Figure 7
From: Sparse coding of the modulation spectrum for noise-robust automatic speech recognition
State probability traces for the digit sequence ‘3 6 7’ at 5-dB SNR. (a) Traces obtained by using the activation weights of the full modulation spectrum exemplars only. The Viterbi decoder returns the incorrect sequence ‘3 3 7’. (b) Traces obtained from fusing the probability estimates obtained with the full modulation spectrum and the probability estimates obtained from nine modulation bands (weights obtained with a genetic algorithm). The Viterbi decoder now returns the correct sequence ‘3 6 7’.