Statistical Parametric Speech Synthesis Using Deep Gaussian Processes

utteranceVOCPIC-GPDNNDGP
A
B
C
D
E


MethodsMeaning
VOCVocoder speech by STRAIGHT.
PIC-GPHMM-Gaussian process hybrid model using partially independent conditional approximation.
DNNDeep neural network.
DGPProposed method based on deep Gaussian processes.