Single-channel dereverberation by feature mapping using cascade neural networks for robust distant speaker identification and speech recognition

EURASIP Journal on Audio, Speech, and Music Processing

Table 4 Baseline for the speaker identification experiments using the real reverberant data

Experiment scheme	Method	Speaker identification rate (%)
Experiment scheme	Method	P01	P03	P05	P02	P04	Avg. known	Avg. unknown	Avg. all
I	Baseline	45.0	49.0	45.5	44.0	46.0	46.5	45.0	45.9
I	CMN	83.0	88.0	91.5	93.0	90.0	87.5	91.5	89.1
II	Baseline	45.0	50.5	54.0	55.5	49.5	49.8	52.5	50.9
II	CMN	84.0	89.5	91.0	91.5	89.5	88.2	90.5	89.1

The known environments include P01, P03, and P05, while the unknown environments include P02 and P04.