The Research Of Segmented Audio Retrieval Algorithm Based On Audio Fingerprint

With the rapid development of the Internet,multimedia information,especially audio information,is growing explosively.The traditional text-based retrieval methods can not meet the people's demand for audio information retrieval.The research of audio retrieval technology based on audio signal is advancing rapidly,and especially the audio fingerprint retrieval technology is the hotspot of the research.Audio fingerprint is a digital summary extracted from the audio signal,and what it is compared to is its corresponding small digital fingerprint,rather than the big audio data itself directly.Therefore,the audio fingerprint search technology can not only greatly reduce the search volume,but also can significantly improve the retrieval efficiency.With the development of technology,its application scene has entered the music retrieval,copyright protection,advertising broadcast,television interaction and other fields.Therefore,it is of great significance to study the audio fingerprint retrieval technology.This paper is based on the Shazam algorithm,and through the analysis of the process of the audio fingerprint extraction,then proposes an improved audio fingerprint extraction algorithm,which improves the accuracy of audio retrieval.On the basis of improving the audio fingerprint extraction algorithm,a segmented audio retrieval algorithm is proposed,which ensures the accuracy of audio retrieval and greatly improves the retrieval speed.The main work is as follows:An audio fingerprinting extraction algorithm based on triangular combination is proposed.In this paper,the advantages and disadvantages of the audio fingerprint extraction process in Shazam algorithm are analyzed in detail.In the Shazam algorithm,the peak point pairs of the spectrum form audio fingerprints,and this paper optimize it as that an anchor point corresponding to two target peaks is combined to triangular combination and then forms the audio fingerprint.This can not only increase fingerprint information,reduce the amount of fingerprint extraction,but also enhance the robustness of audio fingerprints,thus it improves the accuracy and robustness of audio retrieval,and through simulation proves that the improved algorithm has higher performance of the search.A segmented audio retrieval method based on audio fingerprint is proposed.Based on the improved audio fingerprint algorithm,this method optimizes the process of the audio fingerprint extraction and match,and uses the idea of audio segmentation and matching threshold to segment the longer audio segments,and then extracts the audio fingerprints for segment and match them.If the match value is greater than the matching threshold,the search ends,and you do not have any processing on the remaining fragments,otherwise,you need to use other fragments.The improved method,while ensuring a relatively high accuracy,can greatly shorten the retrieval time,and finally through the simulation experiment it is also proved the conclusion.