Go to the source code of this file.
|Element of successor word list for LM factoring computation. More...|
|LM cache for the 1st pass. More...|
|Transition arc holding cell. More...|
|HMM state on tree lexicon. More...|
|wchmm-specific work area More...|
|Whole lexicon tree structure holding all information. More...|
|Number of arcs in an arc cell. |
This file defines structure for word-conjunction HMM, aka tree lexicon for recognition of 1st pass. Words in the dictionary are gathered to build a tree lexicon. The lexicon is built per HMM state basis, with their HMM output probabilities, transition arcs, and other informations for search such as successor word lists and maximum uni-gram scores for LM factoring, word head/tail marker, phoneme start marker, and so on.
Definition in file wchmm.h.