Sound environment recognition method based on background noise minimum statistic feature

Abstract

The invention provides a sound environment recognition method based on a background noise minimum statistic feature, and relates to a classified recognition technique of sound scenes. The method aims to solve the problems that in sound environment recognition, due to the facts that a sound environment is complicated in structure and easy to mix, and acoustic features are not determined and are high in dimensionality, feature extraction and statistic model building are difficult, and recognition performance is poor. The method for extraction, model building and recognition of a noise minimum statistic feature is put forward. According to the method, in the stage of feature extraction and model building, frequency-domain smoothness and time-domain smoothness are carried out on an energy spectrum of a sound signal, a minimum statistic is tracked, the minimum statistic is converted to a logarithm domain to carry out standardized and dimensionality-reduction processing, and accordingly the noise minimum statistic feature of the sound environment is extracted, and model building is carried out on the noise minimum statistic by means of a Gaussian mixing model; in the recognition stage, feature extraction is carried out on input sounds, the likelihood value of the extracted minimum statistic feature under each model is calculated and category decision making is carried out on the likelihood values.

Description

technical field [0001] The invention relates to the classification and recognition technology of sound scenes, belongs to the field of intelligent information processing, and specifically relates to a method for modeling and recognizing different sound environments based on the background noise characteristics of the sound environment. Background technique [0002] With the rapid development of computing technology, communication technology and network technology, a variety of intelligent devices continue to emerge, and the study of intelligent human-computer interaction is becoming more and more important in real life. Sound is one of the most important information that people can obtain. Research on sound perception technology will undoubtedly help users effectively use sound information and provide related services. Therefore, in recent years, the research on sound perception technology has received extensive attention from the academic community. The goal of sound perce...

AI Technical Summary

Problems solved by technology

[0007] The present invention aims to solve the problem of poor recognition performance due to the complex sound environment structure, easy confusion, uncertain acoustic features, and high dimensionality in sound environment recognition, such as the difficulty in extracting effective features, the difficulty in establishing a statistical model, and poor feature robustness.

Images