AUGENVERFOLGUNG FÜR KONTEXTABHÄNGIGE SPRACHERKENNUNG

摘要

A method for discriminating between an instance of a voice command and an instance of speech dictation can include identifying a focus point in a user interface; defining a surrounding region about the focus point; identifying user interface objects in the surrounding region; further identifying among the identified user interface objects those user interface objects which are configured to accept speech dictated text and those user interface objects which are not configured to accept speech dictated text; computing a probability based upon those user interface objects which have been further identified as being configured to accept speech dictated text and those user interface objects which have been further identified as not being configured to accept speech dictated text; receiving speech input; and, biasing a determination of whether the speech input is a voice command or speech dictation based upon the computed probability. Additionally, the method can include identifying a focus point outside of the user interface; and, biasing a determination of whether the speech input is a voice command or speech dictation based upon a default probability.