Hwa Jeon Song

A new method for speaker adaptation using bilinear model

In this paper, a novel method for speaker adaptation using bilinear model is proposed. Bilinear model can express both characteristics of speakers (style) and phonemes across speakers (content) independently in a training database. The mapping from each speaker and phoneme space to observation space is carried out using bilinear mapping matrix which is independent of speaker and phoneme space. We apply the bilinear model to speaker adaption. Using adaptation data from a new speaker, speaker-adapted model is built by estimating the style(speaker)-specific matrix. Experimental results showed that the proposed method outperformed eigenvoice and MLLR. In vocabulary-independent isolated word recognition for speaker adaptation, bilinear model reduced word error rate by about 38% and about 10% compared to eigenvoice and MLLR respectively using 50 words for adaptation.

Speech/Music Discrimination for Robust Speech Recognition in Robots

Automatic speech recognition (ASR) is one indispensable technology to communicate with a service robot. In real-world environments, ASR faces many kinds of sound sources and they should be discriminated to improve ASR performance. In ASR systems, speech is usually detected from the input signal by voice activity detection (VAD) scheme. Speech and music, how ever, are not easily discriminated by the VAD because they share similar characteristics such as periodicity. In this paper, we adopt a speech/music discriminator into the front-end of the ASR system in order to disable music stream not to be an input for the ASR system. Our speech/music discriminator employs the mean of minimum cepstral distances (MMCD) as a feature parameter. Experimental result shows the MMCD parameter outperforms the conventional feature parameter, spectral flux.

Bilinear Model-Based Maximum Likelihood Linear Regression Speaker Adaptation Framework

This letter proposes a novel framework for speaker adaptation, using bilinear model-based maximum likelihood linear regression (MLLR) method. First, a set of speaker models is decomposed into the style factor identified as each speakers characteristics and the common content factor across the speakers, by the bilinear model. Then, using some adaptation data from a new speaker, the speaker-specific model is generated by properly adjusting the dimensionality of the content factor and estimating a new style factor simultaneously. Experimental results show that the proposed framework outperforms MLLR with fewer number of parameters to be estimated.

Implementation of CNN in the view of mini-batch DNN training for efficient second order optimization

This paper describes some implementation schemes of CNN in view of mini-batch DNN training for efficient second order optimization. This uses same procedure updating parameters of DNN to train parameters of CNN by simply arranging an input image as a sequence of local patches, which is actually equivalent with mini-batch DNN training. Through this conversion, second order optimization providing higher performance can be simply conducted to train the parameters of CNN. In both results of image recognition on MNIST DB and syllable automatic speech recognition, our proposed scheme for CNN implementation shows better performance than one based on DNN.