Limits...
Intensity Variation Normalization for Finger Vein Recognition Using Guided Filter Based Singe Scale Retinex.

Xie SJ, Lu Y, Yoon S, Yang J, Park DS - Sensors (Basel) (2015)

Bottom Line: However, the capacities and percentages of finger tissues (e.g., bone, muscle, ligament, water, fat, etc.) vary person by person.This usually causes poor quality of finger vein images, therefore degrading the performance of finger vein recognition systems (FVRSs).The experimental results on two public datasets demonstrate the effectiveness of the proposed method in enhancing the image quality and finger vein recognition accuracy.

View Article: PubMed Central - PubMed

Affiliation: Institute of Remote Sensing and Earth Science, College of Science, Hangzhou Normal University, Hangzhou 311121, China. shanj_x@hotmail.com.

ABSTRACT
Finger vein recognition has been considered one of the most promising biometrics for personal authentication. However, the capacities and percentages of finger tissues (e.g., bone, muscle, ligament, water, fat, etc.) vary person by person. This usually causes poor quality of finger vein images, therefore degrading the performance of finger vein recognition systems (FVRSs). In this paper, the intrinsic factors of finger tissue causing poor quality of finger vein images are analyzed, and an intensity variation (IV) normalization method using guided filter based single scale retinex (GFSSR) is proposed for finger vein image enhancement. The experimental results on two public datasets demonstrate the effectiveness of the proposed method in enhancing the image quality and finger vein recognition accuracy.

Show MeSH
Image samples from our available database MMCBNU_6000 [21]: (a–d) are finger vein image samples collected from four volunteers P1–P4, respectively. Each row shows six images from six different fingers of one individual.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4541924&req=5

sensors-15-17089-f002: Image samples from our available database MMCBNU_6000 [21]: (a–d) are finger vein image samples collected from four volunteers P1–P4, respectively. Each row shows six images from six different fingers of one individual.

Mentions: Figure 2 shows four groups of finger vein images coming from MMCBNU_6000. Each row depicts six images taken from six fingers of a volunteer. The figure clearly illustrates that the images from different individuals displayed different finger structures and thicknesses. In addition, there were various image contrasts for each individual due to intensity variation. The first individual (P1) had the thickest fingers. Images (Figure 2a) collected from P1 showed good image contrast between the venous and non-venous areas. The individuals of P2 and P3 have thinner fingers than P1; therefore, the volume of each tissue was larger in P1 than those in P2 and P3. Thus, Figure 2b,c displayed higher brightness on the whole, as compared with the image brightness shown in Figure 2a. Furthermore, the images captured from P2 and P3 displayed different global and local image contrast due to the effect of intensity variation. Since the bone in the finger joint is articular cartilage and can easily be penetrated by infrared light [15], the joint part in the image is always shown in brighter gray values. This resulted in brighter local areas in each of the captured images shown in Figure 2. However, due to intensity variation, the local image contrast in the finger joint parts was much more obvious in Figure 2b than those in other images. Images displayed in Figure 2c showed lower global image contrast than those in the other images. Figure 2d shows good global image contrast; the regions between the venous and non-venous areas were not clear. This may have resulted from the presence of thick fat or muscle near the finger veins. Although the thickness of fingers from individual P2 and P4 are almost same, the image contrasts are different due to the intensity variation.


Intensity Variation Normalization for Finger Vein Recognition Using Guided Filter Based Singe Scale Retinex.

Xie SJ, Lu Y, Yoon S, Yang J, Park DS - Sensors (Basel) (2015)

Image samples from our available database MMCBNU_6000 [21]: (a–d) are finger vein image samples collected from four volunteers P1–P4, respectively. Each row shows six images from six different fingers of one individual.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4541924&req=5

sensors-15-17089-f002: Image samples from our available database MMCBNU_6000 [21]: (a–d) are finger vein image samples collected from four volunteers P1–P4, respectively. Each row shows six images from six different fingers of one individual.
Mentions: Figure 2 shows four groups of finger vein images coming from MMCBNU_6000. Each row depicts six images taken from six fingers of a volunteer. The figure clearly illustrates that the images from different individuals displayed different finger structures and thicknesses. In addition, there were various image contrasts for each individual due to intensity variation. The first individual (P1) had the thickest fingers. Images (Figure 2a) collected from P1 showed good image contrast between the venous and non-venous areas. The individuals of P2 and P3 have thinner fingers than P1; therefore, the volume of each tissue was larger in P1 than those in P2 and P3. Thus, Figure 2b,c displayed higher brightness on the whole, as compared with the image brightness shown in Figure 2a. Furthermore, the images captured from P2 and P3 displayed different global and local image contrast due to the effect of intensity variation. Since the bone in the finger joint is articular cartilage and can easily be penetrated by infrared light [15], the joint part in the image is always shown in brighter gray values. This resulted in brighter local areas in each of the captured images shown in Figure 2. However, due to intensity variation, the local image contrast in the finger joint parts was much more obvious in Figure 2b than those in other images. Images displayed in Figure 2c showed lower global image contrast than those in the other images. Figure 2d shows good global image contrast; the regions between the venous and non-venous areas were not clear. This may have resulted from the presence of thick fat or muscle near the finger veins. Although the thickness of fingers from individual P2 and P4 are almost same, the image contrasts are different due to the intensity variation.

Bottom Line: However, the capacities and percentages of finger tissues (e.g., bone, muscle, ligament, water, fat, etc.) vary person by person.This usually causes poor quality of finger vein images, therefore degrading the performance of finger vein recognition systems (FVRSs).The experimental results on two public datasets demonstrate the effectiveness of the proposed method in enhancing the image quality and finger vein recognition accuracy.

View Article: PubMed Central - PubMed

Affiliation: Institute of Remote Sensing and Earth Science, College of Science, Hangzhou Normal University, Hangzhou 311121, China. shanj_x@hotmail.com.

ABSTRACT
Finger vein recognition has been considered one of the most promising biometrics for personal authentication. However, the capacities and percentages of finger tissues (e.g., bone, muscle, ligament, water, fat, etc.) vary person by person. This usually causes poor quality of finger vein images, therefore degrading the performance of finger vein recognition systems (FVRSs). In this paper, the intrinsic factors of finger tissue causing poor quality of finger vein images are analyzed, and an intensity variation (IV) normalization method using guided filter based single scale retinex (GFSSR) is proposed for finger vein image enhancement. The experimental results on two public datasets demonstrate the effectiveness of the proposed method in enhancing the image quality and finger vein recognition accuracy.

Show MeSH