[ITK Community] [Insight-users] Cost function value increases

Brian Helba brian.helba at kitware.com
Fri Dec 6 12:44:47 EST 2013


Hi again Hana,

Sorry for the delay, I'm just catching up on missed emails from the U.S.
holiday. I'm not quite sure what your issue with header files is, can you
please clarify?

Thanks,
Brian


On Sat, Nov 30, 2013 at 7:55 AM, Hana Cherif <hanacherif62 at gmail.com> wrote:

> I get that my program does not take into account the header files from my
> optimizer , as when I tried to modify the header file optimizer so that
> it displays me an error !! when I did the compilation it works without
> any error !
>
>  Would you give me some hint of that please, thanks
>
>
> 2013/11/28 Brian Helba <brian.helba at kitware.com>
>
>> I would try some much larger maximum step values. You program appears to
>> be spending iterations while not making very much progress at the start.
>>
>> Also, when you're tuning things, it can be helpful to print the moving
>> image parameters on each iteration, in addition to the cost function.
>> On Nov 27, 2013 5:42 PM, "Hana Cherif" <hanacherif62 at gmail.com> wrote:
>>
>>>
>>> Hello every body !
>>>
>>>
>>> After execution of my images' registration program , using this
>>> components:
>>>
>>> Regular Step Gradient Descent Optimizer,
>>> Mean Squares Image  To Image Metric
>>> Bspline interpolator
>>>
>>>  I get this results::
>>>
>>>
>>> Iteration number:     Cost function value
>>>
>>>
>>> 0                        2732.74
>>> 1                        2700.84
>>> 2                        2699.27
>>> 3                        2701.48
>>> 4                        2703.19
>>> 5                        2700
>>> 6                        2704.12
>>> 7                        2705.1
>>> 8                        2705.27
>>> 9                        2705.4
>>> 10                        2702.91
>>> 11                        2705.15
>>> 12                        2713.1
>>> 13                        2720.48
>>> 14                        2720
>>> 15                        2715.81
>>> 16                        2704.64
>>> 17                        2687.46
>>> 18                        2666.14
>>> 19                        2639.91
>>> 20                        2613.11
>>> 21                        2587.55
>>> 22                        2564.89
>>> 23                        2546.22
>>>
>>> I use this parameters:
>>>  optimizer->SetMaximumStepLength(0.10 );
>>>  optimizer->SetMinimumStepLength( 0.05);
>>>  optimizer->SetRelaxationFactor(0.5);
>>>
>>> My question is : as you see i the results the value of the cost function
>>> increases i want to know why can be the reason of this results please!! is
>>> it just a probleme of parameters??
>>>
>>> Thanks !
>>>
>>>
>>>
>>> _____________________________________
>>> Powered by www.kitware.com
>>>
>>> Visit other Kitware open-source projects at
>>> http://www.kitware.com/opensource/opensource.html
>>>
>>> Kitware offers ITK Training Courses, for more information visit:
>>> http://www.kitware.com/products/protraining.php
>>>
>>> Please keep messages on-topic and check the ITK FAQ at:
>>> http://www.itk.org/Wiki/ITK_FAQ
>>>
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.itk.org/mailman/listinfo/insight-users
>>>
>>> _______________________________________________
>>> Community mailing list
>>> Community at itk.org
>>> http://public.kitware.com/cgi-bin/mailman/listinfo/community
>>>
>>>
>


-- 
Brian Helba
Medical Imaging
Kitware, Inc.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/community/attachments/20131206/1da9b97d/attachment.html>
-------------- next part --------------
_____________________________________
Powered by www.kitware.com

Visit other Kitware open-source projects at
http://www.kitware.com/opensource/opensource.html

Kitware offers ITK Training Courses, for more information visit:
http://www.kitware.com/products/protraining.php

Please keep messages on-topic and check the ITK FAQ at:
http://www.itk.org/Wiki/ITK_FAQ

Follow this link to subscribe/unsubscribe:
http://www.itk.org/mailman/listinfo/insight-users


More information about the Community mailing list