-1
$\begingroup$

I am testing the linear grey level method stated in enter link description here.Now it is clear that this method have a major problem: If there are already pixels of values 0 or 1, there won't be any changes in the image.

As a solution, I apply a threshold to my image to put for example pixels with values above threshold 1 to threshold1 and pixels with values below threshold 2 to thresold 2.

Now this seems to be an unatural way to deal with this problem.Are there more methods to bypass the maximum and minimum issue?

$\endgroup$
  • $\begingroup$ If you wish to downvote,please explain at least your reasons. $\endgroup$ – ChiPlusPlus Oct 13 '17 at 20:01
0
$\begingroup$

If we stick to point-wise processing, you can use a soft transfer function to increase the contrast, like:

Gimp Curves
Figure 1. Gimp curves too. Horizontal axis is the input value and vertical axis is the output value. The input histogram is shown in the background.

This way there will be still some contrast left in the dark and bright values that would get clipped if the linear portion of the curve was extended to encompass all input values. I don't have formulas for automating this.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.