Last decade has witnessed the booming development of smartphones. With the powerful computing and sensing capabilities and a large storage of a modern smartphone, instead of just making phone calls, a rich set of complex applications, such as taking photos, investing in stocks, sending emails and banking, are made possible to run on such devices. The security problem of private information (e.g., personal photos, contact list, emails and bank accounts) stored on smartphones therefore is of great importance to smartphone users. Either physiological characteristics (e.g., fingerprints and face recognition) or behavioral characteristics (e.g., voices, typing and stokes on touch screens) can be utilized to label or describe individuals. In this research initiative, we investigate new authentication schemes which are more difficult to compromise through shoulder surfing or biometrics hacking attacks and much easier to use.