Frame accuracies with multi-resolution features and FC-sigmoid-RBM-pretrain DNNs with hidden layers of 2048 units.

In all cases features are MFCC+Splice+LDA+fMLLR. Input Dim. is the dimension of the input of the network including feature splicing. Param. is the number of trainable parameters of the network.所有场景下的特征均采用梅尔频率倒谱系数（Mel-Frequency Cepstral Coefficients，MFCC）+特征拼接（Splice）+线性判别分析（Linear Discriminant Analysis，LDA）+特征最大似然线性回归（feature Maximum Likelihood Linear Regression，fMLLR）组合得到。输入维度（Input Dim.）指包含特征拼接操作在内的网络输入维度。参数数量（Param.）指网络的可训练参数量。