中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

41 records (took 0.024 seconds)

    METHOD, APPARATUS, AND RADIO FOR OPTIMIZING HIDDEN MARKOV MODEL SPEECH RECOGNITION
    36.
    发明专利
    METHOD, APPARATUS, AND RADIO FOR OPTIMIZING HIDDEN MARKOV MODEL SPEECH RECOGNITION 未知

    公开(公告)号：CA2189249A1

    公开(公告)日：1996-10-03

    申请号：CA2189249

    申请日：1996-01-29

    Applicant: MOTOROLA INC

    Inventor： KUSHNER WILLIAM M ,  SRENGER EDWARD ,  HARTMAN MATTHEW A

    IPC: G10L15/14 ,  G10L9/00

    Abstract: In a statistical based speech recognition system, one of the key issues is the selection of the Hidden Markov Model that best matches a given sequence of feature observations. The problem is usually addressed by the calculation of the maximum likelihood, ML, state sequence by means of a Viterbi or other decoder. Noise or inadequate training can produce an ML sequence associated with a Hidden Markov Model other than the correct model. The method of the present invention provides improved robustness by combining the standard ML state sequence score (416) with an additional path core (418) derived from the dynamics of the ML score as a function of time. These two scores, when combined, form a hybrid metric (420) that, when used with the decoder, optimizes selection of the correct Hidden Markov Model (422).

    Method of storing reflection coefficient vectors
    38.
    发明专利
    Method of storing reflection coefficient vectors 未知

    公开(公告)号：FR2709366A1

    公开(公告)日：1995-03-03

    申请号：FR9410202

    申请日：1994-08-23

    Applicant: MOTOROLA INC

    Inventor： GERSON IRA A ,  JASIUK MARK A ,  HARTMAN MATTHEW A

    IPC: G10L19/06 ,  G10L25/06 ,  G10L9/14 ,  G10L9/18 ,  H04B7/00 ,  H04B7/24

    Abstract: Speech coder with Vector-Sum Excited Linear Predictive (VSELP) Coding having enhanced quality and reduced complexity with respect to the prior art. The VSELP uses a coding table having a predefined structure such that the calculations required for the process of searching in the coding table can be reduced significantly, as well as a vector quantifier with one or more segments of the reflection coefficients on the basis of a Fixed-Point Meshing Technique (FLAT). Moreover, this speech coder uses a prequantifier so as to reduce the complexity of searching in the vector coding table, and a high-resolution scalar quantifier so as to reduce the quantity of storage memory in the reflection coefficient vector coding tables. Thus a high-quality speech coder is obtained with reduced calculations and storage conditions.

Patent Agency Ranking