Could someone point me in the direction of how to solve this?

I = [I1, . . . , In] is a 1D grayscale image and D = [D1, . . . , Dn] represents the second derivative of I. I am given the four pixel intensities I1, I2, In−1, In] and the second derivative values D3, . . . , Dn−2. How would I compute the rest of I’s intensities?

What I have tried so far: Each Ix intensity can be approximated via Taylor series: Ix = I(0) + x(dI(o)/dx) + (1/2)x^2(d^2I(x)/dx^2).

I am sure the trick is in using the sliding window algorithm for fitting a 2nd degree polynomial, and that a matrix is involved in solving. However, I am unsure how big the sliding window should be (3, 5 pixels?), etc.

  • $\begingroup$ Hint: What's the definition of the second derivative? Note that you can use LaTeX here to typeset mathematics in a more readable way. See here for a short introduction. $\endgroup$
    – D.W.
    Feb 28, 2016 at 4:16

1 Answer 1


I am sure these are finite differences, so the first "derivatives" are just

$B_i := I_{i+1} - I_i$

and the second "derivatives" are the differences of those:

$D_i := B_{i+1} - B_i$

Write it out and solve for $I_i$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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