This paper presents a robust and structured pipeline for finger vein extraction and identification using a combination of classical image processing and deep learning. The system begins with preprocessing grayscale images to reduce noise and enhance clarity, followed by morphological operations and Otsu thresholding for precise vein pattern extraction. A Fast Non-Local Means (FNL) filter is applied to preserve edge details, and the image is transformed into the YUV color space to enhance luminance without distorting chromatic information. Histogram equalization and color inversion further improve contrast, while morphological operations refine the vein structures. The proposed VEFVI algorithm effectively extracts veins, and a CNN-based model (CCN-50) is trained for multi-class subject identification using the SDUMLA dataset. Experimental results demonstrate high accuracy, robustness under varying conditions, and suitability for real-time biometric applications. The proposed method achieves a Rank-1 identification rate of 96.2% and an Equal Error Rate (EER) of 2.84%, outperforming traditional baselines.
L. Wang, G. Leedham, and S. Y. Cho, “Minutiae feature analysis for infrared hand vein pattern biometrics,” Pattern Recognition, vol. 41, no. 3, pp. 920-929, 2008.
A. Kumar and Y. Zhou, “Human identification using finger images,” IEEE Transactions on Image Processing, vol. 21, no. 4, pp. 2228-2244, 2012.
Y. Li, M. Zhou, and L. Wang, “Adaptive repeated line tracking with threshold optimization for finger vein recognition,” IEEE Access, vol. 7, pp. 114092-114101, 2019, doi: 10.1109/ACCESS.2019.2935473.
J. Chen and Z. Xu, “Multi-scale maximum curvature method for finger vein image enhancement under non-uniform illumination,” Sensors, vol. 20, no. 3, p. 726, 2020, doi: 10.3390/s20030726.
S. Kim, H. Lee, and Y. Park, “Robust finger vein recognition using hybrid deep LBP features with CNN fusion,” Pattern Recognition Letters, vol. 145, pp. 47-54, 2021, doi: 10.1016/j.patrec.2021.02.004.
H. D. Nguyen and H. T. Nguyen, “Finger vein recognition using deep learning,” International Journal of Intelligent Engineering and Systems, vol. 14, no. 1, pp. 220-230, 2021.
M. T. Islam, M. M. Rahman, and M. A. Amin, “Efficient finger vein recognition using attention-enhanced lightweight CNN,” Computers in Biology and Medicine, vol. 142, p. 105222, 2022.
Y. Zhang and H. Wang, “Improved U-Net architecture for robust finger vein segmentation in varying conditions,” Biomedical Signal Processing and Control, vol. 79, p. 104060, 2023.
M. R. Haque, S. Ali, and M. Kabir, “Finger vein biometric authentication using adaptive histogram equalization and spatial attention,” Expert Systems with Applications, vol. 215, p. 119370, 2023.
Y. Yin, L. Liu, and X. Sun, “SDUMLA-HMT: A multimodal biometric database,” Lecture Notes in Computer Science, vol. 7098, pp. 260-268, 2011.
H. Kim, J. Park, J. Shim, and Y. Lee, “Application and optimization of a fast non-local means noise reduction algorithm in pediatric abdominal virtual monoenergetic images,” Electronics, vol. 13, no. 23, p. 4684, 2024.
N. K. Baloch, Z. Bhutto, A. S. Chan, M. L. Memon, K. Saleem, M. H. Shaikh, and S. Ahmed, “Finger-vein image dual contrast enhancement and edge detection,” International Journal of Computer Science and Network Security, vol. 19, no. 11, pp. 184-192, 2019.
Y. Ding, K. Wang, X. Wu, X. Xiang, X. Tang, and Y. Zhang, “Study for lightweight finger vein recognition based on a small sample,” Scientific Reports, vol. 14, no. 1, p. 12002, 2024.
N. K. B. Noroz, S. A. Ahmed, R. K. Kumar, D. S. B. Bhatti, and Y. R. Rehman, “Finger-vein image dual contrast adjustment and recognition using 2D-CNN,” Sukkur IBA Journal of Computing and Mathematical Sciences, vol. 6, no. 1, pp. 16-25, 2022.
B. Z. Kamil, D. A. Salman, and S. A. Kareem, “Detect interest region of finger vein based on K-means clustering,” Unpublished manuscript.
·
This work is licensed under a
Creative Commons Attribution-ShareAlike 4.0 International License
