An Innovative SIFT-Based Method for Rigid Video Object Recognition

Figure 3

The region selection modes of the target object.

(a)

(b)

(c)

(d)

(e)

(f)