[Bug?] Lack of the threshold judgement for the number of inliers during the monocular RANSAC?

[RonghaiHe]

I have noticed that the monocular RANSAC (geometricVerificationNister) in the loop_closure_detector.cpp just judges if the percentage of inliners > threshold (happens in this line),
while the stereo RANSAC (recoverPose) judges both the number of inliers and the percentage of them (happens in these lines).

Compared with the intra-loop closure detection part in the Kimera-VIO, the mono RANSAC also judges if the number of inliers satisfies the threshold (happens in these lines).

So is it a bug that the threshold judgement of the number of inliers is lack during monocular RANSAC? Thank you for your attention!