Page 138 - Classification Parameter Estimation & State Estimation An Engg Approach Using MATLAB

DISCRETE STATE VARIABLES 127

(a)

(b)

Figure 4.20 Detected license plate pixels using offline estimation. (a) Individually
estimated states. (b) Jointly estimated states

MATLAB functions for HMM
The MATLAB functions for the analysis of hidden Markov models are
found in the Statistics toolbox. There are five functions:

hmmgenerate: Given P t ( j ) and P z ( j ), generate a sequence of states
and observations.
hmmdecode: Given P t ( j ), P z ( j ) and a sequence of observations,
calculate the posterior probabilities of the states.
hmmestimate: Given a sequence of states and observations, estimate
P t ( j ) and P z ( j ).