JP2016207207A - Method and apparatus for recognizing fingerprint - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide techniques of authenticating a user using partial bio-images.SOLUTION: A fingerprint recognition method includes: receiving an input image corresponding to bio-information on a user; partitioning the input image into blocks; comparing one or more of the blocks with one or more enrolled partial images corresponding to already enrolled bio-information; and recognizing, identifying and/or authenticating the user on the basis of a result of the comparison.SELECTED DRAWING: Figure 1

Description

  The description of the present invention relates to a fingerprint recognition method and apparatus.

  The authentication technology based on biometric authentication authenticates a user using a fingerprint, iris, voice, face, blood vessel, or the like. The biometric characteristics used for authentication differ from person to person and do not have the inconvenience of possession, but also have the advantage that they do not need to be changed throughout their lives because the risk of theft or counterfeiting is low. In particular, fingerprint recognition methods are currently most frequently commercialized for various reasons such as convenience, security, and economy. For example, the fingerprint recognition method can enhance security for user equipment and easily provide various application services such as mobile payment.

  The present invention has been made in view of the above-described conventional problems, and an object of the present invention is to provide a fingerprint recognition method.

  One embodiment of the present invention corresponds to a step of receiving an input image corresponding to a user's biometric information, a step of dividing the input image into blocks, and biometric information already registered with one or a plurality of the blocks. There is provided a biometric image recognition method comprising: comparing one or more registered partial images; and recognizing, identifying, and / or authenticating the user based on a result of the comparison.

  The embodiment of the present invention can provide a technique for authenticating a user using a partial biological image.

It is a figure explaining the fingerprint image which concerns on one Embodiment of this invention. It is a figure explaining the fingerprint image which concerns on one Embodiment of this invention. It is a figure explaining the fingerprint recognition method concerning one embodiment of the present invention. It is a figure explaining the fingerprint recognition method concerning one embodiment of the present invention. It is a figure explaining the fingerprint recognition method which divides | segments the input partial image which concerns on one Embodiment of this invention into a some block, and compares with a registration partial image. 5 is an operation flowchart illustrating a fingerprint recognition method according to an exemplary embodiment of the present invention. 5 is an operational flowchart illustrating a user authentication method according to an embodiment of the present invention. 6 is an operation flowchart illustrating a frequency-based matching method according to an exemplary embodiment of the present invention. It is a figure explaining Log-Polar conversion concerning one embodiment of the present invention. It is a figure explaining the operation | movement which processes the block of a registration image and an input image in order to calculate the block score which concerns on one Embodiment of this invention. 5 is an operational flowchart illustrating a user authentication method according to another embodiment of the present invention. It is a block diagram which shows the electronic system which concerns on one Embodiment of this invention.

  The specific structural or functional descriptions of the embodiments of the present invention are disclosed merely for the purpose of illustrating the embodiments and may be implemented in various forms. Therefore, the embodiments of the present invention are not limited to specific disclosure forms, and the scope of the present specification includes modifications, equivalents, or alternatives included in the technical idea.

  Terms such as first or second may be used to describe various components, but such terms are to be interpreted for the purpose of distinguishing one component from another. Should be. For example, the first component may be named the second component, and similarly the second component may be named the first component.

  When a component is referred to as being “coupled” to another component, it may be directly coupled to or connected to the other component, but there is another component in between It should be understood that it can.

  The singular expression includes the plural unless the context clearly dictates otherwise. In this specification, terms such as “including” or “having” are intended to indicate that there is a feature, number, step, action, component, part, or combination thereof described in the specification. And one or more other features, numbers, steps, operations, components, parts, or combinations thereof should not be understood as pre-excluding the possibility of existence or addition Don't be.

  Unless defined otherwise, all technical and scientific terms used herein are understood to be generally understood by those having ordinary skill in the art to which the embodiments belong. Have the same meaning. Terms commonly defined as commonly used should be construed as having a meaning consistent with the meaning possessed in the context of the related art and are ideal unless explicitly defined in this application. It is not interpreted as having a formal or overly formal meaning.

  The embodiments of the present invention described below are used to recognize a user's fingerprint. Hereinafter, the operation of recognizing a user's fingerprint includes an operation of authenticating or identifying the user. The operation of authenticating the user may include, for example, an operation of determining whether or not the user is a registered user. In this case, the result of the operation of authenticating the user can be output as true or false. The operation of identifying a user may include, for example, an operation of determining which user the user corresponds to among a plurality of users already registered. In this case, the result of the operation for identifying the user can be output from the ID of any one of the already registered users. If the user does not correspond to any user among a plurality of registered users, a signal notifying that the user is not identified can be output.

