利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

16 条记录 (耗时 0.026 秒)

    Word dependent N-best search method
    1.
    发明授权
    Word dependent N-best search method 失效
    词依赖N最佳搜索方法

    公开(公告)号：US5241619A

    公开(公告)日：1993-08-31

    申请号：US720652

    申请日：1991-06-25

    申请人： Richard M. Schwartz ,  Stephen C. Austin

    发明人： Richard M. Schwartz ,  Stephen C. Austin

    IPC分类号： G10L15/08 ,  G10L15/14 ,  G10L15/18

    CPC分类号： G10L15/08 ,  G10L15/142 ,  G10L15/197

    摘要： As a step in finding the one most likely word sequence in a spoken language system, an N-best search is conducted to find the N most likely sentence hypotheses. During the search, word theories are distinguished based only on the one previous word. At each state within a word, the total probability is calculated for each of a few previous words. At the end of each word, the probability score is recorded for each previous word theory, together with the name of the previous word. At the end of the sentence, a recursive traceback is performed to derive the list of the N best sentences.

    摘要翻译： 作为在口语系统中找到最可能的单词序列的步骤，进行N最佳搜索以找到N个最可能的句子假设。 在搜索期间，词理论仅基于前一个单词来区分。 在一个单词内的每个状态下，为每个前几个单词计算总概率。 在每个单词的结尾处，对于每个前一个单词理论以及前一个单词的名称记录概率得分。 在句子的末尾，执行递归追溯以导出N个最佳句子的列表。

    Single tree method for grammar directed, very large vocabulary speech recognizer
    2.
    发明授权
    Single tree method for grammar directed, very large vocabulary speech recognizer 失效
    单树方法用于语法指导，非常大的词汇语音识别器

    公开(公告)号：US5621859A

    公开(公告)日：1997-04-15

    申请号：US183719

    申请日：1994-01-19

    申请人： Richard M. Schwartz ,  Long Nguyen

    发明人： Richard M. Schwartz ,  Long Nguyen

    IPC分类号： G10L15/02 ,  G10L15/14 ,  G10L15/18 ,  G10L5/06 ,  G10L9/00

    CPC分类号： G10L15/142 ,  G10L15/197

    摘要： The invention provides a method of large vocabulary speech recognition that employs a single tree-structured phonetic hidden Markov model (HMM) at each frame of a time-synchronous process. A grammar probability is utilized upon recognition of each phoneme of a word, before recognition of the entire word is complete. Thus, grammar probabilities are exploited as early as possible during recognition of a word. At each frame of the recognition process, a grammar probability is determined for the transition from the most likely preceding grammar state to a set of words that share at least one common phoneme. The grammar probability is combined with accumulating phonetic evidence to provide a measure of the likelihood that a state in the HMM will lead to the word most likely to have been spoken. In a preferred embodiment, phonetic context information is exploited, even before the complete context of a phoneme is known. Instead of an exact triphone model, wherein the phonemes previous and subsequent to a phoneme are considered, a composite triphone model is used that exploits partial phonetic context information to provide a phonetic model that is more accurate than aphonetic model that ignores context. In another preferred embodiment, the single phonetic tree method is used as the forward pass of a forward/backward recognition process, wherein the backward pass employs a recognition process other than the single phonetic tree method.

    摘要翻译： 本发明提供了一种在时间同步过程的每个帧处采用单个树结构语音隐马尔可夫模型（HMM）的大词汇语音识别方法。 在识别整个单词完成之前，在识别单词的每个音素时使用语法概率。 因此，在识别单词时尽可能早地利用语法概率。 在识别过程的每个帧处，确定从最可能的前一语法状态到共享至少一个公共音素的一组单词的转换的语法概率。 语法概率与积累的语音证据相结合，以提供HMM中的状态将导致最有可能发音的单词的可能性的量度。 在优选实施例中，甚至在已知音素的完整语境之前利用语音上下文信息。 考虑到音素之前和之后的音素被认为是一种精确的三音节模型，而是使用复合三音模型，它利用部分语音上下文信息来提供比忽略上下文的无声模型更准确的语音模型。 在另一个优选实施例中，使用单个语音树方法作为前向/后向识别过程的前向，其中后向通过采用除了单个语音树方法之外的识别过程。

    Information retrieval system
    4.
    发明授权
    Information retrieval system 失效
    信息检索系统

    公开(公告)号：US06405188B1

    公开(公告)日：2002-06-11

    申请号：US09127685

    申请日：1998-07-31

    申请人： Richard M. Schwartz ,  David R. H. Miller ,  Timothy R. Leek

    发明人： Richard M. Schwartz ,  David R. H. Miller ,  Timothy R. Leek

    IPC分类号： G06F1730

    CPC分类号： G06F17/30687 ,  Y10S707/99932 ,  Y10S707/99933 ,  Y10S707/99934 ,  Y10S707/99935 ,  Y10S707/99936 ,  Y10S707/99945

    摘要： Methods and systems for providing an improved IR system that performs information retrieval by using probabilities. When performing information retrieval, the improved IR system utilizes both the prior probability that a document is relevant independent of the query as well as the probability that the query was generated by a particular document given that the particular document is relevant. By using these probabilities, the improved IR system retrieves documents in a more accurate manner than conventional systems which are based on an ad hoc approach.

