0
$\begingroup$

I know similar questions have already been asked, but as I understand, little differences in my goal make a difference on a decision on which algorithm to use. As I am a total novice in this area, I am asking which algorithm would be best for measuring the similarity of two very similar images - e.g. two photos of faces which were taken from the same angle and are grayscale. Basically, this algorithm would need to quantify the difference or similarity of these two very similar pictures.

What software can I use to do that? Basically, I need something that inputs 2 pictures and then produces a number which tells me how different/similar they are.

EDITED: Picture examples

Hey something as similar as these two pictures Car1 and Car2. And which software can be used for these analyses?

$\endgroup$
  • $\begingroup$ The algorithm choice will depend on your specific application. Will you mainly be identifying cars ? any kind of images ? What do you consider as similars ? Is background relevant ? You should try to narrow a little bit more your need. If you are looking for something that can be as generic as possible and can assess similarities like a human would do, I believe the way to go would be neural network. $\endgroup$ – Pier-Yves Lessard Apr 21 '17 at 10:35
1
$\begingroup$

This measure really depends on how your images look like. A very basic thing would be the calculate the point-wise difference between both images and summing them up. The smaller this value is, the more equal they are. Possibly, you can perform some Gaussian Smoothing and contrast equalization before. But, that's really depending on the images.

$\endgroup$
  • $\begingroup$ I edited my original post to give you an impression about the images I am using. And which software can be used for such analyses? $\endgroup$ – User33268 Jan 20 '17 at 13:51

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.