Speech recognition apparatus, speech recognition method, and speech recognition program

摘要

A apparatus includes: a storage unit to store a model representing a relationship between a relative time and an occurrence probabilities; a first detection unit to detect first speech period of a first speaker; a second period detection unit to detect second speech period of a second speaker; a unit to calculate a feature value of the first speech period; a detection unit to detect a word using the calculated feature value; an adjustment unit to make an adjustment such that in detecting a word for a reply by the detection unit, the adjustment unit retrieves an occurrence probability corresponding to a relative position of the reply in the second speech period, and adjusts a word score or a detection threshold value for the reply; and a second detection unit to re-detect, using the adjusted word score or the adjusted detection threshold value, the detected word by the detection unit.

主权项

1. A speech recognition apparatus comprising:
a reply probability storage unit configured to store a reply probability model representing a relationship between a relative time defined with respect to a speech period of one speaker and an occurrence probabilities of a reply occurring in an utterance of another different speaker; a first speech period detection unit configured to detect an speech period of a first speaker from a voice of the first speaker; a second speech period detection unit configured to detect an speech period of a second speaker from a voice of the second speaker different from the first speaker; a feature value calculation unit configured to calculate a feature value of the speech period of the first speaker detected by the first speech period detection unit; a first detection unit configured to detect a word using the feature value of the speech period of the first speaker calculated by the feature value calculation unit; an adjustment unit configured to make an adjustment such that in a case where the word detected by the first detection unit is a reply, the adjustment unit refers to the reply probability model stored in the reply probability storage unit to retrieve an occurrence probability corresponding to a relative position of the reply with respect to the speech period of the second speaker detected by the second speech period detection unit, the adjustment unit adjusts a word score for the reply or a detection threshold value for the reply depending on the retrieved occurrence probability; and a second detection unit configured to perform a re-detection using the word score for the reply or the detection threshold value for the reply adjusted by the adjustment unit, in terms of the word detected by the first detection unit.