• 제목/요약/키워드: mapping matrix

검색결과 179건 처리시간 0.02초

한국어 격리단어 인식 시스템에서 HMM 파라미터의 화자 적응 (Speaker Adaptation in HMM-based Korean Isoklated Word Recognition)

  • 오광철;이황수;은종관
    • 대한전기학회논문지
    • /
    • 제40권4호
    • /
    • 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.

지적 구조 분석을 위한 MDS 지도 작성 방식의 비교 분석 (A Comparison Analysis of Various Approaches to Multidimensional Scaling in Mapping a Knowledge Domain's Intellectual Structure)

  • 이재윤
    • 한국문헌정보학회지
    • /
    • 제41권2호
    • /
    • pp.335-357
    • /
    • 2007
  • 다차원척도법(MDS)은 지적 구조의 시각화를 위해서 오랫동안 사용되어 왔다. 그러나 MDS는 지적 구조를 시각적으로 표현하는데 있어서 세부 구조의 표현력이 취약하다는 약점을 가지고 있다. 이 연구에서는 상관계수 행렬의 가공 방식과 MDS 알고리즘을 조합한 여섯 가지 MDS 지도 작성 방식을 파악한 다음, 실제 지적 구조 데이터에 적용하여 비교해보았다. 실험 결과에서 가장 나쁜 방식으로 파악된 것은 가장 널리 사용되고 있는 방식으로서, 상관계수행렬로부터 유클리드 거리를 산출한 후 ALSCAL 알고리즘으로 MDS 지도를 작성하는 방식인 것으로 나타났다. 반면에 가장 좋은 방식은 상관계수를 z점수로 표준화하여 유클리드 거리를 산출한 후 PROXSCAL 알고리즘를 사용하는 방식이었다. 결론적으로 MDS 처리 과정을 주의깊게 구성한다면 더 구체적이고 명확한 지적 구조를 파악할 수 있음이 확인되었다.

행렬-스타 그래프를 전위 그래프에 일-대-일 사상하는 알고리즘 (An Algorithm for One-to-One Mapping Matrix-star Graph into Transposition Graph)

  • 김종석;이형옥
    • 한국정보통신학회논문지
    • /
    • 제18권5호
    • /
    • pp.1110-1115
    • /
    • 2014
  • 행렬-스타 그래프와 전위 그래프는 스타 그래프 부류로 그래프 이론 관점에서 노드 대칭성, 고장 허용도, 재귀적 확장성 등 스타그래프의 장점을 가지고 있는 상호연결망이다. 본 논문에서는 그래프 이론의 인접 성질을 이용하여 행렬-스타 그래프와 전위 그래프 사이의 일-대-일 사상 알고리즘을 제안한다. 행렬-스타 그래프가 전위 그래프에 연장율 n 이하에 사상할 수 있음을 보이며, 평균 연장율이 2 이하임을 보인다.

PROMISE: A QR Code PROjection Matrix Based Framework for Information Hiding Using Image SEgmentation

  • Yixiang Fang;Kai Tu;Kai Wu;Yi Peng;Yunqing Shi
    • KSII Transactions on Internet and Information Systems (TIIS)
    • /
    • 제17권2호
    • /
    • pp.471-485
    • /
    • 2023
  • As data sharing increases explosively, such information encoded in QR code is completely public as private messages are not securely protected. This paper proposes a new 'PROMISE' framework for hiding information based on the QR code projection matrix by using image segmentation without modifying the essential QR code characteristics. Projection matrix mapping, matrix scrambling, fusion image segmentation and steganography with SEL(secret embedding logic) are part of the PROMISE framework. The QR code could be mapped to determine the segmentation site of the fusion image as a binary information matrix. To further protect the site information, matrix scrambling could be adopted after the mapping phase. Image segmentation is then performed on the fusion image and the SEL module is applied to embed the secret message into the fusion image. Matrix transformation and SEL parameters should be uploaded to the server as the secret key for authorized users to decode the private message. And it was possible to further obtain the private message hidden by the framework we proposed. Experimental findings show that when compared to some traditional information hiding methods, better anti-detection performance, greater secret key space and lower complexity could be obtained in our work.

두 보정면과 사교좌표 매핑을 이용한 카메라 보정법 (Camera Calibration with Two Calibration Planes and Oblique Coordinate Mapping)

  • 안정호
    • 한국정밀공학회지
    • /
    • 제16권7호
    • /
    • pp.119-124
    • /
    • 1999
  • A method to find the line of sight ray in space which corresponds to a point in an image plane is presented. The line of sight ray is defined by two points which are the intersections between the two calibration planes and the sight ray. The intersection point is found by the oblique coordinate mapping between the image plane and the calibration plane in the space. The proposed oblique coordinate mapping method has advantages over the transformation matrix method in the required memory space and computation time.

