WO2012138004A1 - Fingerprint authentication device using pca, and method therefor - Google Patents

Abstract

The present invention relates to a fingerprint authentication device, and to a fingerprint authentication method therefor. The fingerprint authentication device according to the present invention includes: a database unit for storing a registered fingerprint image; a unit for inputting a fingerprint image for receiving, from a user, an image of a fingerprint of the user; a feature-based fingerprint-matching unit for extracting a feature from the image of the fingerprint of the user inputted through the unit for inputting a fingerprint image and the registered fingerprint image stored in the database unit, and performing a feature-based fingerprint-matching operation on the basis of the extracted feature; a principal component analysis (PCA)-based fingerprint-matching unit for performing a fingerprint-matching operation through a PCA method on the basis of the feature extracted from the image of the fingerprint of the user and the registered fingerprint image; and an authenticating unit for performing a fingerprint authentication operation on the basis of the fingerprint-matching operation performed by the feature-based fingerprint-matching unit and the fingerprint-matching operation performed by the PCA-based fingerprint-matching unit.

Description

Fingerprint authentication device using PCA and its method

The present invention relates to a fingerprint authentication device using PCA (Principal Component Analysis) and its authentication method. More specifically, by performing two fingerprint comparison processes using the fingerprint feature point and the image information of the fingerprint, combining the results of the two methods enables more accurate authentication, and by using the PCA to calculate the amount of computation required in the image comparison The present invention relates to a fingerprint authentication device and a method for reducing the real-time processing.

Recently, attempts have been made to incorporate biometrics into computer systems. The biggest feature of these biometrics is that there is no fear of loss, theft, forgetting, or duplication by external factors under any circumstances. The advantage is that it can be built completely.

In particular, the user authentication technology using a fingerprint has been actively commercialized in recent years, and the user authentication using such a fingerprint has a merit that it is easy to access and carry by authenticating a user by using a unique characteristic of a person. Various studies are underway and many developments are being made.

In the case of authentication technology using a fingerprint, a user's fingerprint is typically registered in a security token such as a smart card or a USB token, and the user is authenticated by comparing the fingerprint with the fingerprint in the security token without leaking the registration fingerprint. If only the result is transmitted to the outside, it is more secure than a system that manages fingerprint data with a central database or a system that transmits registered fingerprint data to perform external fingerprint comparison from an external device. By blocking the can ensure the stability of the biological data that can cause a big problem in the outflow.

In general, fingerprint authentication methods are classified into image-based fingerprint authentication methods and feature-based fingerprint authentication methods. Feature-based fingerprint authentication is a universal method that consists of two processes: minutiae extraction and matching. Various image processing techniques such as smoothing, separation of foreground and background areas, binarization, and thinning Apply the spatial features of the extracted feature points. That is, the feature-based fingerprint authentication method is a method of comparing an authentication fingerprint and a registration fingerprint by using a feature point, and a method of checking similarity between two feature points is mainly used.

The image-based fingerprint authentication method uses ridge information. The overall directionality of the fingerprint image is applied by applying techniques such as Gaber filter, Fast Fourier Transform (FFT), slope, directional histogram, and projection. It is one of the classic ways of using information. In a typical ridge-based fingerprint authentication method, the entire fingerprint area is divided into small areas, the direction value of the ridges existing in the small areas is extracted to complete the directional map of the fingerprint, and the two directional maps are convolved to determine the degree of similarity. .

However, in the case of the feature-based fingerprint authentication method, false rejection of a fingerprint may occur due to a missing feature point, as well as a feature point distribution extracted from two different fingerprints similarly due to the missing and incorrectly extracted feature point. Since this may cause a wrong authentication for another fingerprint, there is a problem that there is a limit in the recognition performance, such as FAR.

In addition, the ridge-based fingerprint authentication method requires a lot of storage space for storing the image because the entire fingerprint image is used, and the performance of the fingerprint authentication process is slow.

