Whether actual trials met the conditional target; this is conditional on the assumed sampling/support model.
True when an admitted homography explains the forbidden support population.
Correspondence count meeting the declared forbidden planar overlap criterion.
Last admitted homography, or null when no candidate survived numerical and support admission.
Attempted hypotheses rejected for numerical admission or degeneracy.
Number of correspondences in the admitted hypothesis population.
Hypothesis-support refits refused by numerical admission.
Conditional trial requirement at the forbidden planar-support threshold, or null when unavailable.
Residual-gated original correspondence indices and their pixel quantiles for the current model.
Original correspondence ordinals passing the declared residual gate, in stable input order.
Median accepted residual in pixels, or null for an empty accepted population.
Ninetieth-percentile accepted residual in pixels, or null for an empty accepted population.
Actual random-sample hypotheses attempted under the caller's explicit work ceiling.
Search a planar explanation of the final fixed epipolar-inlier population. Four-point homographies compete with symmetric pixel transfer. Absence of a dominant model requires the conditional trial target for the forbidden overlap fraction, not the much weaker best homography's own support ratio. A found forbidden overlap is sufficient refusal evidence and ends the search.