Biometric Identification Using Visual System Classification on Handheld

Biometric Identification Using Visual System Classification on Handheld Devices Robb Zucker

Overview • The human eye and visual system • Classification potential based on visual system abnormalities and pathology • Proposed experiment

Human Visual System

Problems with the visual system • The National Institute for Health (NIH) lists as many as 15 common causes for decreased visual acuity – Some physical abnormalities to the visual system are inherent from birth (congenital) – Others, more commonly, occur with age (presbyopia)

Light sensitivity • Light sensitivity decreases as we age, as early as age 20. The intensity of illumination for light to just be seen is doubled every 13 years thereafter • Due to: – resting diameter of pupil decreases (senile miosis) – Lens opacification – Vitreous humor opacification

The dress… • Blue with black trim? • White with gold trim? Many factors effect how we see or perceive colors and shapes

The Experiment • Create an app that forces users into difficult reading situations • Capture both the light sensitivity rating and the orientation for each user • Classification using K-Nearest Neighbor algorithm

Device sensors and controls • INPUTS – Sensor. TYPE_LIGHT: Ambient light level in SI lux units – Sensor. TYPE_ORIENTATION: values[0]: Azimuth values[1]: Pitch, rotation around x-axis values[2]: Roll, rotation around the x-axis • OUTPUTS – Settings. System. SCREEN_BRIGHTNESS

Effective Brightness Correct backlight brightness value for ambient light passively illuminating the device

As effective brightness is increased, in what orientation are users holding the device?

Expected results For a given brightness level person A person B person C Machine-Learning classification system based on k Nearest Neighbor (k-NN)

Application • Additional security safeguard in a broader security system • As a compliment to challenge question • As a compliment to user defined security icon

References • • • [1]Vision and Perception - Visual Processing http: //medicine. jrank. org/pages/1805/Vision-Perception-Visual-processing. html [2] University of Calgary http: //ucalgary. ca/pip 369/mod 9/aging/sensitivity [3] Unar, J. A. , Woo Chaw Seng, and Almas Abbasi. "A review of biometric technology along with trends and prospects. " Pattern recognition 47. 8 (2014): 2673 -2688. [4] News report via internet: http: //fox 13 now. com/2015/02/26/what-color-is-this-dress-viral-photo-stirsintense-internet-debate/ [5] Ross, Arun, and Anil Jain. Multimodal biometrics: An overview. na, 2004. [6] National Institute of Health website: http: //www. nlm. nih. gov/medlineplus/ency/article/003029. htm [7] Ullah, Abrar, et al. "Graphical and text based challenge questions for secure and usable authentication in online examinations. " Internet Technology and Secured Transactions (ICITST), 2014 9 th International Conference for. IEEE, 2014.