Figure 8
From: Phone lattice reconstruction for embedded language recognition in LVCSR
EERs of A3 system as a function of percentage of links remaining after the pruning.
Figure 8 | EURASIP Journal on Audio, Speech, and Music Processing