A Bayesian approach is very useful when prior knowledge of a process is available which needs to be combined with the sensed data to make statistical inferences about the parameters of the process.
Statistical approaches to image analysis using the Bayesian
paradigm have been popular in recent years
because of their capability to integrate low-level image
analysis and high-level tasks, as well as fuse data from different sensors.
Deformable template matching can be formulated
using the Bayesian framework [27,39,36].
In fact, in most cases, we can encode the structure and constraints about the deformable
template in terms of the prior and select an appropriate
likelihood function based on the sensor process
in a Bayesian scenario so that the corresponding Maximum A Posteriori (MAP)
estimate is equivalent to the solution of
the original deformable matching problem.