the CFR(contour fluctuation ratio) feature. The CFR is the ratio of the line length to the curve length of a contour segment. The line length means the distance of two end points on a contour segment, and the curve length means the sum of distance of all adjacent two points on a contour segment. We should acquire rotation and scale invariant contour segments because each CFR is computed from contour segments. By using the interleaved contour segment of which length is proportion to the entire contour length and which is generated from all the points on contour, we could acquire rotation and scale invariant contour segments. The CFR can describes the local or global feature of contour shape according to the unit length of contour segment. Therefore we describe the shape of objects with the feature vector which represents the distribution of CFRs, and calculate the similarity by comparing the feature vector of corresponding unit length segments. We implemented the proposed method and experimented with rotated and scaled 165 fish images of fifteen types. The experimental result shows that the proposed method is not only invariant to rotation and scale but also superior to NCCH and TRP method in the clustering power.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.4A
• /
• pp.364-370
• /
• 2002

In this paper, we propose a VLSI array architecture for high speed processing of FBMA. First of all, the sequential FBMA is transformed into a single assignment code by using the index space expansion, and then the dependance graph is obtained from it. The two dimensional VLSI array is derived by projecting the dependance graph along the optimal direction. Since the candidate blocks in the search range are overlapped with columns as well as rows, the processing elements of the VLSI array are designed to reuse the overlapped data. As the results, the number of data inputs is reduced so that the processing performance is improved. The proposed VLSI array has (N$^2$+1)${\times}$(2p+1) processing elements and (N+2p) input ports where N is the block size and p is the maximum search range. The computation time of the rat reference block is (N$^2$+2(p+1)N+6p), and the block pipeline period is (3N+4p-1).