Palmprint Recognition Using Bifurcation Line Direction Coding

Palmprint recognition is a challenging task due to the variability in image quality, scale, and angle. Traditional methods often rely on single line features, which may not effectively capture the local bifurcation information critical for accurate recognition. To address these challenges, this paper proposes the Bifurcation Line Direction Coding (BLDC) method, which better captures the local bifurcation characteristics of palmprint images. Our approach leverages an improved Gabor filter for image preprocessing, where the main direction guides the branch direction. By considering both the maximum and minimum direction subscripts of the main direction, we generate robust feature codes that accurately represent the palmprint’s unique structure. Extensive experiments across five challenging palmprint datasets demonstrate that BLDC outperforms existing methods in recognition rate and feature extraction speed, providing a more effective solution for palmprint recognition in practical applications.