CPC G10L 15/187 (2013.01) [G10L 15/02 (2013.01); G10L 15/22 (2013.01); G10L 2015/025 (2013.01); G10L 2015/223 (2013.01)] | 10 Claims |
1. A speech recognition method comprising:
obtaining a speech recognition model by:
detecting proper nouns from training data,
replacing, using dictionary data corresponding to a category indicated by an identifier pair among sets of dictionary data corresponding to different categories, each proper noun with a phonetic symbol sequence associated with each proper noun to obtain preprocessed training data, and
training the speech recognition model based on the preprocessed training data;
receiving speech data;
obtaining, from the received speech data, a candidate text including at least one word and a phonetic symbol sequence associated with a pronunciation of a target word included in the received speech data, using the trained speech recognition model;
replacing the phonetic symbol sequence included in the candidate text with a replacement word corresponding to the phonetic symbol sequence; and
determining a target text corresponding to the received speech data based on a result of the replacing,
wherein the at least one word includes at least one sub-word, and
the candidate text includes the at least one sub-word, the phonetic symbol sequence, and an identifier pair indicating a start and an end of the phonetic symbol sequence,
wherein the identifier pair indicates a category of the target word associated with the phonetic symbol sequence, and further comprising
replacing the phonetic symbol sequence included in the candidate text with the determined replacement word.
|