The present invention was devised to solve the above problems, and after the two fingerprint comparison process using the fingerprint feature point and the image information of the fingerprint to combine the results of the two methods to enable more accurate authentication, and also PCA It is an object of the present invention to provide a fingerprint authentication device and a method for reducing the amount of calculation required for image comparison to enable real-time processing.

Fingerprint authentication device according to the present invention for achieving the above object, the database unit for storing the registered fingerprint image; A fingerprint image input unit configured to receive a user fingerprint image from a user; A feature point based fingerprint matching unit extracting a feature point from the user fingerprint image inputted through the fingerprint image input unit and the registered fingerprint image stored in the database unit, and performing a feature point based fingerprint matching based on the extracted feature point; A PCA-based fingerprint matching unit performing fingerprint matching based on a principal component analysis (PCA) method based on feature points extracted from the user fingerprint image and the registered fingerprint image; And an authentication performing unit configured to perform fingerprint authentication based on fingerprint matching by the feature-point based fingerprint matching unit and fingerprint matching by the PCA-based fingerprint matching unit.

The fingerprint authentication device may further include a similarity coupling unit that combines the similarity by the feature-based fingerprint matching unit and the similarity by the PCA-based fingerprint matching unit. In this case, the authentication performing unit performs fingerprint authentication when the similarity combined by the similarity combining unit is equal to or greater than a set value.

The fingerprint authentication apparatus may further include a hash table generator for generating a geometric hash table by extracting a region of interest (RoI) for extracting image feature information after applying geometric hashing to a feature point of the user fingerprint image. have.

The fingerprint authentication apparatus may further include a feature point corrector configured to correct an authentication fingerprint by the authentication performer and a feature point of the registered fingerprint based on at least one of the detected position and the direction difference.

On the other hand, the fingerprint authentication method according to the invention, storing the registered fingerprint image; Receiving a user fingerprint image from a user; Extracting a feature point from the user fingerprint image input through the fingerprint image input unit and the registered fingerprint image stored in the database unit, and performing feature-point based fingerprint matching based on the extracted feature point; Performing fingerprint matching by a principal component analysis (PCA) method based on feature points extracted from the user fingerprint image and the registered fingerprint image; And performing fingerprint authentication based on fingerprint matching by the feature-point based fingerprint matching performing step and fingerprint matching by the PCA-based fingerprint matching performing step.

The fingerprint authentication method may further include combining the similarity by the feature-point based fingerprint matching step and the similarity by the PCA-based fingerprint matching step. In this case, the fingerprint authentication performing step may perform fingerprint authentication when the similarity combined by the similarity combining step is equal to or greater than a set value.

The fingerprint authentication method may generate a geometric hash table by extracting a RoI for extracting image feature information after applying geometric hashing to a feature point of the user fingerprint image.

The fingerprint authentication method may correct feature points of the authentication fingerprint and the registration fingerprint by the authentication performing step based on at least one of the detected position and the direction difference.

According to the present invention, more accurate authentication is possible by performing two fingerprint comparison processes using fingerprint feature points and fingerprint image information, and by using PCA, the computation amount required for image comparison is reduced, thereby enabling real-time processing. .

1 is a view schematically showing a fingerprint authentication device according to an embodiment of the present invention.

2 is a diagram showing an example by bin accumulation as an example of feature point correction.

3 is a diagram illustrating an example of applying a geometric hashing method to feature points.

4 is a diagram illustrating an example of extracting RoI from a fingerprint.

FIG. 5 is a diagram illustrating an example of generating a RoI geometric hash table without requiring a correction process by using a hashing hash on a single fingerprint.

6 is a flowchart illustrating the generation and excess of RoI geometric hash table in a fingerprint image.

7 is a flowchart illustrating a fingerprint authentication method according to the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description, the detailed description can be omitted for techniques well known to those skilled in the art.

In addition, in describing the components of the present invention, different reference numerals may be given to components having the same name according to the drawings, and the same reference numerals may be given even though they are different drawings. However, even in such a case, it does not mean that the corresponding components have different functions according to the embodiments, or does not mean that they have the same functions in different embodiments, and the functions of the respective components may be implemented. Judgment should be made based on the description of each component in the example.

