Finger vein recognition using local line binary pattern.

Area of Science:

• Biometrics
• Computer Vision
• Pattern Recognition

Background:

• Biometric personal verification is crucial for security.
• Existing methods like fingerprint and palm print have vulnerabilities.
• Finger vein patterns offer enhanced security as they are internal biometric traits.

Purpose of the Study:

• To propose a novel personal verification method using finger vein biometrics.
• To introduce and evaluate a new texture descriptor, Local Line Binary Pattern (LLBP), for feature extraction.
• To compare the performance of LLBP against existing methods like Local Binary Pattern (LBP) and Local Derivative Pattern (LDP).

Main Methods:

• Feature extraction using the proposed Local Line Binary Pattern (LLBP) descriptor.
• LLBP utilizes a straight-line neighborhood shape for texture analysis, differing from LBP's square shape.
• Implementation of a personal verification system based on extracted finger vein features.

Main Results:

• The proposed finger vein verification method using LLBP demonstrated superior performance.
• LLBP showed improved accuracy and robustness compared to LBP and LDP.
• Experimental results confirm the effectiveness of LLBP for finger vein recognition.

Conclusions:

• The Local Line Binary Pattern (LLBP) is an effective texture descriptor for finger vein biometrics.
• Finger vein verification using LLBP offers a more secure and accurate alternative to traditional methods.
• The proposed method advances the field of secure personal identification systems.