Abstract
In this paper, we deem to utilize the specifics provided by a modulated frequency features for Query by Humming (QBH) music retrieval system based on hum. Initially music signal is transformed to an abstract domain using Empirical Mode Decomposition (EMD). Then, we impart the dimension reduced outcome as precious source for feature extraction process using a Modulation Frequency (MF) criterion. In order to evaluate the suggested approach, experiments are conducted on a database of 1495 song fragments which are manually extracted from 1200 songs and 200 hum recordings. The proposed approach effectively retrieves the desired song based on Humming Query (HQ) and affirms the importance of EMD and MF feature space.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call