Finding best match by minMaxLoc returned values OpenCV

Question

I have one source picture and 5 other samples and I want to find the sample that matches the source image best. I concluded that should use template matching since histogram comparison didn't fit the problem well. I use cv::minMaxLoc to determine min and max values of the result matrix but I misunderstand how to compare this values to properly find the "best-matching" picture.
Thank you!

PS: I have read OpenCV documentation.

Adi Shavit · Answer 1 · 2014-05-05T05:47:32.797

2

Here's a tutorial explaining the use of cv::matchTemplate(). In addition to the min and max, cv::MinMaxLoc() returns the top-left corner of the template position for the best match.

Note that if you expect more than one max/min then cv::MinMaxLoc() is not sufficient since it returns only a single point.

edited May 05 '14 at 05:47

answered May 04 '14 at 14:56

Adi Shavit

16,743
5
67
137

what about min and max value – user3577807 May 04 '14 at 20:11
Please see the docs for `cv::MinMaxLoc()` here: http://docs.opencv.org/modules/core/doc/operations_on_arrays.html#minmaxloc – Adi Shavit May 05 '14 at 05:47
I think this is not an acceptable answer. The user ask distances between template and sources. She wants to select most similar source. – zakjma Nov 02 '15 at 08:25

Finding best match by minMaxLoc returned values OpenCV

1 Answers1