04-12-2012, 05:40 PM
Handwritten Signature Recognition
Handwritten Signature.pdf (Size: 1.04 MB / Downloads: 61)
Introduction
• Recognition: Verification/Identification.
• Behavioral biometric (e.g., speaker, gait, …).
• Challenges of signature biometric:
• Strongly affected by user-dependencies.
• Different features/thresholds for different users
• Highly intrinsic (i.e., not due to acquisition conditions)
intra-user variability.
• Statistical/elastic matching, template/decision
adaptation.
• Detecting “true” signature.
• Performance evaluation is carried out on random
impostor and different levels of skilled forgeries
(e.g., over the shoulder, professional, etc.).
Advantages of signature verification:
• User-friendly.
• Well accepted socially and legally.
• Non invasive.
• Already acquired in a number of applications.
• Acquisition hardware:
• Off-line: ubiquitous (pen and paper).
• On-line: inexpensive and already integrated in some
devices (Tablet PC).
• If compromised, can be changed.
• Long experience in forensic environments.
Disadvantages:
• High user intra-variability.
• Forgeries.
• Higher error rates than other traits.
• Affected by the physical and emotional state of the
user.
• Large temporal variation.
Some Factors Affecting the
Performance
• User-dependent decision thresholds.
• Simulated with a posteriori score alignment between
subjects (dashed lines in the following slides)
• Number of training signatures.
• Multi-session.
• Pressure signal.
• Acquisition hardware:
• MCYT (high quality pen tablet with paper)
• PRIP (Tablet PC)
• SVC (low quality pen tablet without visual feedback)