Phonetic speech retrieval is often used to augment word based retrieval in spoken document retrieval systems. However, a known problem with phonetic word spotting is the alamingly high number of false positives as the collection scales up. In this paper, we address this issue with the use of a new indexing and ranking scheme using stochastic phonetic edit distance, and an improved thresholding algorithm. We conduct an extensive set of experiments using a large unbiased query set, and show the improved accuracy when phonetic retrieval is combined with speech recognition retrieval. Using a hundred hours of HUE34 data with ground truth transcript as a benchmark, and several thousands query words, we show improvement of up to 15% in precision compare
to speech recognition alone.
By: Amon Amir, Alon Efrat, Savitha Srinivasan
Published in: RJ10215 in 2001
LIMITED DISTRIBUTION NOTICE:
This Research Report is available. This report has been submitted for publication outside of IBM and will probably be copyrighted if accepted for publication. It has been issued as a Research Report for early dissemination of its contents. In view of the transfer of copyright to the outside publisher, its distribution outside of IBM prior to publication should be limited to peer communications and specific requests. After outside publication, requests should be filled only by reprints or legally obtained copies of the article (e.g., payment of royalties). I have read and understand this notice and am a member of the scientific community outside or inside of IBM seeking a single copy only.
Questions about this service can be mailed to reports@us.ibm.com .