Speech recognition method based on isolated words and range hood using speech recognition method

Abstract

The invention relates to a speech recognition method based on isolated words. The speech recognition method comprises the following steps of training template speech to form a speech template library;comparing speech to be recognized with the speech template library and computing to recognize the speech to be recognized; extracting the speech of each isolated word in the template speech and the speech to be recognized by using an endpoint detection algorithm, and acquiring a feature parameter of each isolated word via computing; storing the feature parameters of all different isolated words in the template speech to form an isolated word speech feature parameter library; storing a feature parameter vector corresponding to each piece of template speech, and thus forming the speech templatelibrary; and acquiring the feature parameter vector corresponding to the speech to be recognized, judging whether the feature parameter vector corresponding to the speech to be recognized exists in the speech template library, and if yes, recognizing the speech to be recognized. According to the speech recognition method based on the isolated words, the data processing capacity can be greatly reduced. The range hood using the speech recognition method based on the isolated words is low in cost and high in speech command recognition rate.

Description

technical field [0001] The invention relates to the technical field of speech recognition, in particular to a speech recognition method based on isolated words, and to a range hood applied with the speech recognition method. Background technique [0002] With the vigorous development of voice recognition technology, more and more home appliances have voice control functions. In addition, due to the different areas of use, there are many dialects in various places, and the Mandarin of the local people is not standard. Existing speech chips often generate a fixed template library to recognize speech, which leads to the unsatisfactory effect of common speech recognition modules in recognizing dialects. Usually, this kind of semantic recognition template library requires an external storage module due to the large amount of data, which increases additional costs. [0003] Application publication number is CN106997762A (application number is 201710134617.6) Chinese invention pa...

AI Technical Summary

Problems solved by technology

However, the scheme only discloses the implementation principle of the method, and does not propose a specific implementation scheme.

In the specific operation, if the existing training, feature extraction and feature matching methods are used, there is still a large amount of data processing, and the cost is high, so it is not suitable for use in a large number of ordinary household appliances.

Images