LUIFT: LUminance Invariant Feature Transform

Figure 10

Office dataset results. For the corridor scene set: (a) Feature detector repeatability, (b) correct feature points detected, and (c) feature descriptor recall vs 1-precision curve. For the desktop scene set: (d) feature detector repeatability; (e) correct feature points detected; and (f) feature descriptor recall vs 1-precision curve.

(a)

(b)

(c)

(d)

(e)

(f)