KR20100018279A - Method for capturing bio-metrics information using finger-vein recognition - Google Patents

Abstract

PURPOSE: A bio recognizing method using finger-vein recognition is provided to determine a blood vessel area using the average and standard deviation of brightness and assign a recognition weight value to the determined blood vessel area, thereby quickly and accurately performing bio recognition. CONSTITUTION: In a local area, a feature value for finger-vein recognition is extracted(S30). An area including a vein and a skin area are classified based on the average and standard deviation of brightness(S40). A feature value in which the relatively higher first weight value is extracted is assigned to the area including the vein(S60). The second weight value is given in case on a skin area(S70).

Description

Method for capturing bio-metrics information using finger-vein recognition}

The present invention relates to a bio-metrics method, and in particular, the process of extracting a blood vessel region is omitted, and the average and standard deviation of brightness are calculated for a local region of a predetermined size in the acquired finger vein image. The present invention relates to a method for faster and more accurate bio-recognition by determining a blood vessel region and assigning a recognition weight thereto.

In general, bio-recognition is a method of recognizing an individual using human physiological or behavioral characteristics. Typical technologies include fingerprints, faces, irises, finger veins, signatures, and speaker recognition. Among these methods, finger vein recognition has the advantage of high recognition performance, less discomfort and inconvenience of the user, and faster recognition time.

In the prior art, a method of detecting a location of a finger vein and extracting a feature based on the location of a finger vein can be obtained in the following data.

[1] T. Yanagawa, S. Aoki, and T. Ohyama, “Human Finger Vein Images Are Diverse And Its Patterns Are Useful for Personal Identification”, MHF 2007-12, pp. 1-7, 2007.

[2] N. Miura, A. Nagasaka, and T. Miyatake, "Feature Extraction of Finger-Vein Patterns Based on Repeated Line Tracking and Its Application to Personal Identification", Machine Vision and Applications, vol. 15, pp. 194-203, 2004.

[3] Z. Zhang, S. Ma, and X. Han, "Multiscale Feature Extraction of Finger-Vein Patterns Based on Curvelets and Local Interconnection Structure Neural Network", The 18th International Conference on Pattern Recognition (ICPR'06), pp. . 145-148, 2006.

[4] N. Miura, A. Nagasaka, and T. Miyatake, "Extraction of Finger-Vein Patterns Using Maximum Curvature Points in Image Profiles", IEICE TRANS. INF. & SYST., Vol. E90-D, no. 8, pp. 1185-1194, 2007.

In other words, the concept that the human finger can be identified by the finger vein has been proved by Yanagawa's research [1]. According to Yanagawa's research, each person's vein pattern of ten fingers is different, and one finger vein pattern is suitable for personal authentication with similar degrees of freedom as the highly reliable iris pattern [1].

In addition, Miura proposed a method to extract vein patterns with unbounded borders by vein line tracing starting from various locations on the finger image, and the experimental results showed an average recognition error of 0.145%. ms) [2].

Zhang proposed a finger vein pattern extraction method based on curvelet and local interconnection structure Neural Networks, and the average error rate was 0.128% based on Miura's matching method [3].

Meanwhile, a recent study by Miura suggested a method for efficiently extracting vein patterns of various thicknesses, based on the fact that the thickness of the pattern may vary due to changes in vein blood flow depending on weather or human condition. Recognition error of 0.0009% showed excellent performance [4].

The conventional camera-based finger vein recognition method does not detect the position of the finger vein, and studies have been conducted to extract the feature and use it for recognition. After extracting the position of the finger vein from the skin region of the acquired image, The method used for recognition is used.

However, these conventional methods require a long calculation time for extracting vein regions having inconspicuous boundaries, and there is a problem in that vein regions may not be accurately extracted by dark regions caused by thick skin or finger skeletons. In addition, the ratio of the area occupied by the vein pattern compared to the skin area is very low, and there are not many intersections or end points of the vein pattern, and thus high recognition rate cannot be guaranteed.

In order to solve these problems, studies on accurate and rapid vein area extraction have been continuously conducted, but have not solved the fundamental problem according to these problems.

Accordingly, the present invention is to solve the above problems, reflecting whether the vein is included in the unit of the local area without any additional process of detecting the finger vein pattern, and designation independent of the influence of uneven illumination The purpose is to provide a fast and accurate biometric recognition method using a Mac.

