Robust Finger-vein ROI Localization Based on the 3σ Criterion Dynamic Threshold Strategy

Area of Science:

• Biometrics
• Image Processing
• Computer Vision

Background:

• Accurate region of interest (ROI) localization is crucial for effective finger-vein identification.
• Challenges like uneven illumination, noise, and finger distortion hinder precise ROI definition.

Purpose of the Study:

• To propose a robust finger-vein ROI localization method.
• To improve matching accuracy in finger-vein identification systems.

Main Methods:

• Utilized the Kirsch edge detector for horizontal-like edge detection in finger-vein images.
• Applied a 3 σ criterion dynamic threshold strategy across image quadrants for enhanced edge information.
• Defined finger boundaries by labeling the longest connected components to localize the ROI.

Main Results:

• The proposed method demonstrated competitive ROI localization performance across multiple datasets.
• Satisfactory matching results were achieved on diverse finger-vein image datasets.
• The method proved effective even with challenging image acquisition conditions.

Conclusions:

• The 3 σ criterion dynamic threshold strategy offers a robust solution for finger-vein ROI localization.
• The developed method enhances the accuracy and reliability of finger-vein identification systems.
• This approach addresses key limitations in existing ROI localization techniques.