[Insight-users] xxxxxSPAMxxxxx Cost Function Definition

[RCL]

Hello,

I'm new using ITK and I'm trying to model the temporal intensity variations
in contrast enhanced MR images. The function of the model has been already
defined and I need to minimize a cost function to obtain the model
parameters that best fit my data. I want to use the
itk::LevenbergMarquardtOptimizer to minimize my cost function. I know that
the SetCostFunction requires a variable of the type of
MultipleValuedCostFunction, but from what I've read, these functions are
only a sum of Gaussians (please correct me if this is not the case). Is
there any way I can define a cost function with this form:

sin(pi/2 - atan(A/(2*a - 2*b)*(a*exp(-a*(t + 1/A/(a*exp(-a*t) -
b*exp(-b*t))*(I + A/(a - b)*(exp(-a*t) - exp(-b*t)))*(a - b))) - b*exp(-b*(t
+ 1/A/(a*exp(-a*t) - b*exp(-b*t))*(I + A/(a - b)*(exp(-a*t) - exp(-b*t)))*(a
- b)))) + A*(a*exp(-a*t) - b*exp(-b*t))/(2*a - 2*b)))*(I + (A*(1/exp(a*t) -
1/exp(b*t)))/(a - b))

in order to obtain the model parameters "A", "a" and "b" ? 

Thanks,

RCL
-- 
View this message in context: http://itk-insight-users.2283740.n2.nabble.com/Cost-Function-Definition-tp5325258p5325258.html
Sent from the ITK Insight Users mailing list archive at Nabble.com.