In addition, in describing the embodiments of the present invention, when it is determined that the detailed description of the related well-known configuration or function may obscure the gist of the present invention, the detailed description may be omitted.

In addition, if a component is described as being "connected", "coupled" or "connected" to another component, the component may be directly connected or connected to the other component, but there is another It is to be understood that the components may be "connected", "coupled" or "connected".

1 is a view schematically showing a fingerprint authentication device according to an embodiment of the present invention. Referring to the drawings, the fingerprint authentication apparatus 100 includes a database unit 110, a fingerprint image input unit 120, a feature point based fingerprint matching unit 130, a PCA based fingerprint matching unit 140, an authentication performing unit 150, The similarity combiner 160, a hash table generator 170, and a feature point corrector 180 may be included.

The database unit 110 stores the registered fingerprint image. In this case, the database unit 110 may store the feature point information analyzed with respect to the registered fingerprint image and the image feature information of the fingerprint for PCA (Principal Component Analysis). Principal component analysis in statistics is one of the techniques for analyzing data sets. Principal component analysis shows that when the data is mapped onto one axis, the axis with the largest variance comes to the first coordinate axis, the second largest axis comes second, the third largest axis comes third, and so on. Linearly transform the data into a new coordinate system. Thus, various applications are possible by placing the "most important" component of the data on each axis. Principal component analysis is sometimes referred to as the Karhunen-Loeve transform (KLT), named after Kari Karunen and Michelle Loeb, or the Hotelling transform, named after Harold Hotelling. Principal component analysis is characterized by an optimal linear transformation that preserves the subspace with the largest variance. However, there is a disadvantage in that it requires more computation time than other methods such as discrete cosine transform. Unlike other linear transformations, principal component analysis does not have a fixed basis vector, which depends on the nature of the data.

One fingerprint curve representing a fingerprint is called a ridge, which is a continuous ridge, an end that breaks in the middle, and a bifurcation where two or more ridges meet, which are called minutiae. do. In order to apply the PCA technique, an average image of a region of interest (RoI) of a fingerprint and a collection of eigenvalues and eigenvectors, which are representable values and vectors of the RoI images, are called image feature information. RoI generally represents a data sample selected for another additional purpose within a series of data sets with a particular purpose. In an embodiment of the present invention, a region of interest (RoI) may be designated for a fingerprint image, and filtering may be applied only to that region. At this time, the RoI extracts some features from the input image to form an initial specific map, and then performs spatial feature analysis method that performs coordinate-centered calculation based on azimuth with respect to the configured specific map, and the current frame (t) and the previous one. After receiving a continuous contrast image sequence consisting of frames (t-1) and extracting time-specific motion information, a temporal feature analysis method may be used to extract a motion vector from the image sequence using a block matching technique.

Since the feature point and the image feature information differ from person to person, the database unit 110 analyzes the location and number of feature points on the fingerprint image obtained from the registered subscriber, extracts the image feature information, and then extracts the extracted image feature information from the database. Save it.

The fingerprint image input unit 120 receives a user fingerprint image from a user for authentication.

The feature point-based fingerprint matching unit 130 extracts the feature point and the image feature information from the fingerprint image of the user for authentication input through the fingerprint image input unit 120 and the registered fingerprint image stored in the database unit 110, and extracts the extracted feature point. A feature-based fingerprint matching is performed based on.

