Opencv Repeatability Doesn't make sense?

Question

I am trying to evaluate SIFT and SURF detectors based on repeat criteria.

I find out below that the method can find Repeatability, Match SIFT and SURF

cv::evaluateFeatureDetector(img_1c, img_2c, h12, &key_points_1, &key_points_2, repeatability, corrCounter); 

Some of the results are listed below:

 Number Repeatibility Correspond Keypoint 1st Keypoint 2th 1to2 0.7777778 140 224 180 1to3 0.7125 114 224 161 1to4 0.704918 86 224 123 1to5 0.6853933 61 224 89 1to6 0.6521739 45 224 69 

to repeat the first line, you can calculate as → (root) / min (keypoint1st, keypoint2th) = (140/180) = 0.7777778 but for other lines this value is different from what I calculate using the above formula.

can anyone tell why this is happening?

Sincerely.

Answer 1

I think I might have found the problem here.

Repeat frequency is the number of repeating points between two images, taking into account the total number of selected key points. To measure the number of repeating points, it is necessary to take into account that the observed scenes differ from each other due to their altered state of the image (point of view, rotation + scale, blur, etc. In the case of Mikolajik's data set).

Key points that are not found in both images can repeatability, therefore, only points that lie in the overall scene of the part affect repeatability.

What happens when you look at all the detected key points in the second image to calculate repeatability and in fact you should use only the key points in the homography .

Hope this helps.

Answer 2

In fact, OpenCV 2.49 uses this line of code in evaluation.cpp :

 repeatability = minCount ? (float)correspondencesCount / minCount : -1; 

which is the number of matches divided by the smaller number of Keypoint 1st or Keypoint 2th , so I assume your error might be somewhere else. Can you post your full code where you install img_1c, etc., and call a function and display the results?