How to find local sub-pixel maxima in image?

I need to find the x- and y- coordinates of local maxima of objects in a microscope image (in this case, balls of fluorescent DNA that manifests as a dots of a few pixels in the image). Please see example image below. Note that these were not taken with any kind of super-resolution localization method. I can find the centroid and brightest pixel for each image, but I would like to find the sub-pixel location of maxima. I am aware of Gaussian fitting methods (maximum likelihood estimation, etc.) where one would upsample and find peaks, but I am not sure what is best/most appropriate algorithm. Any helpful hints would be appreciated.

Example image here: https://www.dropbox.com/s/oa6fa9esew0gtps/loc_max.tif?dl=0

Answer 1

[Image Analyst]

I'm thinking that it might be

GMModel = fitgmdist(XYZ, 3);

but I'm not sure. I'd need data to try it out with. It requires the Statistics and Machine Learning Toolbox. I guess I could try it with fspecial() which can generate a Gaussian, but I don't have time right now. Hopefully it's a start. If you run into trouble, then post your array. Hopefully you can at least get a single spot by itself and we don't have to find those in an image of hundreds of them - I hope you've done that part already.

Answer 2

[Walter Roberson]

Your images are normal lossy JPEG images. You cannot meaningfully find sub-pixel locations in them. You are not working with original data; you are working on the reconstruction of the JPEG model of the data.

[Jonathan Cheng]

Originally, I couldn't attach a .tif image. I've attached original .tif in the link. Thanks.