Patent №
US 9,704,505
Granted
2017-07-11
Filed 2014
Owner
CANON KABUSHIKI KAISHA
Lab
—
AI components
2
speech · hardware
Assignment
Recorded
Dataset
AIPD
2023_r1 edition
Application
14527682
An audio separating apparatus obtains a matrix by performing time-frequency conversion on an input audio signal. The audio separating apparatus divides the obtained matrix into at least a basis matrix and an activity matrix, and classifies base spectra that configure the respective columns of the basis matrix into first base spectra corresponding to a target sound and second base spectra corresponding to a non-target sound.
AI classification
Ownership
CANON KABUSHIKI KAISHA
assignment · 356230283
Assignors
TAWADA, NORIAKI
On an employer assignment, the assignors are typically the inventors.