Fast fuzzy keyword spotting using syllable confusion network indexing

This paper presents a fast fuzzy search algorithm to extract keyword candidates from Syllable confusion networks (SCNs) in Mandarin spontaneous speech. Since the recognition accuracy of spontaneous speech is quite poor, Syllable confusion matrix (SCM) is applied to compensate for the recognition errors and to improve recall. In order to scale up to large collections and support quick query response, an efficient vocabulary-independent index structure is designed, which selects individual arcs of syllable confusion network as indexing unit. An inverted search algorithm that use syllable confusion matrix to calculate relevance score and search in this index structure is proposed. In experiments performed on a telephone conversational task, the Equal error rate (EER) was reduced by about 33% relative over the baseline where keywords are directly extracted from phoneme lattices. Additionally, it only took computer one or two seconds to search 100 keywords in one hour speech data.