Add missing sqrt to magSpectrum(). by yo1990 · Pull Request #20102 · opencv/opencv

yo1990 · 2021-05-17T07:02:07Z

resolves #20100

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

I agree to contribute to the project under Apache 2 License.
To the best of my knowledge, the proposed patch is not based on a code under GPL or other license that is incompatible with OpenCV
The PR is proposed to proper branch
There is reference to original bug report and related work
There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
The feature is well documented and sample code can be built with the project CMake

Add some std::sqrt() to magSpectrum().
#20100

force_builders=ARMv7,Custom
build_image:Custom=centos:7
buildworker:Custom=linux-1,linux-4,linux-6

llschloesser · 2021-05-21T13:21:03Z

For phase correlation, the | | in DFT' (F G* / |F G*|) means absolute value and not magnitude, right? Is magSpectrums() the wrong operation to use for phase correlation?

llschloesser · 2021-05-21T13:32:34Z

For phase correlation, the | | in DFT' (F G* / |F G*|) means absolute value and not magnitude, right? Is magSpectrums() the wrong operation to use for phase correlation?

In my code I replaced:
magnitude_of_spectrums(cross_power_, cross_power_magnitude_);
with
cv::absdiff(cross_power_, cv::Scalar::all(0), cross_power_magnitude_);
and observed no loss of accuracy/quality. Same good results as before. What am I not understanding?

llschloesser · 2021-05-21T18:51:40Z

For phase correlation, the | | in DFT' (F G* / |F G*|) means absolute value and not magnitude, right? Is magSpectrums() the wrong operation to use for phase correlation?

In my code I replaced:
magnitude_of_spectrums(cross_power_, cross_power_magnitude_);
with
cv::absdiff(cross_power_, cv::Scalar::all(0), cross_power_magnitude_);
and observed no loss of accuracy/quality. Same good results as before. What am I not understanding?

In the book Image Processing, Analysis, and Machine Vision, the magnitude function | F(u,v) | is defined as sqrt(R(u,v)^2 + I(u,v)^2). So it seems that the magnitude function on a spectrum is indeed the right operation and is denoted with single bars rather than double?

modules/imgproc/src/phasecorr.cpp

vpisarev · 2021-05-29T03:17:09Z

👍

yo1990 · 2021-05-29T06:06:59Z

Thank you.
I agree with your corrections; absolute value is more appropriate, not magnitude.

asmorkalov assigned vpisarev May 21, 2021

vpisarev reviewed May 25, 2021

View reviewed changes