Abnormal emotion automatic detection and extraction method and system on basis of short-time analysis

Abstract

The invention discloses an abnormal emotion automatic detection and extraction method and an abnormal emotion automatic detection and extraction system on the basis of the short-time analysis. The method comprises the following steps of: extracting an emotion characteristic sequence from a voice signal to be detected; calculating the likelihood of the emotion characteristic sequence and an abnormal emotion model in a preset emotion model and calculating the likelihood of the emotion characteristic sequence and a non-abnormal emotion model in the preset emotion model; according to the likelihood of the emotion characteristic sequence and the abnormal emotion model and the likelihood of the emotion characteristic sequence and the non-abnormal emotion model, calculating the likelihood ratio; and judging whether the likelihood ratio is greater than a set threshold value, determining the voice signal to be detected is abnormal emotion voice if yes, or determining the voice signal to be detected is a non-abnormal voice signal. Due to the utilization of the abnormal emotion automatic detection and extraction method and the abnormal emotion automatic detection and extraction system, the automatic high-efficiency judgment on the abnormal emotion in the voice signal can be implemented and the automatic processing efficiency of mass customer service data is improved.

Description

technical field [0001] The present invention relates to a method and system for pattern recognition and signal detection, in particular to a method and system for automatic detection and extraction of abnormal emotions based on short-term analysis. Background technique [0002] Emotion refers to a strong emotional state caused by the subjective, and is often accompanied by psychological changes. In the field of customer service in the actual call center, banking, medical and other service industries, as the most direct face of customers, customer service personnel are not only the most direct medium for the transmission of customer emotions, but their negative emotions will also directly affect effective communication with customers And it is very likely to generate unnecessary complaints. Effective supervision and management of customer service quality can detect problems in time, sum up experience, make suggestions, and then urge improvement to improve service quality. ...

AI Technical Summary

Problems solved by technology

[0007] First of all, the emotional features related to the statistics extracted in the traditional algorithm weaken the characteristics of short-term changes in speech emotions and destroy the true distribution of the feature parameters themselves.

Especially in real speech, speech segments with abnormal emotions often only account for a small part of the speech to be measured, so the characteristics based on statistics make other silence, noise and non-abnormal emotional speech segments greatly offset a small amount of abnormal emotions The specific characteristics of the speech segment lead to a further reduction in feature discrimination, which in turn causes a sharp drop in detection effect

[0008] Secondly, the method of training the emotional model based on manually labeled data, in the actual application environment where the amount of abnormal emotional voice data is small, the model is difficult to accurately simulate the real distribution, which affects the system performance

[0009] Finally, in the traditional pattern matching algorithm, the extracted emotional features are matched with multiple emotional models preset by the system to calculate their similarity, and the model with the maximum likelihood is selected as the matching object. Abnormal emotions only occupy a small proportion of the speech segment to be detected. When , because the speech segment with strong distinguishing effect is too short, the likelihood score will be dominated by the features of the non-abnormal emotional speech segment, resulting in wrong judgment and affecting the effect of abnormal emotion detection

Images