    摘要翻译： 用于提供通过使用概率执行信息检索的改进的IR系统的方法和系统。 当执行信息检索时，改进的IR系统利用了独立于查询的文档相关的先验概率，以及考虑到特定文档是相关的，由特定文档生成查询的概率。 通过使用这些概率，改进的IR系统以比基于特别方法的传统系统更准确的方式检索文档。

    Client/server speech processor/recognizer
    6.
    发明授权
    Client/server speech processor/recognizer 失效
    客户端/服务器语音处理器/识别器

    公开(公告)号：US5960399A

    公开(公告)日：1999-09-28

    申请号：US997912

    申请日：1997-12-24

    申请人： Christopher B. Barclay ,  Richard M. Schwartz ,  David G. Stallard ,  Lawrence R. Smith

    发明人： Christopher B. Barclay ,  Richard M. Schwartz ,  David G. Stallard ,  Lawrence R. Smith

    IPC分类号： G10L15/02 ,  G10L15/28 ,  H04L29/06 ,  H04L29/08 ,  H04M3/493 ,  G10L9/06

    CPC分类号： H04L65/602 ,  G10L15/02 ,  H04L29/06 ,  H04M3/4938 ,  G10L15/30 ,  H04L29/06027 ,  H04L67/42 ,  H04L69/329 ,  H04M2201/40

    摘要： A real-time or streaming speech processing system and method is disclosed with capabilities distributed between and client and a server where the server may be reached via the Internet. The speech processing entails digitizing and converting the utterances to features extracted to help the processing. The features are sent via a communications channel to the server where the recognition occurs. The features extracted allow low bandwidth channels to be used with still maintaining real-time response. The recognizer will determine the most likely text representing the utterances and return the text to the client. The system can be used to identify and/or verify who is speaking.

    摘要翻译： 公开了一种实时或流式语音处理系统和方法，其具有分布在客户端之间的能力和可以经由因特网到达服务器的服务器。 语音处理需要将话语数字化并转换为提取的特征以帮助处理。 这些功能通过通信通道发送到发生识别的服务器。 提取的功能允许使用低带宽通道，仍然保持实时响应。 识别器将确定表示话语的最可能的文本，并将文本返回给客户端。 该系统可用于识别和/或验证谁在说话。

    Language-independent and segmentation-free optical character recognition system and method
    10.
    发明授权
    Language-independent and segmentation-free optical character recognition system and method 失效
    语言无关和无分割的光学字符识别系统和方法

    公开(公告)号：US5933525A

    公开(公告)日：1999-08-03

    申请号：US630162

    申请日：1996-04-10

    申请人： John I. Makhoul ,  Richard M. Schwartz

    发明人： John I. Makhoul ,  Richard M. Schwartz

    IPC分类号： G06K9/00 ,  G06K9/48 ,  G06K9/80

    CPC分类号： G06K9/00879

    摘要： A language-independent and segment free OCR system and method comprises a unique feature extraction approach which represents two dimensional data relating to OCR as one independent variable (specifically the position within a line of text in the direction of the line) so that the same CSR technology based on HMMs can be adapted in a straightforward manner to recognize optical characters. After a line finding stage, followed by a simple feature-extraction stage, the system can utilize a commercially available CSR system, with little or no modification, to perform the recognition of text by and training of the system. The whole system, including the feature extraction, training, and recognition components, are designed to be independent of the script or language of the text being recognized. The language-dependent parts of the system are confined to the lexicon and training data. Furthermore, the method of recognition does not require pre-segmentation of the data at the character and/or word levels, neither for training nor for recognition. In addition, a language model can be used to enhance system performance as an integral part of the recognition process and not as a post-process, as is commonly done with spell checking, for example.

    摘要翻译： 一种独立于语言的和无分段的OCR系统和方法包括一种独特的特征提取方法，其将与OCR相关的二维数据表示为一个独立变量（具体地说是在线的方向上的文本行内的位置），使得相同的CSR 基于HMM的技术可以以直接的方式进行调整，以识别光学字符。 经过寻线阶段，随后简单的特征提取阶段，系统可以利用市售的CSR系统，很少或根本没有修改，可以对系统进行文本识别和培训。 整个系统，包括特征提取，训练和识别组件，被设计为独立于被识别的文本的脚本或语言。 系统的语言相关部分仅限于词典和训练数据。 此外，识别方法不需要在字符和/或词级别对数据进行预分割，既不用于训练也不用于识别。 另外，语言模型可以用来提高系​​统性能，作为识别过程的一个组成部分，而不是作为后处理，例如通常用拼写检查来完成的。