Features from accelerated segment test: Difference between revisions

Revision as of 17:46, 12 June 2013

Revision as of 00:06, 13 June 2013

Line 25:

* b, I_p→x≥ I_p + t (brighter)

Then choosing an x could(same for all p) partitions P (the set of all pixels of all training images) into 3 different subsetsubsets, P_d, P_s, P_b where:

* P_d = {p ∈ P : S_p→x = d }

* P_s = {p ∈ P : S_p→x = s }