In general, the fingerprint feature information used in the fingerprint recognition system uses an ending point through which the ridge passes and a bifurcation point where one ridge is divided in the fingerprint image. One feature point extracted from the fingerprint image may have a coordinate of the feature point and type information of the feature point and may be represented by (x, y, θ, type). On the other hand, even if the fingerprint of the same person every time the fingerprint is input to the fingerprint input device, the coordinate value of the feature point is translated and the direction is rotated. In this way, all the feature points extracted from the two fingerprints can be identified based on the difference in position and direction, and all pairs of feature points determined to be similar can be found, or scores can be obtained based on the similarity to determine whether the threshold is greater than or equal to the threshold. This is a method of comparing absolute phases assuming that fingerprints are input in the same position and the same direction to be input, and it is necessary to correct the two fingerprints to have the same criteria before checking the similarity of the position and the direction. That is, even in the same pair of feature points, an absolute coordinate value is different in direction according to the input time, and thus, an alignment process is required in which the two fingerprints are moved and rotated by a changed amount. To this end, the feature point corrector 130 may be configured to calculate the position, direction, distance, and angle of each feature point extracted from the registered fingerprint image stored in the database unit 110 and the fingerprint image of the user input through the fingerprint image input unit 120. Correct the difference for at least one.

The feature point-based fingerprint matching unit 130 extracts feature point information for authentication from the user fingerprint image input through the fingerprint image input unit 120, and then stores the feature point information on the registered fingerprint image previously stored and registered in the database unit 110. Read and load the data into a memory to compare and analyze the registered feature information and the feature information of the user fingerprint image. In this case, the feature point corrector 180 extracts a difference of at least one of a position, a direction, a distance, and an angle between the feature point information of the registered fingerprint loaded in the memory and the feature point information of the authentication fingerprint to correct the feature point. As shown in FIG. 2B, an area for accumulating the difference between the position, the direction, and the distance between the two feature points may be set, and the positions bx1, by1, and the direction unit of each bin may be determined. Here, the unit of the initial bin with respect to the position and the direction may be set to be large and then reduced to the minimum unit while changing sequentially. In FIGS. 2A and 2B, Range_X1 is associated with a detectable positional shift in the X direction to the first area of the X axis, Range_X2 is set to the area around the maximum bin at the first stage as the second area of the X axis, and Range_Y1 Is associated with the detectable positional movement in the Y direction to the first region of the Y axis, and Range_Y2 is the second region of the Y axis, set as the region around the maximum bin at the first stage. In addition, bx1 is the first unit of the X-axis, bx2 is the second unit of the X-axis, smaller than bx1, by1 is the first unit of the Y-axis, and by2 is the second unit of the Y-axis, which is smaller than by1. ΔX is the amount of change in the X-axis of the pair of feature points, ΔY is the amount of change in the Y-axis of the pair of feature points, and B means a bin having a maximum value. In addition, a bin means an area of each feature point expressed in binary.

It is determined whether all of the feature points between the feature point information of the user fingerprint image and the feature point information of the loaded registered fingerprint image are considered.If all feature points are not taken into consideration, a pair of authentication fingerprint feature points and a registration fingerprint feature point are generated. The direction difference can be measured. In this case, by checking whether the direction difference between the two feature points is out of the allowable range, if the deviation is not out of the allowable range, the position difference and the distance difference between the two feature points may be measured. Similarly, if the position difference and distance difference between the two feature points do not deviate from the allowable range, the level of the specific bin of the set area corresponding to the obtained position, direction and distance difference is raised or accumulated, and all the feature points, ie, the direction, After considering the feature point pairs of the position and distance difference, the bin having the maximum level is obtained, and whether the unit for the position, direction, and distance of the bin is the minimum unit is determined. If the unit is the minimum unit, the difference in the position, direction, and distance corresponding to the maximum bin value, that is, the position, direction, and distance difference corresponding to B shown in FIG. 2B may be extracted. At this time, the feature point corrector corrects the feature points of the authentication fingerprint and the enrolled fingerprint based on the detected position, direction, and distance difference, and then measures the similarity between the two corrected feature points, if the measured similarity is greater than the preset threshold. It is determined that this is the same, and the authentication process of the user is performed according to the determined result.

