The embodiment of the invention discloses a voice recognition method and device, a storage medium, and electronic equipment. The method comprises the steps: obtaining first voice data; inputting the first voice data into a pre-built screening model for screening, and obtaining a voice segment which is outputted by the screening model, wherein preset voice features in the voice segment is filtered; recognizing the voice segment, and obtaining characters corresponding to the voice segment. According to the above technical scheme, the method can effectively reduce the calculation burden in a voice recognition process, and improves the recognition speed.

Voice recognition method and device, storage medium, and electronic equipment

The invention discloses a speaker recognition method based on multi-stream hierarchical fusion transform features and a long-short-term memory network The method comprises: a Filterbank feature and aMel frequency cepstral coefficient feature are extracted from a speech sample as two feature flows respectively; the two feature flows are inputted into two deep belief networks with bottleneck layers respectively to carry out feature conversion, thereby obtaining two bottleneck feature flows; the two bottleneck feature flows are spliced and the spliced feature flows are inputted into a third deep belief network with a bottleneck layer to carry out feature conversion, so that a feature after fusion conversion; and then a speaker to which the voice sample belongs is determined by using a long-short-term memory network According to the invention, fusion conversion is carried out on inputted acoustic features by using multiple deep belief networks; and compared with the single acoustic feature or the single feature after neural network conversion, the provided feature after fusion conversion is able to depict characteristic differences of different speakers, so that an excellent effectis obtained in speaker identification

Speaker recognition method based on multi-stream hierarchical fusion transform features and long-short-term memory network

The invention discloses a speech signal recognition method, device and apparatus. The method comprises: collecting an input speech signal, performing spectrum analysis on the speech signal, comparingthe spectrum analysis result with the spectrum data of the useless vocabulary in a preset useless vocabulary spectrum library, screening out the useless speech signal, deleting the useless speech signal in the speech signal, and performing character recognition on the rest of the speech signals. The method provided by the invention effectively solves the problems of low speech recognition speed and low working efficiency caused by the recognition of unwanted speech signals excluding control commands in the speech recognition technology.

Speech signal recognition method, device and apparatus

The invention is a divisional application of the patent 201710854125.4 and discloses a sample generation method, a model training method and device, a medium and electronic equipment. The sample generation method comprises the following steps that a sent short message input by a voice mode and/or a stored memorandum which is input through a voice mode are acquired; according to the text content ofthe short message and/or the memorandum, a standard voice data sample input by a user is acquired; the voice characteristic vector sequence of the standard voice data sample is obtained; according tothe historical voice data of the short message and/or the memorandum, the personalized voice characteristics without actual meanings and the occurrence positions of the voice characteristics are determined; and the voice characteristics are added into the voice characteristic vector sequence on the basis of the occurrence position to obtain a training sample of a screening model. By adopting thetechnical scheme, the calculation amount in the voice recognition process can be effectively reduced, and the recognition speed is increased.

Sample generation method, model training method and device, medium and electronic equipment

Disclosed are a speech recognition method and apparatus. The method comprises: acquiring a feature classification result of a speech signal to be recognized, wherein the feature classification result contains a pronunciation for describing a pronunciation feature of each speech signal frame and the probability of mapping each speech signal frame to a corresponding pronunciation (S201); if the probability of mapping any speech signal frame to a corresponding pronunciation satisfies a predetermined filtering rule, determining the pronunciation as a candidate pronunciation (S202); if the probability of mapping any of a predetermined number of speech signal frames adjacent to the speech signal frame to the candidate pronunciation satisfies the predetermined filtering rule, deleting the candidate pronunciation from pronunciations contained in the feature classification result (S203); if none of the probabilities of mapping the predetermined number of speech signal frames adjacent to the speech signal frame to the candidate pronunciation satisfies the predetermined filtering rule, retaining the candidate pronunciation in the pronunciations contained in the feature classification result (S204); and recognizing the speech signal based on a filtered feature classification result (S205). During the process of speech signal recognition, a recognition operation related to a pronunciation filtered out is not required to be executed, thereby improving the efficiency of speech recognition.

Speech recognition method and apparatus

The embodiments of the invention provide a speech recognition method and a speech recognition device. The method comprises the following steps: acquiring a characteristic classification result of a speech signal to be recognized, wherein the characteristic classification result includes pronunciations used to describe the pronunciation characteristics of speech signal frames and the probability that the speech signal frames are mapped to the corresponding pronunciations; removing the pronunciations contained in the characteristic classification result through filtering based on the probability contained in the characteristic classification result; and recognizing the speech signal based on the characteristic classification result after filtering. According to the embodiments of the invention, there is no need to implement recognition operation related to the removed pronunciations in the process of speech signal recognition, for example, there is no need to search a recognition network for a path related to the removed pronunciations. Therefore, the time consumed by the speech recognition process can be reduced, and the efficiency of speech recognition can be improved.

Li Zhongjie

Papers

Speech recognition method and speech recognition device

Speech recognition method and apparatus