<?xml version="1.0" standalone="yes"?>
<Paper uid="H94-1068">
  <Title>vIicrophone Arrays and Neural Networks for Robust Speech Recognition</Title>
  <Section position="10" start_page="345" end_page="345" type="concl">
    <SectionTitle>
7. CONCLUSION AND
DISCUSSION
</SectionTitle>
    <Paragraph position="0"> The above evaluation results suggest that the system of microphone array and neural network processors can * effectively mitigate environmental acoustic interference * without retraining the recognizer, elevate word recognition accuracies of HMM-based and/or DTW-based speech recognizers in variable acoustic environments to levels comparable to those obtained for close-talking, high-quality speech * achieve word recognition accuracies, under unmatched training and testing conditions, that exceed those obtained with a retrained speech recognizer using array speech for both retraining and testing, i.e., under em matched training and testing conditions Similar results have also been achieved for studies on speaker recognition \[9, 10\].</Paragraph>
    <Paragraph position="1"> In future work, we expect to extend the comparative evaluations of different neural network architectures, so that the performance of neural network equalization can be addressed in terms of word recognition accuracy. We also want to extend the evaluation experiments to continuous speech. For comparison, the DECIPHER system will be included, and possibly other advanced ARPA speech recognizers. The CAIP Center has concomitant NSF projects on developing 2-D and 3-D microphone arrays. These new array microphones have better spatial volume selectivity and can provide a high signal-to-noise ratio. They will be incorporated into this study. Further work will compare the system of microphone array and neural network with other existing noise compensation algorithms, such as Codebook Dependent Cepstrum Normalization (CDCN) \[17\] and Parallel Model Combination (PMC) \[18\].</Paragraph>
  </Section>
class="xml-element"></Paper>