Improving dysarthric speech recognition using empirical mode decomposition and convolutional neural network

EURASIP Journal on Audio, Speech, and Music Processing

Table 1 Results of enhancement techniques using a sentence from dysarthric speaker JF extracted from the Nemours corpus

	Speech enhancement technique
Metrics	Spectral subtraction	Wiener filtering	EMDH
SNRseg	2.4146	11.1018	15.8155
PESQ	2.1160	3.2983	4.3640