How to identify Gaussian type maxima in an image

I need to identify the correct maximum in a 2D image, which may not may not be a global maxima and is a kind of gaussian or asymmetric type.

Some example are attached.

A method other than imregionalmax or findpeaks is needed

Any suggestion

Thanks

[DGM]

What does this image mean? You want the "correct" maximum, but not necessarily the global maximum or local maximum? ... or are you just trying to avoid existing tools because of homework?

[SRT]

The right maxima. it can be global or local. The image shows the examples of right maximas and wrong maximas.

findpeaks in 2D is slow

imregionalmax has maximum connectivity of 8 pixels in 2D, the 2D maxima can have spread over 10 - 50 pixels.

using imregionalmax i can generate a list of candidate. I wanna check each one now. I may like to filter the one which are non symmetrical maximas.

[DGM]

Oh so it's more about finding the maximum that also locally fits a gaussian profile.

Answer 1

[Matt J]

using imregionalmax i can generate a list of candidate. I wanna check each one now. I may like to filter the one which are non symmetrical maximas.

One possibility would be to pull out the surrounding pixels of these candidates - in a neighborhood you think broad enough - and do a Gaussian fit, e.g., with this FEX download,

https://www.mathworks.com/matlabcentral/fileexchange/69116-gaussfitn

You can then assess the quality of the fit from the optimal fval to see if it is an admissible maximum.