Speech recognition control method and device

Abstract

The invention relates to the technical field of speech recognition and discloses a speech recognition control method and device, wherein the speech recognition control method comprises: storing voiceprint characteristics of a first audio signal acquired to a local voiceprint characteristic database, acquiring voiceprint characteristics of a second audio signal, matching the voiceprint characteristics of the second audio signal with voiceprint characteristics of each first audio signal in the local voiceprint characteristic database, and if it is determined that the voiceprint characteristics of the second audio signal gain are fully matched with those of any first audio signal in the local voiceprint characteristic database, transmitting corresponding control commands to a controlled device. The speech recognition control method and device have the advantages that the problem is effectively avoided that since existing speech recognition devices recognize speech through a networked language semantic database, response is too slow in places with poor network signal or zero WiFi and user experience is poor; the problem is effectively avoided that incorrect speech of local accent leadsto inaccurate recognition; therefore, speech recognition rate is effectively increased.

Description

technical field [0001] The invention relates to the technical field of voice recognition, in particular to a voice recognition control method and device. Background technique [0002] Existing speech recognition devices are widely used in home appliances, which can control televisions, VCDs, air conditioners, electric fans and curtains, etc. through speech recognition. At present, the existing speech recognition devices generally need to be connected to the Internet, so as to rely on the language and semantic database of the network for speech recognition. However, due to the large differences in dialects in various places, even if it is the same language, the pronunciation will also have different local accents. Therefore, As a result, it is difficult for existing speech recognition devices to distinguish audio signals of different local accents, resulting in a low recognition rate of existing speech recognition devices. Contents of the invention [0003] The object of t...

AI Technical Summary

Problems solved by technology

At present, the existing speech recognition devices generally need to be connected to the Internet, so as to rely on the language and semantic database of the network for speech recognition. However, due to the large differences in dialects in various places, even if it is the same language, the pronunciation will also have different local accents. Therefore, As a result, it is difficult for existing speech recognition devices to distinguish audio signals of different local accents, resulting in a low recognition rate of existing speech recognition devices

Images