Another object of the present invention is based on a method that does not go through the extraction step of the vein, the designation that can improve the efficiency of the user and attempt to improve the efficiency of the finger vein recognition system by reducing the time required for recognition by reducing the processing time The present invention provides a biometric recognition method using MAC recognition.

Still another object of the present invention is to provide a biometric recognition method using finger vein recognition, which does not use the feature extraction method based on the exact location of the vein, and thus may not be greatly affected by the shadows on thick skin or skeleton. .

A feature of the biorecognition method using the finger vein recognition according to the present invention for achieving the above object is of a predetermined size based on the finger vein image acquired using a biometric information acquisition device having an infrared light and a camera. Extracting feature values for finger vein recognition in a local region, a region including a vein based on an average and a standard deviation of brightness values of each local region calculated from the finger vein images acquired in units of the local region; Dividing the skin into regions, and in the case of the region including the vein, a first weight is given to the feature value extracted from the region, and in the case of the skin region, the feature value extracted from the region. Giving a second weight lower than the first weight to and based on a feature value reflecting the weight. To each other, matching the pre-stored specification of the user acquires the vein image in which the biometric data is to include a step of performing individual authentication.

Another feature of the biorecognition method using the finger vein recognition according to the present invention for achieving the above object is a predetermined size based on the finger vein image acquired using a biometric information acquisition device having an infrared illumination and a camera Extracting feature values for finger vein recognition in a local region of the region, and an area including a vein based on the average and standard deviation of brightness values of each local region calculated from the finger vein images acquired in units of the local region; And dividing the skin into a skin region and an undetermined region, and in the case of the region including the vein, a first weight is given to the feature value extracted from the region, and in the case of the skin region, Gives a second weight relatively low to the feature value extracted from, and in the case of an undetermined region, is lower than the first weight, And assigning a third weight higher than the second weight, and performing personal authentication by matching previously stored bioinformation with the acquired finger vein image based on the feature value reflecting the weight. It features.

Another feature of the biorecognition method using the finger vein recognition according to the present invention for achieving the above object is predetermined based on the finger vein image acquired using a biometric information acquisition device having an infrared light and a camera Extracting feature values for finger vein recognition in a local area of size, and including veins based on the mean and standard deviation of brightness values of each local area calculated from the finger vein images acquired by the local area; And a feature value extracted for each local area from the acquired finger vein image of the user and a feature extracted for each local area from the finger vein image of the user previously stored in the storage for registration. Comparing the values with each other to calculate a similarity value, and in the case of the region including veins in the classification step, Giving a relatively high first weight to the similarity value calculated in the reverse region, and giving a second weight lower than the first weight to the similarity value calculated in the corresponding region in the case of the skin region, and reflecting the weight. And matching the previously stored bio information and the acquired finger vein image based on the similarity value to perform personal authentication.

Preferably, the feature value may be extracted by using a local binary pattern (LBP) or wavelet transform method.

Preferably, the vein is divided into a region including the vein and the skin region when the average of the brightness values of the local region is larger than a first defined value, and the standard deviation is smaller than a second value defined, and divided into skin regions. When the average of the values is smaller than the first value and the standard deviation is larger than the second value, the values are divided into regions including veins.

Preferably, the vein and skin regions may be classified into at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier. It is characterized by using.

Bio-recognition method using finger vein recognition according to the present invention as described above has the following effects.

First, since it is based on the method that does not go through the vein extraction step, it is possible to reduce the time required for recognition by reducing the processing time to improve the convenience of the user attempting the recognition and the efficiency of the finger vein recognition system.

Second, since the vein pattern is not extracted and only the validity is reflected, the vein pattern is not significantly influenced by the shadow or the noise of the image, and since the vein is included in the local area unit, the recognition accuracy is high.

Third, since it is based on the feature extraction method that is not affected by the brightness value for the local area, it can be used to compare uneven lighting conditions or images with different brightness.

Fourth, without accurately extracting the position of the vein, it is determined whether the vein is included in the unit of the local area to reflect the validity as a weight, it is possible to improve the recognition accuracy.

Other objects, features and advantages of the present invention will become apparent from the following detailed description of embodiments with reference to the accompanying drawings.

