1902 Unlock with Your Heart:Heartbeat-based Authentication on Commercial mobile Phones Lei Wang,Kang Wang,Ke Sun,Wei Wang,Chen Tian,Lei Xie,Qing Gu Presenter:Dr.Lei Xie State Key Laboratory for Novel Software Technology Nanjing University UBICOMP2018 Oct 10th,2018
Unlock with Your Heart: Heartbeat-based Authentication on Commercial Mobile Phones Lei Wang, Kang Wang, Ke Sun, Wei Wang, Chen Tian, Lei Xie, Qing Gu Presenter: Dr. Lei Xie UBICOMP2018 Oct 10th , 2018 State Key Laboratory for Novel Software Technology Nanjing University
Motivation Unlock your mobile device:authentication system to identify the user
Motivation Unlock your mobile device: authentication system to identify the user
Motivation Traditional metrics:token&password Easy to leakage and lost! ● 25932四 L
Motivation Traditional metrics: token&password Easy to leakage and lost!
Motivation Biometric features:fingerprint,face,and voices. Vulnerable to spoofing and replaying! A USER LOGI (⑧VoiceKey2.0
Motivation Biometric features: fingerprint, face, and voices. Vulnerable to spoofing and replaying!
Motivation Another biometric feature:vibration of chest in response to heartbeat(i.e.,SCG) Unique for each person ·Against spoofing ·Linked to"liveness
Motivation Another biometric feature: vibration of chest in response to heartbeat(i.e.,SCG) • Unique for each person • Against spoofing • Linked to “liveness
Can we enable the heartbeat-based authentication system on the commercial mobile devices?
Can we enable the heartbeat-based authentication system on the commercial mobile devices?
Basic idea We utilize the built-in accelerometer to capture the heartbeat vibration and perform user authentication. ax
Basic idea We utilize the built-in accelerometer to capture the heartbeat vibration and perform user authentication
Challenges Unlock With Your Heart 1.Human heartbeat patterns contain intrinsic Heart Rate variability (HRV) Hear Rate Estimaticn:66.616470 2.Reliable features with different durations should be extracted 3.Human heartbeat patterns change under various conditions
Challenges 1. Human heartbeat patterns contain intrinsic Heart Rate Variability (HRV) 2. Reliable features with different durationsshould be extracted 3. Human heartbeat patterns change under various conditions
Challenge 1 Unlock With Your Heart SCG Signal Fine Segmentation How to segment heart duration with r the effect of the variability of heart Hear Rate Estimation:66.616470 rate?
Challenge 1 How to segment heart duration with the effect of the variability of heart rate?
Basic idea ATC MC Heartbeat ● Observation:Although heart rate has Cycle intrinsic variability,the interval between AO and RF is stable. AC ● Solution:Segment the heartbeat Heartbeat cycle 0.05 -Volunteer A based on the stable interval between 30.04 Volunteer B AO and RF. 0o9 Volunteer C Volunteer D Volunteer E i0.01 0 0 0 50 Deviation(ms) Deviation of AO and RF from mean value
Basic idea ⚫ Observation: Although heart rate has intrinsic variability, the interval between AO and RF is stable. ⚫ Solution: Segment the heartbeat based on the stable interval between AO and RF. Deviation of AO and RF from mean value Heartbeat cycle