The PCA-based fingerprint matching unit 140 performs fingerprint matching by a principal component analysis (PCA) method based on feature points extracted from a user fingerprint image and a registered fingerprint image. The fingerprint matching method based on principal component analysis calculates the eigenvalues and eigenvectors from the covariance matrix obtained by converting them into one-dimensional fingerprint vectors and calculates the eigenvectors corresponding to the eigenvalues arranged in size order. The calculated eigenvectors are basis vectors representing one fingerprint image. In the actual recognition step, the fingerprint is recognized by comparing the feature vectors obtained by linearly transforming the input fingerprint image with the feature vectors obtained in advance. Meanwhile, other methods related to PCA may be used, for example, independent component analysis (ICA) and Kemel PCA based methods. Bartlett et al. Showed that ICA was superior to PCA when cosine was used as a similarity measure, and Yang showed that the Kemel Eigenface method is superior to the existing Eigenface method using Kemel PCA for feature extraction and recognition. However, both ICA and Kemel PCA methods have a disadvantage in that the computational amount is larger than that of the PCA method.

3 shows a process of extracting image feature information for PCA. The image feature information always needs the same RoI and the corrected image. Therefore, the feature point extraction is performed from the fingerprint input through the fingerprint input process 310 (320), and the correction process of the fingerprint is solved using the geometric hashing (330). Geometric hashing is a method of pre-forming a feature point table in which correction is completed by applying all correction values that can be generated by the extracted feature points as shown in FIG. 4. To this end, the hash table generator 170 may apply the geometric hashing to the feature points of the user fingerprint image and then extract the RoI for extracting the image feature information to generate the geometric hash table. In this way, if the corrected feature point table is prepared in advance, fingerprint authentication is possible by only comparing the feature point tables without additional correction process. After applying geometric hashing 330 to a feature point, RoI extraction for image feature information extraction is performed (340). Fingerprints are rotated and moved each time, so you cannot always get the same image. Therefore, as shown in FIG. 5, it is preferable to extract RoI using only a partial image around a feature point. When the RoI is extracted 340 using the corrected feature point through the geometric hashing 330, the same RoI is extracted from the image having the same feature point. The geometric hashing table 330 and RoI extraction 340 generate a geometric hashing table as shown in FIG. 6. If n feature points are extracted from one fingerprint, the feature has n correction values in the geometric hashing process 330. Therefore, the n number of tables are generated RoI n. When the geometric hashing table as shown in FIG. 6 is applied to the PCA technique which is generally used for pattern recognition, an average image, an eigenvalue, and an eigenvector, which are image feature information, are generated. Since the PCA has a characteristic of converting high-dimensional information into low-dimensional information, the data size of the image characteristic information can be reduced.

The image feature jumbo generated in this manner is stored in the database unit 110 and fingerprint matching is performed by the PCA-based fingerprint matching unit 140 on the user fingerprint image input through the fingerprint image input unit 120.

The authentication performing unit 150 performs fingerprint authentication by two fingerprint comparison processes based on the fingerprint matching by the feature-based fingerprint matching unit 130 and the fingerprint matching by the PCA-based fingerprint matching unit 140.

The similarity of the fingerprints extracted from the feature point-based fingerprint matching unit 130 and the PCA-based fingerprint matching unit 140 may be combined in the similarity combining unit 160 to perform fingerprint authentication in the authentication performing unit 150. In this case, the authentication performing unit 150 may be implemented to perform fingerprint authentication when the similarity combined by the similarity combining unit 160 is equal to or greater than a set value. In this case, the similarity combiner 160 sets the similarity range of the feature-based fingerprint matching by the feature-based fingerprint matching unit 130 and the similarity range of the fingerprint matching by the PCA-based fingerprint matching unit 140, respectively. Fingerprint matching may be performed by the authentication performing unit 150 by selecting a fingerprint matching that is included in the range of similarity by the fingerprint matching method. For example, the similarity range of the fingerprint registration by the feature-based fingerprint matching unit 130 is set to 90% to 100%, and the similarity range of the fingerprint matching by the PCA-based fingerprint matching unit 140 is 80% to 100%. In this case, fingerprint authentication may be performed only for fingerprint matching having a similarity of 90% to 100%. Alternatively, the similarity of the fingerprint matching determined by the feature point-based fingerprint matching unit 130 and the similarity of the fingerprint matching determined by the PCA-based fingerprint matching unit 140 may be used as judgment data for performing fingerprint authentication. have. For example, when it is determined that the same fingerprint by the fingerprint matching is 1, if the similarity of the fingerprint authentication performance criteria by the authentication performing unit 150 is 1.8 to 2.2, the feature-based fingerprint matching unit 130 When the similarity determined by 0.9 and the similarity by the PCA-based fingerprint matching unit 140 are 0.93, the sum of the similarities by the two fingerprint matching methods is 1.83, so that fingerprint authentication by the authentication performing unit 150 may be performed. have.