Referring to the accompanying drawings, a preferred embodiment of a biometric recognition method using finger vein recognition according to the present invention will be described. However, the present invention is not limited to the embodiments disclosed below, but can be embodied in various forms, and only the present embodiments are intended to complete the disclosure of the present invention and to those skilled in the art to fully understand the scope of the invention. It is provided to inform you.

First Example

1 is a flowchart illustrating a biometric recognition method using finger vein recognition according to a first embodiment of the present invention.

Referring to FIG. 1, first, a finger vein image of a user is acquired using a biometric information acquisition device including an infrared light and a camera (S10). At this time, the biometric information acquisition device is a well-known technique, and a detailed description of the method for acquiring a finger vein image of the user is omitted.

After normalizing the acquired finger vein image to a constant size (S20), a feature value for finger vein recognition using a local binary pattern (LBP) or wavelet transform method in a local region of a predetermined size It is extracted (S30).

The average vein image and the standard deviation of the brightness value of each local area are analyzed by analyzing the finger vein image acquired in the local area unit (S40). In this case, it is preferable to improve the accuracy of the analysis by allowing the local area to be analyzed while being overlapped with another neighboring local area.

Calculation of the mean and standard deviation of the brightness values is shown in FIG. 3, where the skin region 10 has a generally bright and uniform brightness value, and the region 20 containing veins has a dark and uneven brightness value. Has a distribution. Therefore, the skin region 10 has a large average of brightness values, a small standard deviation, and the region 20 including veins has a small average of brightness values and a large standard deviation.

Based on this characteristic, the skin region 10 and the vein region 20 in the finger vein image are analyzed, respectively, and the distribution in the plane with the standard deviation and the mean of brightness values as two axes as shown in FIG. 4. Have

However, since the characteristics of the skin region 10 and the region 20 including the vein may not be clearly distinguished, the two distributions cannot be clearly distinguished by linear dividing lines. Therefore, the distribution can be distinguished using at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier that finds an optimal nonlinear divider. have.

The vein obtained based on the dividing line 30 obtained experimentally through at least one of the support vector machine (SVM) classifier, neural network or K-means clustering, support vector data description (SVDD), and polynomial classifier. By analyzing the local area of the image, the average and standard deviation of the brightness values are calculated, and if the area is located at the top of the previously obtained dividing line, the area is included in the vein, and if it is located at the bottom, it is determined as the skin area (S50).

Through the area determined based on the average and standard deviation of the brightness values (S50), if the area containing the vein is determined as a valid area (valid), a relatively high weight to the feature value extracted from the area Grant and apply it when matching for performing personal authentication (S60).

In addition, through the region determined based on the average and the standard deviation of the brightness value (S50), in the case of the skin region, it is determined as an invalid region (invalid) and gives a relatively low weight to the feature value extracted from the region. And, when matching for performing personal authentication is applied to this (S70).

As an embodiment, in the case of the region including the vein through the determined region, the feature value extracted from the region is used as it is when matching for performing personal authentication, and in the case of the skin region, the characteristic value extracted from the region is used. Multiply by zero to avoid using feature values in matching to perform personal authentication.

In this way, when different feature values are applied according to the image, the reflectance of the feature values extracted from the region where the vein is determined to be located is increased, and the feature values of the other regions are lowered. You can get the feature value.

Then, personal authentication is performed by matching the previously stored bio information with the acquired finger vein image based on the feature value reflecting the weight (S80).

2nd Example

2 is a flowchart illustrating a biorecognition method using finger vein recognition according to a second embodiment of the present invention.

Referring to FIG. 2, first, a finger vein image of a user is acquired using a biometric information acquisition device including an infrared light and a camera (S100). At this time, the biometric information acquisition device is a well-known technique, and a detailed description of the method for acquiring a finger vein image of the user is omitted.

After normalizing the acquired finger vein image to a constant size image (S200), a feature value for finger vein recognition using a local binary pattern (LBP) or wavelet transform method in a local region of a predetermined size Extract the (S300).

In addition, the average vein image and the standard deviation of brightness values of each local region are calculated by analyzing the acquired vein image in the unit of the local region (S400). In this case, it is preferable to improve the accuracy of the analysis by allowing the local area to be analyzed while being overlapped with another neighboring local area.

