Segmented quality inspection method and system for recorded data

Abstract

The invention provides a segmented quality inspection method and system for recorded data, which can be used in the technical field of cloud computing or other fields, and the method comprises the following steps: converting audio data into text data; performing similarity detection on the text data and a preset standard verbal skill template, and outputting a similarity; comparing the output similarity with a set threshold value, and if the similarity is larger than the set threshold value, judging that quality inspection of the audio data is qualified; converting the audio data into the text data, the text data and the preset standard verbal skill template being subjected to similarity detection, extra manpower does not need to be input for manual rechecking, and the service value of the quality inspection system is improved.

Description

technical field [0001] The present application relates to the technical field of cloud computing, and in particular to a method and system for segmented quality inspection of recorded data. Background technique [0002] In order to protect the rights and interests of consumers, regulatory agencies such as the China Banking and Insurance Regulatory Commission and the China Securities Regulatory Commission require financial institutions to record synchronous audio and video recordings of the sales process (abbreviation: double recording) when selling their own or consignment wealth management, funds, insurance and other products to prevent misleading sales, private Sales of "flying orders" and so on. According to regulatory requirements and their own business processes, financial institutions formulate the double-recording steps and standard speech templates for each step when selling related products, and check whether the double-recording videos meet the requirements one by ...

AI Technical Summary

Problems solved by technology

[0006] The disadvantage of this method is that, on the one hand, it relies on the uniqueness of the start and end keywords. When the words are short or the keywords of different quality inspection points are similar, the segmentation effect is not good; / The customer’s voice is not loud enough, the local accent is heavy, etc., coupled with the limitations of the speech recognition technology itself, resulting in some missing or errors in the converted text, which will also cause segmentation faults at this time, resulting in wrong quality inspection results. Additional manpower is required for manual review, which reduces the business value of the quality inspection system

[0007] (2) Video part: For each video quality inspection point, or rely on the segmentation results of the voice to check, at this time there are also situations where inaccurate segmentation leads to quality inspection errors; or do not implement segmentation, but double-record from the entire segment The video is inspected one by one after drawing frames into pictures. At this time, there are disadvantages of high computing resource consumption and long time-consuming quality inspection.

Images