The concept of this scheme is that the special features of
the scene observed at a specific position are exploited for
describing and subsequently identifying that position. In Fig. 3,
the position of a user can be inferred based on the similarity of
his observed scene characteristics.