Calculation of the mean and standard deviation of the brightness value, as shown in Figure 3, the skin region 10 has a generally bright and uniform brightness value, the region 20 containing veins of the dark and non-uniform brightness value of Has a distribution. Therefore, the skin region 10 has a large average of brightness values, a small standard deviation, and the region 20 including veins has a small average of brightness values and a large standard deviation.

Based on this characteristic, the skin region 10 and the vein region 20 in the finger vein image are analyzed, respectively, as shown in FIG. 5. Has

However, since the characteristics of the skin region 10 and the region 20 including the vein may not be clearly distinguished, the two distributions cannot be clearly distinguished by linear dividing lines. Therefore, the distribution can be distinguished using at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier that finds an optimal nonlinear divider. (S500).

In this case, the following two methods are used to classify each distribution using at least one of the support vector machine (SVM) classifier, neural network or K-means clustering, support vector data description (SVDD), and polynomial classifier. Method can be used.

The first method is to classify a region using at least one of a hierarchical support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier.

That is, first, the divided non-venous area is divided into two areas based on the first reference value preset in the classifier, and then the divided non-venous areas are divided again into two areas based on the second reference value preset in the classifier. Area: skin area). It should be noted that the first classifier and the second classifier are not limited in order.

Preferably, the first reference value is a value between -1 and -0.7 and is divided into non-venous areas when the output value of the classifier is included in the first reference value, and when the output value of the classifier is included outside the first reference value. do. The second reference value is a value between 1 and 0.5. When the output value of the classifier is included in the second reference value, the second reference value is divided into skin regions. When the output value of the classifier is included outside the first reference value and the second reference value, the second reference value is divided into an indeterminate area. . In this case, it is to be understood that the first reference value and the second reference value may be various embodiments within the scope of the technical idea of the present invention as preferred embodiments.

The second method is to classify the area according to a value calculated through at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier. It is a way.

That is, after calculating a value in each local area through at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier, the calculated value is calculated. According to the value, it is classified into vein area, skin area and microcrystalline area. For reference, local areas with values calculated between -1 and -0.7 are classified as vein areas, values calculated between 0.5 and 1 are classified as skin areas, and values calculated between 0.5 and -0.7 are classified as undetermined areas. . At this time, the range of the value of each area may be changed within the scope of the technical idea of the present invention, but is not limited thereto.

The region containing the vein through the values obtained using at least one of the SVM (Support Vector Machine) classifier, neural network or K-means clustering, Support Vector Data Description (SVDD), and polynomial classifier (first). The region 20 is classified into an undetermined region (second region) 40 that is not clearly distinguished from the skin region (third region) 10 but overlaps (S500).

In the case of the region (first region) 20 including the vein based on the average and the standard deviation of the brightness values of each local region, the weight is given to the feature value extracted from the region relatively. This is applied when matching for performing personal authentication (S600).

In addition, in the case of the skin region (third region) 10 based on the calculated average and standard deviation of the brightness values of each local region, the weight is relatively low to the feature values extracted from the corresponding region, and personal authentication is performed. This is applied when performing matching (S700).

In addition, in the case of the undetermined region (second region) 40 based on the average and standard deviation of the brightness values of each local region, the weighted value of the first region 20 is applied to the feature value extracted from the region. It is low and gives a weight higher than the weight of the third area 10, and applies it when matching for performing personal authentication (S800).

For example, when the determined region is the first region 20 including the vein, the extracted feature value is multiplied by a weight of 2, and in the case of the third region 40 which is the skin region, the weight is multiplied by 1 and an undecided region. In the case of the second area 10, the weight is multiplied by 1.5 and used for matching to perform personal authentication.

In this way, when different feature values are applied according to the image, the reflectance of the feature values extracted from the region where the vein is determined to be located is increased, and the feature values of the other regions are lowered. You can get the feature value.

Then, personal authentication is performed by matching the previously stored bio information with the acquired finger vein image based on the feature value reflecting the weight (S900).

In such a method, it is possible to obtain a more reliable feature value by increasing the reflectance of the feature value extracted from the region where the vein is determined to be located, and reducing the reflectance of the feature value of the other region. .

3rd Example

On the other hand, unlike the form of multiplying the weight of the feature value calculated from the finger vein image of the user acquired using the recognition information acquisition device as in the first and second embodiments, the feature value and the pre-stored image for registration By comparing the feature values and multiplying the similarity value calculated by the weight, the similarity can be made larger in the case of an image of the same person, and the similarity can be made smaller in the case of an image of another person.

