Skip to main content
EURASIP Journal on Audio, Speech, and Music Processing

Table 5 Word recognition accuracy [%] corresponding to each SNR

From: Unsupervised domain adaptation for lip reading based on cross-modal knowledge distillation

Method clean 30dB 20dB 10dB 0dB
Known class adap. 52.22 47.50 47.55 47.33 47.13
Unknown class adap. 52.04 47.18 47.21 47.10 46.75
  1. The results of “clean” correspond to those in Table 3
Back to article page