A New Approximate Matching Algorithm and Its Application in Internet Music Search by Humming

摘要

This paper has proposed a new approximate matching algorithm-similarity matching, and use the characteristics of algorithm to establish a system of internet music search by humming. The author compared the similarity matching algorithm and dynamic time warping (DTW) algorithm, which is most commonly used to query by humming, by the system of internet music search by humming. On the two standard of the query hit ratio and query speed, we got the result that similarity matching algorithm's comprehensive efficiency is superior, is one of QBH (query by humming) algorithm, which is applicable to the large-scale music library such as internet music search.