That is, the third embodiment first acquires a finger vein image of a user by using a biometric information acquisition device having an infrared light and a camera. At this time, the biometric information acquisition device is a well-known technique, and a detailed description of the method for acquiring a finger vein image of the user is omitted.

After normalizing the acquired finger vein image to a constant size image, the feature value for finger vein recognition is extracted using a local binary pattern (LBP) or wavelet transform method in a local region of a predetermined size. do.

The average vein image of the local area is analyzed to calculate the average and standard deviation of the brightness values of each local area. In this case, it is preferable to improve the accuracy of the analysis by allowing the local area to be analyzed while being overlapped with another neighboring local area.

Calculation of the mean and standard deviation of the brightness value, as shown in Figure 3, the skin region has a generally bright and uniform brightness value, the region containing the vein has a dark and non-uniform distribution of brightness values. Therefore, the skin region is characterized by a large average of brightness values, a small standard deviation, and a region including a vein having a small average of brightness values and a large standard deviation.

Based on this feature, the region including the skin region and the vein in the finger vein image is analyzed, respectively, and has a standard deviation and mean of brightness values in a plane having two axes as shown in FIG. 4.

However, since the characteristics of the skin region 10 and the region 20 including the vein may not be clearly distinguished, the two distributions cannot be clearly distinguished by linear dividing lines. Therefore, the distribution can be distinguished using at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier that finds an optimal nonlinear divider. have.

Based on the experimentally obtained dividing line 30 through the SVM, the local region of the acquired vein image is analyzed to calculate the average and standard deviation of the brightness values, and the vein is included when the region is located above the previously obtained dividing line. If it is located at the bottom, it is determined as the skin area.

Thereafter, the feature value extracted for each local area from the finger vein image of the user acquired using the biometric information acquisition device and the feature value extracted for each local area from the finger vein image of the user previously stored in the storage unit for registration. Are compared with each other to yield a similarity value.

The area determined based on the average and standard deviation of the brightness values is determined to be a valid area in the case of a vein-containing area, and a relatively high weight is given to the similarity value calculated in the corresponding area. This applies when matching for personal authentication.

In addition, through the area determined based on the average and standard deviation of the brightness value, it is determined to be an invalid area (invalid) in the case of the skin area, and a relatively low weight is given to the similarity value calculated in the corresponding area. This applies when matching to perform authentication.

For example, in the case of the region including the vein through the determined region, the similarity value calculated in the corresponding region is used as it is when matching for performing personal authentication, and in the case of the skin region, the similarity value calculated in the corresponding region is used. By multiplying by zero, the similarity value is not used when matching to perform personal authentication.

As such, when different similarity values are applied according to images, the similarity is larger in the case of the image of the same person, and the similarity can be made smaller in the case of the image of another person, thereby obtaining a more reliable similarity value.

Personal authentication is performed by matching previously stored bioinformation with acquired finger vein images based on similarity values reflecting weights.

Although the technical spirit of the present invention described above has been described in detail in a preferred embodiment, it should be noted that the above-described embodiment is for the purpose of description and not of limitation. In addition, those skilled in the art will understand that various embodiments are possible within the scope of the technical idea of the present invention. Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

1 is a flowchart illustrating a biometric recognition method using finger vein recognition according to a first embodiment of the present invention.

2 is a flowchart illustrating a biometric recognition method using finger vein recognition according to a second embodiment of the present invention.

3 is an embodiment showing a region including and not including a vein according to the present invention

FIG. 4 is an embodiment classified into two stages by SVM and distribution of the mean and standard deviation extracted using the method of FIG.

FIG. 5 is an embodiment classified into three stages by SVM and distribution of mean and standard deviation extracted using the method of FIG.