  • PDF

Identification of Superior Single Nucleotide Polymorphisms (SNP) Combinations Related to Economic Traits by Genotype Matrix Mapping (GMM) in Hanwoo (Korean Cattle)

  • Lee, Yoon-Seok;Oh, Dong-Yep;Lee, Yong-Won;Yeo, Jung-Sou;Lee, Jea-Young
    • Asian-Australasian Journal of Animal Sciences
    • /
    • 제24권11호
    • /
    • pp.1504-1513
    • /
    • 2011
  • It is important to identify genetic interactions related to human diseases or animal traits. Many linear statistical models have been reported but they did not consider genetic interactions. Genotype matrix mapping (GMM) has been developed to identify genetic interactions. This study uses the GMM method to detect superior SNP combinations of the CCDC158 gene that influences average daily gain, marbling score, cold carcass weight and longissimus muscle dorsi area traits in Hanwoo. We evaluated the statistical significance of the major SNP combinations selected by implementing the permutation test of the F-measure. The effect of g.34425+102 A>T (AA), g.8778G>A (GG) and g.4102+36T>G (GT) SNP combinations produced higher performance of average daily gain, marbling score, cold carcass weight and the longissimus muscle dorsi area traits than the effect of a single SNP. GMM is a fast and reliable method for multiple SNP analysis with potential application in marker-assisted selection. GMM may prospectively be used for genetic assessment of quantitative traits after further development.

VLIW 시뮬레이터 상에서의 디지털 신호처리 행렬 연산에 대한 병렬화 알고리즘 (A Parallelising Algortithm for Matrix Arithmetics of Digital Signal Processings on VLIW Simulator)

  • 송진희;전문석
    • 한국정보처리학회논문지
    • /
    • 제5권8호
    • /
    • pp.1985-1996
    • /
    • 1998
  • 본 논문에서는 행렬 또는 벡터 곱셈을 선형 프로세서나 VLIW 시뮬레이터로 분할 및 배치하는 알고리즘을 제안한다. 먼저 입력 행렬이나 벡터를 임의 크기의 프로세서 배열에 배치하는 기법에 대해 논의하고, 문제 크기를 프로세서 배열 크기로 분할하는 알고리즘을 보인다. 이 알고리즘을 VLIW 시뮬레이터 상에서 실행하고 알고리즘의 효율성을 보이도록한다. 그 결과 우리가 설계한 VLIW 시뮬레이터 상에서의 수행이 선형 프로세서 상에서 보다 병렬화 성능이 향상됨을 알 수 있었다.

  • PDF

2D Mesh SIMD 구조에서의 병렬 행렬 곱셈의 수치적 성능 분석 (An Analytical Evaluation of 2D Mesh-connected SIMD Architecture for Parallel Matrix Multiplication)

  • 김정길
    • 정보통신설비학회논문지
    • /
    • 제10권1호
    • /
    • pp.7-13
    • /
    • 2011
  • Matrix multiplication is a fundamental operation of linear algebra and arises in many areas of science and engineering. This paper introduces an efficient parallel matrix multiplication scheme on N ${\times}$ N mesh-connected SIMD array processor, called multiple hierarchical SIMD architecture (HMSA). The architectural characteristic of HMSA is the hierarchically structured control units which consist of a global control unit, N local control units configured diagonally, and $N^2$ processing elements (PEs) arranged in an N ${\times}$ N array. PEs are communicating through local buses connecting four adjacent neighbor PEs in mesh-torus networks and global buses running across the rows and columns called horizontal buses and vertical buses, respectively. This architecture enables HMSA to have the features of diagonally indexed concurrent broadcast and the accessibility to either rows (row control mode) or columns (column control mode) of 2D array PEs alternately. An algorithmic mapping method is used for performance evaluation by mapping matrix multiplication on the proposed architecture. The asymptotic time complexities of them are evaluated and the result shows that paralle matrix multiplication on HMSA can provide significant performance improvement.

  • PDF

3상 계통에서 전류/전압과 순시전력간의 맵핑이론 (The Mapping Theory between Current/Voltage and Instantaneous Powers in Three-phase Systems)

  • 김효성;최재호
    • 전력전자학회:학술대회논문집
    • /
    • 전력전자학회 1997년도 전력전자학술대회 논문집
    • /
    • pp.228-232
    • /
    • 1997
  • The relation between instantaneous active/reactive powers and currents is defined by voltage mapping matrix in three-phase four-wire systems. Control strategies for an active filter without energy storage components are proposed on the basis of mapping matrices. It can compensate for the zero-sequence current, irrespectively of whether or not a zero-sequence voltage exists in a three-phase four-wire system.

  • PDF

A study on the effect of the condition number in the magnetic field mapping of the Air-Core solenoid

  • Huang, Li;Lee, Sangjin
    • 한국초전도ㆍ저온공학회논문지
    • /
    • 제17권2호
    • /
    • pp.31-35
    • /
    • 2015
  • Mapping is a useful tool in the magnetic field analysis and design. In some specific research area, such as the nuclear magnetic resonance (NMR) or the magnetic resonance imaging (MRI), it is important to map the magnetic field in the interesting space with high accuracy. In this paper, an indirect mapping method in the center volume of an air-core solenoid is presented, based on the solution of the Laplace's equation for the field. Through the mathematical analysis on the mapping calculation, we know that the condition number of the matrix, generated by the measurement points, can greatly affect the error of mapping result. Two different arrangement methods of the measurement points in field mapping are described in this paper: helical cylindrical line (HCL) method and parallel cylindrical line (PCL) method. According to the condition number, the HCL method is recommended to measure the field components using one probe. As a simple example, we mapped the magnetic fields in a MRI main magnet system. Comparing the results in the different methods, it is feasible and convenient to apply the condition number to reduce the error in the field mapping calculation. Finally, some guidelines were presented for the magnetic field mapping in the center volume of the air-core solenoid.