Audio classification by comparison of feature sections and integrated features to known references

摘要

To classify moving images using audio signals. An audio signal is acquired, a section feature relating to an audio frequency distribution is extracted with respect to each of a plurality of sections each having a predetermined length contained in the acquired audio signal, each extracted section feature is compared with each of reference section features to calculate a section similarity indicating a degree of correlation between each section feature and each reference section feature. An integrated feature relating to the plurality of sections and being calculated based on the section similarity calculated with respect to each of the plurality of sections is extracted from the acquired audio signal. The extracted integrated feature is compared with each of one or more reference integrated features, and the audio signal is classified based on comparison result. Then, classification result is used for moving image classification.

主权项

1. An audio classification device comprising:
a processor; an acquisition unit operable to acquire an audio signal; a section feature extraction unit operable, with respect to each of a plurality of sections each having a predetermined length contained in the audio signal, to extract a section feature relating to an audio frequency distribution; a reference section feature storage unit that stores therein a plurality of reference section features that are each a reference for a comparison with each of the extracted section features; a calculation unit operable, with respect to each of the plurality of sections, to make a comparison between the section feature and each of the reference section features to calculate a section similarity indicating a degree of correlation between the section feature and each of the reference section features; an integrated feature extraction unit operable to extract, from the audio signal, an integrated feature relating to the plurality of sections, the integrated feature being calculated based on the section similarity calculated with respect to each of the plurality of sections; a reference integrated feature storage unit that stores therein one or more reference integrated features that are each a reference in a different category for a comparison with the integrated feature; and a classification unit operable to make a comparison between the integrated feature with each of the one or more reference integrated features, and classify the audio signal based on a result of the comparison, wherein the integrated feature and the reference integrated features are each composed of a combination of respective containing degrees indicating how much of the reference section features are contained therein, and the classification unit specifies, by the comparison, one of the reference integrated features that has a highest similarity to the integrated feature in terms of combination of containing degree, and classifies the audio signal to a category indicated by the reference integrated feature having the highest similarity.