Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > sci.image.processing > #4188 > unrolled thread

interpolation techniques for image reconstruction.

Back to article view | Back to sci.image.processing

Contents

  interpolation techniques for image reconstruction. james ikediashi <jamesikediashi@gmail.com> - 2016-12-03 05:53 -0800
    Re: interpolation techniques for image reconstruction. Martin Leese <please@see.Web.for.e-mail.INVALID> - 2016-12-03 10:44 -0700

#4188 — interpolation techniques for image reconstruction.

hi guys,  i am working on image reconstruction project using interpolation/extrapolation. specifically, i want to be able to add two images, say a father and mother, and use some interpolation/extrapolation technique to get what the offspring will look like. now i have my source images for father and mother. and i have my feature extraction, because my extrapolation will only work on key features such as the nose, the eyes, etc. now my question is what is the best interpolation technique to use? i should mention that i have considered using a bayesian algorithm with arbirtrary weights to signify the dominance of the genes of each parent, and effectively the dominance of their features in the offspring. i really am having difficulties modelling the right algorithm, and my supervisor suggested using a 2-d lagrange interpolation. but i do not think that will give me the result i am looking for. for example, we have both vectors for the eyes for father and mother. the lagrange interpolation will just go pixel by pixel and give me some sort of shape that might not be an eye eventually, this is just my thought, how do i go about this? thanks in anticipation.

[toc] | [next] | [standalone]

#4189

james ikediashi wrote:
> hi guys,  i am working on image reconstruction 
 > project using interpolation/extrapolation.
 > specifically, i want to be able to add two
 > images, say a father and mother, and use some
 > interpolation/extrapolation technique to get
 > what the offspring will look like.
...
 >  now my question is what is the best
 > interpolation technique to use? i should
 > mention that i have considered using a bayesian
 > algorithm with arbirtrary weights to signify
 > the dominance of the genes of each parent, and
 > effectively the dominance of their features in
 > the offspring.

Genetics is not my area, but my
understanding is that genes do not blend.
Darwin thought that they did, but Mendel
showed that they do not.  If a child has one
gene for blue eyes and another for brown
eyes, their eyes are brown.  The child's
eyes are not some blend of blueish-brown.

My guess is you don't need a blending
algorithm, but just a random number
generator generating true/false
(mother/father) for each of your features.

-- 
Regards,
Martin Leese
E-mail: please@see.Web.for.e-mail.INVALID
Web: http://members.tripod.com/martin_leese/

[toc] | [prev] | [standalone]

Back to top | Article view | sci.image.processing

csiph-web