Voice information batch screening method and device, computer equipment and storage medium

Abstract

The embodiment of the invention discloses a voice information batch screening method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining an input folderaddress of a folder where a preset training set is located if an information processing instruction is received; determining a preset threshold, a preset first output folder address and a preset second output folder address according to the information processing instruction; reading the input folder address to obtain all to-be-processed voice information; calling a preset voice screening script to respectively extract feature information of each piece of to-be-processed voice information; sequentially reading all the feature information to judge whether the feature information is matched witha preset threshold value or not; and if yes, storing the to-be-tested voice information corresponding to the feature information into a second output folder corresponding to the second output folderaddress for batch output. According to the invention, unified screening of multiple pieces of to-be-processed voice information in the training set can be realized efficiently and accurately, and errors in the screening process are reduced.

Description

technical field [0001] The invention relates to the field of data processing, in particular to a batch screening method, device, computer equipment and storage medium for voice information. Background technique [0002] In speech recognition projects, it is usually necessary to collect or collect a large amount of speech information from various channels, and use these speech information as training samples in the training set to train the neural network, so as to obtain the corresponding recognition model for feature speech recognition. In order to ensure the smooth training process of the neural network and the accuracy of the acquired recognition model, it is usually necessary to pre-process the acquired voice information before training, such as screening effective voice information, etc., so as to realize the recognition of a large amount of voice information. The preprocessing work needs to be iterated step by step to complete, but the process of repeated iterative pro...

AI Technical Summary

Problems solved by technology

In order to ensure the smooth training process of the neural network and the accuracy of the acquired recognition model, it is usually necessary to pre-process the acquired voice information before training, such as screening effective voice information, etc., so as to realize the recognition of a large amount of voice information. The preprocessing work needs to be iterated step by step to complete, but the process of repeated iterative processing is very prone to operational errors due to the large amount of data, resulting in inaccurate voice information screening

Images