Method and Apparatus for Removing Speech Components in Sound Source

摘要

An embodiment of the present invention utilizes deep learning in removing vocal components in sound source. The present invention relates to a method and apparatus for removing voice components in sound source capable of improving the accuracy for learning by applying different learning models according to attribute information of a sound source signal and calculating the sound source from which voice is removed with more improved quality by primarily receiving the sound source in which vocal components are removed.