edu.cmu.sphinx.linguist.acoustic.tiedstate

Class SenoneHMM

java.lang.Object

edu.cmu.sphinx.linguist.acoustic.tiedstate.SenoneHMM

All Implemented Interfaces:

HMM

public class SenoneHMM
extends java.lang.Object
implements HMM

Represents a hidden-markov-model. An HMM consists of a unit (context dependent or independent), a transition matrix from state to state, and a sequence of senones associated with each state. This representation of an HMM is a specialized left-to-right markov model. No backward transitions are allowed.

Constructor Summary

Constructors

Constructor and Description
SenoneHMM(Unit unit, SenoneSequence senoneSequence, float[][] transitionMatrix, HMMPosition position) Constructs an HMM

Method Summary

Modifier and Type	Method and Description
boolean	equals(java.lang.Object o)
Unit	getBaseUnit() Gets the base unit associated with this HMM
HMMState	getInitialState() Gets the initial states (with probabilities) for this HMM
int	getOrder() Returns the order of the HMM
HMMPosition	getPosition() Retrieves the position of this HMM.
SenoneSequence	getSenoneSequence() Returns the SenoneSequence associated with this HMM
HMMState	getState(int which) Retrieves the hmm state
float[][]	getTransitionMatrix() Returns the transition matrix that determines the state transition probabilities for the matrix.
float	getTransitionProbability(int stateFrom, int stateTo) Returns the transition probability between two states.
Unit	getUnit() Gets the unit associated with this HMM
int	hashCode()
boolean	isComposite() Determines if this HMM is a composite HMM
boolean	isContextDependent() Determines if this HMM corresponds to a context dependent unit
boolean	isFiller() Determines if this HMM represents a filler unit.
java.lang.String	toString() Returns the string representation of this object

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Constructor Detail

SenoneHMM

public SenoneHMM(Unit unit,
                 SenoneSequence senoneSequence,
                 float[][] transitionMatrix,
                 HMMPosition position)

Constructs an HMM

Parameters:

unit - the unit for this HMM

senoneSequence - the sequence of senones for this HMM

transitionMatrix - the state transition matrix

position - the position associated with this HMM

Method Detail

getUnit

public Unit getUnit()

Gets the unit associated with this HMM

Specified by:

getUnit in interface HMM

Returns:

the unit associated with this HMM

getBaseUnit

public Unit getBaseUnit()

Gets the base unit associated with this HMM

Specified by:

getBaseUnit in interface HMM

Returns:

the unit associated with this HMM

getState

public HMMState getState(int which)

Retrieves the hmm state

Specified by:

getState in interface HMM

Parameters:

which - the state of interest

Returns:

hmm state

getOrder

public int getOrder()

Returns the order of the HMM

Specified by:

getOrder in interface HMM

Returns:

the order of the HMM

getSenoneSequence

public SenoneSequence getSenoneSequence()

Returns the SenoneSequence associated with this HMM

Returns:

the sequence of senones associated with this HMM. The length of the sequence is N, where N is the order of the HMM. Note that senone sequences may be shared among HMMs.

isComposite

public boolean isComposite()

Determines if this HMM is a composite HMM

Returns:

true if this is a composite hmm

getTransitionMatrix

public float[][] getTransitionMatrix()

Returns the transition matrix that determines the state transition probabilities for the matrix. Each entry in the transition matrix defines the probability of transitioning from one state to the next. For example, the probability of transitioning from state 1 to state 2 can be determined by accessing transition matrix element[1][2].

Returns:

the transition matrix (in log domain) of size NxN where N is the order of the HMM

getTransitionProbability

public float getTransitionProbability(int stateFrom,
                                      int stateTo)

Returns the transition probability between two states.

Parameters:

stateFrom - the index of the state this transition goes from

stateTo - the index of the state this transition goes to

Returns:

the transition probability (in log domain)

getPosition

public HMMPosition getPosition()

Retrieves the position of this HMM. Possible

Specified by:

getPosition in interface HMM

Returns:

the position for this HMM

isFiller

public boolean isFiller()

Determines if this HMM represents a filler unit. A filler unit is speech that is not meaningful such as a cough, 'um' , 'er', or silence.

Returns:

true if the HMM represents a filler unit

isContextDependent

public boolean isContextDependent()

Determines if this HMM corresponds to a context dependent unit

Returns:

true if the HMM is context dependent

getInitialState

public HMMState getInitialState()

Gets the initial states (with probabilities) for this HMM

Specified by:

getInitialState in interface HMM

Returns:

the set of arcs that transition to the initial states for this HMM

toString

public java.lang.String toString()

Returns the string representation of this object

Overrides:

toString in class java.lang.Object

Returns:

the string representation

hashCode

public int hashCode()

Overrides:

hashCode in class java.lang.Object

equals

public boolean equals(java.lang.Object o)

Overrides:

equals in class java.lang.Object