Table 3.

Average classification accuracy (in %) of time-frequency image representations.

Signal Representation	Sound Event		Speech Command
	Validation	Test	Validation	Test
Spectrogram	92.70	93.77	92.33	91.90
Smoothed-Spectrogram	96.48	97.32	93.79	93.41
Mel-Spectrogram	96.45	96.31	93.64	93.64
Cochleagram	98.35	98.61	94.33	94.13