Andreas Brinch Nielsen

Learn More
This paper explores the vocal and non-vocal music classification problem within popular songs. A newly built labeled database covering 147 popular songs is announced. It is designed for classifying signals from 1sec time windows. Features are selected for this particular task, in order to capture both the temporal correlations and the dependencies among the(More)
The amplitudes are normalized with the amplitude of the first harmonic. We call these features Harmonic Representation (HR). rKOPLS KOPLS projects data in " Kernel " space, where relevant (non-linear) features are extracted using linear OPLS. Assuming a sparse representation of the solution (rKOPLS), complexity of the algorithm can be kept under control.(More)
  • 1