一种结合图像质量评估和图像增强的指静脉识别方法

论文摘要

Biometric recognition is the study of physiological and behavioral characteristic of person to overcome the problem of Information security.In today modern world,variety of biometric technology are employed,like fingerprints,face,iris or behavioral characteristics such as gait and signature,for automatic identification of an individuals.Among all biometric traits,finger vein recognition（FVR）technology is considered one of the new,reliable and superior biometric modality,which attract the attention of researcher around the globe with its unique advantages.Compared to other biometric technology,FVR system have many advantages,such as anti-counterfeiting,user-friendly,live detection of finger vein,required a small device to capture finger vein image,easy to carry the FVR system.However,image quality of finger vein influence the performance of FVR system.Finger vein image is acquire using NIR（near infra-red light）in illumination transaction method.Therefore,most of the finger vein images prone to non-uniform illumination,low contrast,noise,low brightness problem.These low quality images have a great influence on the performance of FVR systems.Therefore,how to effectively evaluate the quality of finger vein image and how to enhance the low quality finger vein image is still a key issue for FVR system.In this paper,a detail review about various steps is presented i.e.image acquisition,preprocessing,feature extraction and matchingMoreover,we list of the some novel finding after critical finding comparative analysis of the highlighted techniques i.e.traditional finger vein recognition method,Machine Learning and Deep Learning Finger vein identification methods.During performance analysis of various,we find that Image quality is also one of the factor that effect the performance of Finger vein recognition system.Finally,we present a novel method to enhance the quality of finger vein images and improve the performance of the FVR system.To deal with the image quality problem,a novel finger-vein image quality assessment method and an enhancement method are proposed.The proposed finger vein recognition Scheme is based on two folds:（i）Image Quality Assessment,and（ii）Image Enhancement.First,the quality of the image is assessed by the decision tree with r-smote technique,to classify the finger vein image into two classes,i.e.High Quality（HQ）and Low Quality（LQ）images.Second,a single scale retinex filter（SSR）with chromaticity preserved algorithm and Gaussian filter are proposed to enhance the low and high quality finger vein images.To evaluate and test the strength of proposed system,two parameters-namely accuracy and equal error rate（EER）for classifier and recognition performance respectively are employed on a dataset of 1052 Finger vein images.It was determined from experiment that proposed image assessment and enhancement method outperformed other enhancement and assessment scheme,by achieving low error rate of 0.0379.The achieved results show the strength of the proposed art is better than already developed methods in FVR domain.Results also conclude that proposed art would be perfect pre-processing tool for finger vein feature based algorithms.

论文目录

• Abstract
• Introduction
•   1.1 Research Background
•   1.2 Finger Vein Recognition
•     1.2.1 Significance of Finger Vein Recognition
•     1.2.2 Research Status of Finger Vein Recognition
•     1.2.3 Materials
•   1.3 Contributions
•   1.4 Thesis Structure
• Chapter 2 Related Work
•   2.1 Steps of Finger Vein Recognition
•   2.2 Image Acquistion
•   2.3 Preprocessing
•     2.3.1 Image Assessment
•     2.3.2 ROI（Region of Interest） Selection
•     2.3.3 Image Enhancement & Normalization
•   2.4 Feature Extraction
•     2.4.1 Vein Based Method
•     2.4.2 Local Binary Based Method
•     2.4.3 Dimensionality Reduction Based Method
•     2.4.4 Minutiae Point-Based Method
•   2.5 Matching
•   2.6 Performance Analysis
•     2.6.1 Conventional Finger Vein Recognition Method
•     2.6.2 Traditional Machine Learning Finger Vein Recognition Methods
•     2.6.3 Finger Vein Recognition Method Using Deep Learning Methods
•   2.7 Problem Statement
•     2.7.1 Image Quality Assessment
•     2.7.2 Image Enhancement
• Chapter 3 Proposed Novel Image Quality Assessment and Enhancement Techniques for Finger Vein Recognition
•   3.1 Introduction
•   3.2 Our Method
•   3.3 Feature Selection
•     3.3.1 Gradient
•     3.3.2 Entropy
•     3.3.3 Information Capacity
•   3.4 Image Quality Evaluation using Decesion Tree with R-SMOTE technique
•   3.5 Finger Vein Image Enahncement Method
•     3.5.1 Single Scale Retinex filter with Choromaticity Preserved Algorithm
•     3.5.2 Gaussian filter
• Chapter 4 Experimental Results
•   4.1 Database and tools
•   4.2 Experiments Setting
•     4.2.1 Experiment 1
•     4.2.2 Experiment 2
•     4.2.3 Experiment 3
•   4.3 Conclusion
• Chapter 5 Conclusions and Discussions
• References
• Acknowledgements
• List of Published Papers
• 学位论文评阅及答辩情况表

文章来源

类型: 硕士论文

作者: KASHIF SHAHEED

导师: 尹义龙,杨公平

来源: 山东大学

年度: 2019

分类: 基础科学,医药卫生科技,信息科技

专业: 生物学,生物医学工程,计算机软件及计算机应用

单位: 山东大学

分类号: TP391.41;R318

总页数: 74

文件大小: 4936K

下载量: 111