Related Experiment Video
Updated: Nov 10, 2025
04:19
A New Best Practice for Validating Tail Vein Injections in Rat with Near-infrared-Labeled Agents
Published on: April 19, 2019
21.8K
A Novel Finger Vein Recognition Method Based on Aggregation of Radon-Like Features
Qiong Yao1, Dan Song1, Xiang Xu1
1Artificial Intelligence and Computer Vision Laboratory, University of Electronic Science and Technology of China, Zhongshan Institute, Zhongshan 528402, China.
Sensors (Basel, Switzerland)
|April 3, 2021
Summary
A new method enhances finger vein (FV) biometrics by combining curvature and radon-like features (RLF) to improve image quality. This approach extracts more robust and genuine vein patterns for accurate individual recognition.
Area of Science:
- Biometrics
- Image Processing
- Pattern Recognition
Background:
- Finger vein (FV) biometrics offers unique, anti-forgery, and bio-assay recognition capabilities.
- Low-quality FV images (low-contrast, blur, noise) hinder effective feature extraction.
- Robust feature extraction from degraded FV images remains a significant challenge.
Purpose of the Study:
- To develop a novel and robust feature extraction method for low-quality finger vein images.
- To enhance the accuracy and reliability of finger vein biometrics.
- To address limitations of existing curvature-based methods in preserving genuine vein patterns.
Main Methods:
- A novel feature extraction method combining curvature and radon-like features (RLF) was developed.
- Mean curvature calculation enhanced the initial vein pattern image.
- A specific RLF implementation aggregated spatial information, highlighting veins and suppressing noise.
Main Results:
- The RLF-based method produced smoother and more genuine vein structure images.
- It effectively preserved inherent vein patterns while eliminating pseudo-vein information.
- Experiments on multiple databases demonstrated improved vein pattern completeness and continuity.
Conclusions:
- The proposed RLF-based feature extraction method significantly enhances finger vein image quality.
- This method offers superior recognition accuracy compared to existing techniques.
- It provides a more robust solution for finger vein biometrics, especially with low-quality images.
