<?xml version="1.0" standalone="yes"?>
<Paper uid="H92-1055">
  <Title>MULTIPLE APPROACHES TO ROBUST SPEECH RECOGNITION</Title>
  <Section position="7" start_page="277" end_page="278" type="concl">
    <SectionTitle>
5. SUMMARY AND CONCLUSIONS
</SectionTitle>
    <Paragraph position="0"> In this paper we describe our current research in acoustical pre-processing for robust speech recognition, as well as our first attempts to integrate pre-processing with other approaches to robust speech recognition. The CDCN algorithm was also applied to the ATIS task for the first time, and provided the best recognition scores for speech collected using the unidirectional desktop PCC160 microphone. We demonstrated that the CDCN algorithm and the Flanagan delay-and-sum microphone array can provide complementary benefits to speech recognition in reverberant environments. We also found that the Seneff auditory model improves recognition accuracy of the CMU speech system in reverberant as well as noisy environ- null ments, but preliminary efforts to combine the auditory model with the CDCN algorithm were inconclusive.</Paragraph>
  </Section>
class="xml-element"></Paper>