Correcting speech recognition errors on a smartphone is a challenging task that requires a lot of user effort. To reduce this user effort, we previously proposed an error correction method based on Long Context Match (LCM) with higher-order N-grams, which we combined with a simple gesture-based user interface. However, LCM was used when there was only one substitution error exists in an utterance. In this paper, we extended LCM to be used when the contexts of the error region contain errors. We examined the error regions which contain only one error and confirmed the effectiveness of the extended LCM-based method.
各種検索
サポート
ヘルプ