7 is a flowchart illustrating a fingerprint authentication method according to the present invention.

1 and 7, the database unit 110 stores a registered fingerprint image (S710). In this case, the database unit 110 may store the feature point information analyzed with respect to the registered fingerprint image and the image feature information of the fingerprint for PCA (Principal Component Analysis).

The fingerprint image input unit 120 receives a user fingerprint image from the user for authentication (S720).

The feature point-based fingerprint matching unit 130 extracts the feature point and the image feature information from the fingerprint image of the user for authentication inputted through the fingerprint image input unit 120 and the registered fingerprint image stored in the database unit 110, and extracts the extracted feature points. Based on the feature point based fingerprint matching is performed (S730).

The PCA-based fingerprint matching unit 140 performs fingerprint matching by a principal component analysis (PCA) method based on the feature points extracted from the user fingerprint image and the registered fingerprint image (S740).

The similarity combiner 160 combines the similarity by the feature-based fingerprint matching unit 130 and the similarity by the PCA-based fingerprint matching unit 140 (S750). The similarity combining unit 160 sets the similarity range of the feature-based fingerprint matching by the feature-based fingerprint matching unit 130 and the similarity range of the fingerprint matching by the PCA-based fingerprint matching unit 140, respectively. By selecting the fingerprint matching all included in the range of similarity by the fingerprint authentication by the authentication performing unit 150 may be performed. For example, the similarity range of the fingerprint registration by the feature-based fingerprint matching unit 130 is set to 90% to 100%, and the similarity range of the fingerprint matching by the PCA-based fingerprint matching unit 140 is 80% to 100%. In this case, fingerprint authentication may be performed only for fingerprint matching having a similarity of 90% to 100%. Alternatively, the similarity of the fingerprint matching determined by the feature point-based fingerprint matching unit 130 and the similarity of the fingerprint matching determined by the PCA-based fingerprint matching unit 140 may be used as judgment data for performing fingerprint authentication. have. For example, if it is determined that 1 is the same fingerprint by fingerprint matching, the similarity of the fingerprint authentication performance criteria by the authentication performing unit 150 is 1.8 to 2.2, the feature-based fingerprint matching unit 130 When the similarity determined by 0.9 and the similarity by the PCA-based fingerprint matching unit 140 are 0.93, the sum of the similarities by the two fingerprint matching methods is 1.83, so that fingerprint authentication by the authentication performing unit 150 may be performed. have.

The authentication performing unit 150 performs fingerprint authentication based on the fingerprint matching by the feature-based fingerprint matching unit 130 and the fingerprint matching by the PCA-based fingerprint matching unit 140. In this case, the authentication performing unit 150 may be implemented to perform fingerprint authentication when the similarity combined by the similarity combining unit 160 is equal to or greater than a set value. That is, it is determined that the user fingerprint image and the registered fingerprint image are matched by the fingerprint matching by the feature point-based fingerprint matching unit 130 or the fingerprint matching by the PCA-based fingerprint matching unit 140, or the similarity coupling unit 160. It can be implemented to perform the fingerprint authentication when the similarity combined by the similarity within the error range with respect to the reference value.