* Explanation of symbols for main parts of the drawings

10: skin area 20: area including vein

30: dividing line 40: undecided area

Claims (12)

Extracting feature values for finger vein recognition in a local region of a predetermined size based on the finger vein image acquired by using a biometric information acquisition device having an infrared light and a camera; Dividing the vein region and the skin region based on the average and standard deviation of the brightness values of each local region calculated through the finger vein images acquired by the local region; As a result of the classification, in the case of the region including the vein, a first weight is given to the feature value extracted from the corresponding region, and in the case of the skin region, the weight is higher than the first weight relative to the feature value extracted from the region. Assigning a low second weight, And performing personal authentication by matching previously stored bioinformation with the acquired finger vein image based on the feature value reflecting the weight. Extracting feature values for finger vein recognition in a local region of a predetermined size based on the finger vein image acquired by using a biometric information acquisition device having an infrared light and a camera; Dividing the region including the vein, the skin region and the undecided region based on the average and standard deviation of the brightness values of each local region calculated through the finger vein images acquired in the unit of the local region; As a result of the classification, in the case of the region including the vein, the first weight is given to the feature value extracted from the corresponding region, and in the case of the skin region, the second weight is relatively low to the feature value extracted from the region. Assigning a third weight lower than the first weight and higher than the second weight in the case of an undetermined region; And performing personal authentication by matching previously stored bioinformation with the acquired finger vein image based on the feature value reflecting the weight. Extracting feature values for finger vein recognition in a local region of a predetermined size based on the finger vein image acquired by using a biometric information acquisition device having an infrared light and a camera; Dividing the vein region and the skin region based on the average and standard deviation of the brightness values of each local region calculated through the finger vein images acquired by the local region; Calculating a similarity value by comparing feature values extracted for each local area from the acquired finger vein image of the user and feature values extracted for each local area from the finger vein image of the user previously stored in a storage unit for registration; Wow, In the case of the region including the vein, the first weight is given a relatively high first weight to the similarity value calculated in the region, and in the case of the skin region, the first weight is lower than the first weight relative to the similarity value calculated in the region. 2 weighting, And performing personal authentication by matching previously stored bioinformation with the acquired user's finger vein image based on the similarity value reflecting the weight. The method according to any one of claims 1 to 3, The feature value is extracted using a local binary pattern (LBP) or wavelet transform (wavelet transform) method, characterized in that the biometric recognition method using finger vein recognition. The method according to any one of claims 1 to 3, The analysis of the finger vein image by the local area unit is a bio-recognition method using the finger vein recognition, characterized in that the analysis is overlapping (overlap) with another neighboring local area. The method according to any one of claims 1 to 3, The division into the area containing the vein and the skin area is If the average of the brightness values of the local area is greater than the first value defined and the standard deviation is less than the defined second value, the average value of the brightness values is smaller than the first value, and the standard deviation is less than the first value. If greater than 2, biorecognition method using finger vein recognition, characterized in that divided into areas containing veins. The method according to any one of claims 1 to 3, The vein area and the skin area may be classified using at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier. Bio-recognition method using finger vein recognition, characterized in that. The method of claim 2, wherein the division between the vein-containing region and the skin region and the undetermined region is Dividing the vein region and the non-vein region based on a first reference value preset in the classifier; And dividing the separated non-venous area into an undetermined area and a skin area based on a second reference value preset in the classifier. The method of claim 8, The first reference value is a value between -1 and -0.7, and the output value of the classifier is included in the first reference value. The first reference value is divided into the non-venous area when the output value of the classifier is included in the first reference value. The second reference value is a value between 1 and 0.5. When the output value of the classifier is included in the second reference value, the second reference value is divided into skin regions. When the output value of the classifier is included outside the first reference value and the second reference value, the second reference value is divided into an indeterminate area. Bio-recognition method using finger vein recognition. The method of claim 2, wherein the division between the vein-containing region and the skin region and the undetermined region is Calculating a value according to each local area through at least one of a support vector machine (SVM) classifier, a neural network or K-means clustering, a support vector data description (SVDD), and a polynomial classifier; Based on the calculated value, the local area having a value calculated in the first calculated value is classified into the vein area, the value calculated in the second calculated value is classified into the skin area, and the value calculated in the third calculated value is calculated. A biometric recognition method using finger vein recognition, characterized in that it comprises the step of classifying into the undetermined region. The method of claim 10, The first calculated value comprises a value between -1 and -0.7, the second calculated value comprises a value between 0.5 and 1, and including a value between 0.5 and -0.7 into a third calculated value. Bio-recognition method using finger vein recognition. The method of claim 2, The first weight is a value obtained by multiplying the extracted feature value by 2, the second weight is a value multiplied by 0 by the extracted feature value, and the third weight is a value obtained by multiplying the extracted feature value by 1 Bio recognition method using finger vein recognition.

Images