Both mono-phone and triphone-based models are trained, and the training of the triphone
model uses decision tree clustering for state binding to reduce the number of states.
The fundamental idea is to explicitly model the context information of eye motions using context-dependent eye motion HMMs; this approach is analogous to the triphone
model used in high-performance speech recognition systems .
As a starting point, they took a large (1.6 MB) text corpus and after automatic phonetic transcription they created phone, diphone and triphone
statistics from it.
'It was a unanimous decision,' said Sri Lankan Cricket board secretary Triphone
The acoustic models were gender-dependent sharedstate triphone
hidden Markov models (HMMs) and were designed using tree-based clustering.
The most common approach is to use triphone
HMM states as (H)LDA classes, which was confirmed to outperform other approaches in .
Near 75000 triphone
models represented by about 40000 physical different HMMs were obtained as a result of training.
Its basic version uses triphones
as a basic speech element to model acoustic events occurring in speech utterance.
So diphones and triphones
are selected carefully that acquire complete information.
Try this group of triphones
with one tetraphone thrown in:
Other types of speaker-dependent systems include those that match phonemes, multiple words or triphones
, and are used for larger vocabularies.