In the above description, all elements constituting the embodiments of the present invention are described as being combined or operating in combination, but the present invention is not necessarily limited to these embodiments. In other words, within the scope of the present invention, all of the components may be selectively operated in combination with one or more. In addition, although all of the components may be implemented as one independent hardware, each or some of the components of the program modules are selectively combined to perform some or all of the functions combined in one or a plurality of hardware It may be implemented as a computer program having a. In addition, such a computer program may be stored in a computer readable medium such as a USB memory, a CD disk, a flash memory, and the like, and read and executed by a computer, thereby implementing embodiments of the present invention. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like.

In addition, all terms including technical or scientific terms have the same meaning as commonly understood by a person of ordinary skill in the art unless otherwise defined in the detailed description. Terms used generally, such as terms defined in a dictionary, should be interpreted to coincide with the contextual meaning of the related art, and shall not be interpreted in an ideal or excessively formal sense unless explicitly defined in the present invention.

The above description is merely illustrative of the technical idea of the present invention, and those skilled in the art to which the present invention pertains may make various modifications and changes without departing from the essential characteristics of the present invention. In addition, the embodiments disclosed in the present invention are not intended to limit the technical spirit of the present invention but to explain, and the scope of the technical spirit of the present invention is not limited by these embodiments. Therefore, the protection scope of the present invention should be interpreted by the claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present invention.

Claims (8)

1. In the fingerprint authentication device,

   A database unit for storing a registered fingerprint image;

   A fingerprint image input unit configured to receive a user fingerprint image from a user;

   A feature point based fingerprint matching unit extracting a feature point from the user fingerprint image inputted through the fingerprint image input unit and the registered fingerprint image stored in the database unit, and performing a feature point based fingerprint matching based on the extracted feature point;

   A PCA-based fingerprint matching unit performing fingerprint matching based on a principal component analysis (PCA) method based on feature points extracted from the user fingerprint image and the registered fingerprint image; And

   An authentication performing unit for performing fingerprint authentication based on fingerprint matching by the feature point based fingerprint matching unit and fingerprint matching by the PCA based fingerprint matching unit.

   Fingerprint authentication device comprising a.
2. The method of claim 1,

   A similarity combiner combining the similarity by the feature point based fingerprint matching unit and the similarity by the PCA based fingerprint matching unit

   More,

   And wherein the authentication performing unit performs fingerprint authentication when the similarity combined by the similarity combining unit is equal to or greater than a set value.
3. The method of claim 1,

   A hash table generator for generating a geometric hash table by extracting RoI for extracting feature information after applying geometric hashing to the feature points of the user fingerprint image.

   Fingerprint authentication device further comprising.
4. The method of claim 1,

   A feature point corrector that corrects feature points of the authentication fingerprint and the registered fingerprint based on at least one of the detected position and the direction difference.

   Fingerprint authentication device further comprising.
5. In the fingerprint authentication method,

   Storing the registered fingerprint image;

   Receiving a user fingerprint image from a user;

   Extracting a feature point from the input user fingerprint image and the stored registered fingerprint image and performing a feature point-based fingerprint matching based on the extracted feature point;

   Performing fingerprint matching by a principal component analysis (PCA) method based on feature points extracted from the user fingerprint image and the registered fingerprint image; And

   Performing fingerprint authentication based on fingerprint matching by the feature-point based fingerprint matching step and fingerprint matching by the PCA-based fingerprint matching step

   Fingerprint authentication method comprising a.
6. The method of claim 5,

   Combining the similarity by the feature-based fingerprint matching step and the similarity by the PCA-based fingerprint matching unit

   More,

   The fingerprint authentication step may include performing fingerprint authentication when the similarity coupled by the similarity coupling unit is equal to or greater than a set value.
7. The method of claim 5,

   And applying a geometric hashing to the feature point of the user fingerprint image and extracting a RoI for extracting the image feature information to generate a geometric hash table.
8. The method of claim 5,

   And a feature point of the authentication fingerprint and the registration fingerprint according to the authentication performing step based on at least one of the detected position and the direction difference.

Images