Re: Digicams With MF Film Quality

acl wrote:

The only thing I wanted to express is that we can model the
situation as signal+noise, so I don't see how any mapping can be linear
in one but nonlinear in the other.

It seems that I am really misunderstanding the whole discussion here.
Assume we know how a Dirac delta function gets imaged (ie we know the
spatial frequency response); we also know everything is convoluted with
this function in the final image. So we invert this process, and this is
also linear. So where is the problem? Unless writing signal+noise is not
a good idea.

Most sophisticated deconvolution methods are non-linear.
The Richardson-Lucy method Roger Clark mentioned is,
for example. There are a number of reasons for this.
It's generally not possible to simply invert the convolution
with the PSF, because deconvolution is an ill-conditioned
problem. So one needs to regularize the process somehow.
One way of doing this is to impose a source model, for example
to model the true image as a sum of delta functions (point sources).
The blurred image is then a sum of PSFs. One constructs a
model by summing PSFs (there are a variety of ways to do this).
Implicit in this, however, is that all the sources are real, and thus
positive. The positivity constraint makes the method nonlinear.

Here's a short page to start with, with a couple of references:

<http://mathworld.wolfram.com/Deconvolution.html>

Read the page on CLEAN if you want to get an idea of how
this gets very complicated very quickly.

.

Relevant Pages

  • Re: Jeff Hawkins Q&A
    ... >> features and aspects of complexity and non linearity but don't really ... Computer power isn't the issue for nonlinear complexity. ... But linear complexity of itself has nothing to ... to the mechanization of intelligence for that reason. ...
    (comp.ai.philosophy)
  • Re: Any ideas how to linearize (and evaluate) this nonlinear equation?
    ... >nutrient, X0 is the total amount of the nutrient contained in all ... >point for the nonlinear one. ... if I had done a linear ... These four kinds of confidence intervals which you ...
    (sci.math.num-analysis)
  • Re: Is the Universe really that complicated?
    ... analysis vs abstraction problem (aka the egg/chicken ... Let's reformulate it into "Is the universe a linear ... nonlinear system and that no problem is really linear ... One odd case would be the matrix algebra where matrix ...
    (sci.math)
  • Re: fitting points to a sin function
    ... this function is linear in A and B and nonlinear in w. ... Even though it is visible that a sin function could fit, ... I have tried also with the fit function of curve fitting ...
    (comp.soft-sys.matlab)
Loading