Behavioral Biometrics

Biometrics is always one of very fascinating domains to watch. I am fortunate to be part of some very exciting landmark programs/studies in UK & India. In my previous post, I shared some thoughts & tips on 'Biometric Spoofing'. Physical Biometrics has dominated the landscape so far, but - we can't ignore the growth & importance of Behavioral Biometrics in times to come. So, this post of mine keeps focus on Behavioral Biometrics coupled with latest trends.

Latest trends

• Research is in progress on new Biometric modals - Hand Geometry, Vein, Retina, Skin, Gait, Signature, Foot, Ear, Odor etc., Currently, though their market share is not significant, still at nascent stage, but - expected to grow in times to come
• Good research is in progress globally on Behavioral Biometrics - Biometrics based on behavioral traits help identify a person by some measurable activity
• Behavioral Biometric Solutions are cheaper than using Physical(also called as Physiological). Right fit for Verification only (1:1)
• Physical Biometrics is facing two major challenges: cost and user acceptance. Better fit for Identification (1:N) than Behavioral
• Behavioral Biometrics coupled with Physical Biometrics would be ideal for economic Multi-modal Solutions

Cost-Accuracy Curve from Frost- Sullivan (Below)

• Voice Verification is more accurate than Signature & Key Stroke Dynamics

• Signature Verification score more than Key Stroke Dynamics in accuracy

• No need for special sensors for Behavioral Biometrics unlike Physical Biometrics

• Behavioral Biometrics are good fit for continuous Biometric Applications, non intrusive too

• Noticed resistance from many users to accept solutions that capture and store things like Fingerprints or Iris patterns, perceive as invasion of their privacy. Behavioral Biometrics eliminates this challenge

• Currently, Behavioral scores over Physical on scalability, cost, ease of use, privacy where as lags behind in accuracy

• Multimodal Biometrics with Physical & Behavioral are going to surge in near future with features from 'best of both the worlds'
• Two of the four behavioral biometrics solutions listed earlier promise to effectively deal with these issues: mouse and keystroke dynamics.
• Signature and Voice dynamics require additional equipment or solution support from Physical Biometrics
• Currently, Signature & Voice dynamics are having relatively high error rates

• Mouse dynamics is in nascent stages of development, need more time to be considered as mature Biometric Solution

• Signature dynamics -- Algorithms are made up of some combination of appearance, shape, timing, and pressure as the user writes his signature
• Voice verification -- Analytics includes tone, pitch, and cadence to create a template to verify user identity
• Mouse dynamics -- Solution includes measurement of distance, speed, and angle contribute to identity analysis

• Keystroke dynamics -- the duration of each key-press and the time between keystrokes is used to help with identification
• Keystroke dynamics is anticipated to become a viable option soon for better Security Solutions

• Keystroke dynamics has advantages like remote authentication and the non-intrusiveness of the technology

• Gait - Walking Style, not very accurate, may change over time due to injury, physical body changes or temporary behavior changes and is not much differentiating among individuals

• Gait is good for verification in low security conditions; economical and can be remotely operated

• Gait Verification depends on factors such as speed, cadence, step length and width, or the angles in different articulations

Image Courtesy - Biometrika, Techrepublic, Hertasecurity

Happy to share more inputs going forward. Any views or thoughts? You are most welcome.

My posts are personal, do not reflect official policy or position of any Organization or Entity that I may be currently employed or associated in past