Personal identification techniques based on operational habit of cellular phone

Biometrics authentication which identifies a legitimate user of a cellular phone has been adopted. However, it is not used so actively, because control procedure is complicated, and a user has a resistance to provide own biological information to the communication terminal. In this paper, we propose personal identification techniques based on operational habit of cellular phone, which authenticate the legitimate user by collecting typing histories of cellular phone in a background process. The typing histories that are recorded in the ring buffer are divided into two profiles. One is "short term profile" which is histories from latest typing, and the other is "long term profile" which are older histories than the short term records. Also, we implement two authentication algorithms; first algorithm retrieves a remarkable key of the short term profile from the long term profile. Second algorithm computes weight values of keys of the short term profile by using long term profile. We evaluate the FAR (False Acceptance Rate) and FRR (False Reject Rate), and we show that our method can apply to personal identification on the cellular phone.