本文介绍了在语音关键词识别的处理方法,对大家解决问题具有一定的参考价值,需要的朋友们下面随着小编来一起学习吧!

问题描述

是任何人都知道一个关键词识别系统的可免费获得,并有可能提供的API ??

Is anyone aware of a Keyword Spotting System that is freely available, and possibly providing APIs ??

CMU斯芬克斯4和MS语音API是语音识别引擎,并不能用于KWS

CMU Sphinx 4 and MS Speech API are speech recognition engines, and cannot be used for KWS.

有一个关键词识别系统,但无有效链接,即使不评价。 (我甚至找不到任何地方的链接与他们联系,为他们的软件)

SRI has a keyword spotting system, but no download links, not even for evaluation. (I even couldn't find anywhere a link to contact them for their software)

我发现一个但它是一个演示和有限的。

I found one here but it's a demo and limited.

推荐答案

CMUSphinx实现了pocketsphinx引擎关键词识别,详情请见:

CMUSphinx implements keyword spotting in pocketsphinx engine, see for details:

要认识到一个单一的关键词就可以去codeR中的关键词的搜索的模式下运行。

To recognize a single keyphrase you can run decoder in "keyphrase search" mode.

从命令行尝试:

pocketsphinx_continuous -infile file.wav -keyphrase "oh mighty computer" -kws_threshold 1e-20

从code:

 ps_set_keyphrase(ps, "keyphrase_search", "oh mighty computer");
 ps_set_search(ps, "keyphrase_search);
 ps_start_utt();
 /* process data */

您还可以找到我们的来源为Python和Android / Java的例子。

You can also find examples for Python and Android/Java in our sources.

阈值必须调整对测试数据的每个关键词短语,以获得正确的平衡漏检和误报。你可以尝试像1E-5值1E-50。

Threshold must be tuned for every keyphrase on a test data to get the right balance missed detections and false alarms. You can try values like 1e-5 to 1e-50.

有关最佳精度,最好是有关键词的3-4个音节。太短短语容易混淆。

For the best accuracy it is better to have keyphrase with 3-4 syllables. Too short phrases are easily confused.

您还可以搜索多个关键词的,创建一个文件keyphrase.list是这样的:

You can also search for multiple keyphrase, create a file keyphrase.list like this:

  oh mighty computer /1e-40/
  hello world /1e-30/
  other_phrase /other_phrase_threshold/

和在去codeR与-kws配置选项中使用它。

And use it in decoder with -kws configuration option.

  pocketsphinx_continuous -inmic yes -kws keyphrase_list

此功能尚未在sphinx4德codeR实施。

This feature is not yet implemented in sphinx4 decoder.

这篇关于在语音关键词识别的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持!

07-08 16:45