Method of identification of a movement for the sake of simplification, to a markov model mask

摘要

The present invention relates to a treatment process for calculating a function relating to a first event, the assembly of the events being constituted by a hidden markov model, the treatment process comprising a step of optimization by minimization of the optimization function in order to obtain a first matrix optimum calculation and a second matrix of optimum calculation, the optimization function as the variable entries the first calculation matrix and the second matrix of calculation, and the optimization function dependent on at least of the time sequence of data supplied to the second event, and a step of calculating the function relating to the first event from the first matrix optimum calculation and the second matrix of optimum calculation.