  Embodiments of the present invention can be realized in various forms of products such as personal computers, laptop computers, tablet PCs, smartphones, televisions, smart home appliances, intelligent automobiles, kiosks, and wearable devices. For example, the embodiments of the present invention may be applied to authenticate a user in a smartphone, a mobile device, a smart home system, or the like. The embodiment of the present invention may be applied to a payment service using user authentication. The embodiment of the present invention may also be applied to an intelligent vehicle system that can authenticate a user and automatically start an engine. Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. The same reference numerals provided in each drawing denote the same members.

  1 and 2 are views for explaining a fingerprint image according to an embodiment of the present invention.

  Referring to FIG. 1, a fingerprint sensor (not shown) according to an embodiment of the present invention detects a user's fingerprint 100. The fingerprint sensor detects the fingerprint 100 through the detection area. Here, the size of the detection area of the fingerprint sensor may be smaller than the size of the fingerprint 100. For example, the detection area of the fingerprint sensor may have a rectangular shape that is smaller than the size of the fingerprint 100. In this case, the fingerprint sensor can detect a part of the fingerprint 100 through the detection area.

  The fingerprint sensor can generate a fingerprint image by capturing the detected fingerprint. When the size of the detection area of the fingerprint sensor is smaller than the size of the fingerprint 100, the fingerprint image generated by the fingerprint sensor may correspond to a partial image including a part of the fingerprint 100. In addition, according to an embodiment of the present invention, the detection area of the fingerprint sensor may be the same as or larger than the size of the fingerprint 100. In this case, the fingerprint image generated by the fingerprint sensor can correspond to the entire image of the fingerprint 100.

  The fingerprint image is used to register or recognize the user's fingerprint 100. For example, a fingerprint image is registered at the registration stage. The registered fingerprint image is stored in a storage destination provided in advance. When the size of the detection area of the fingerprint sensor is smaller than the size of the fingerprint 100, a plurality of fingerprint images corresponding to the partial image of the user's fingerprint 100 can be registered. For example, a plurality of partial images (110 to 170) may be registered. The partial images (110 to 170) may cover a part of each fingerprint 100, and the partial images (110 to 170) may be collected to cover the fingerprint 100 as a whole. Here, the partial images (110 to 170) may overlap each other. Hereinafter, for convenience of explanation, a partial image of a fingerprint already registered is referred to as a registered partial image.

  Further, the input fingerprint image is recognized at the recognition stage. For example, the fingerprint image input at the recognition stage may be compared with the already registered fingerprint image. The user authentication result or the user identification result may be derived depending on whether or not the inputted fingerprint image matches the already registered fingerprint image. If the size of the detection area of the fingerprint sensor is smaller than the size of the fingerprint 100, the input fingerprint image corresponds to a partial image of the user's fingerprint 100. Hereinafter, for convenience of explanation, a partial image of a user's fingerprint is referred to as an input partial image. Although described in detail below, embodiments of the present invention can provide a scheme for recognizing a fingerprint by comparing an input partial image with a registered partial image.

  Although the detection area of the fingerprint sensor in FIG. 1 has been described as having a rectangular shape, the size and shape of the detection area of the fingerprint sensor may be variously modified. For example, referring to FIG. 2, the detection area of the fingerprint sensor is circular. Also in this case, a plurality of partial images (210 to 295) may be registered corresponding to one fingerprint 200 at the registration stage. In the recognition stage, a fingerprint image corresponding to a part of the fingerprint 200 may be compared with a plurality of already registered partial images (210 to 295).

  In some cases, the fingerprint sensor used in the registration stage and the fingerprint sensor used in the recognition stage may be different from each other. For example, a fingerprint sensor having a rectangular detection area shown in FIG. 1 is used in the registration stage, and a fingerprint sensor having a circular detection area shown in FIG. 2 is used in the recognition stage. May be. The same fingerprint sensor may be used for the registration method and the recognition method.

  3 and 4 are diagrams illustrating a fingerprint recognition method according to an embodiment of the present invention. Referring to FIG. 3, a fingerprint recognition apparatus 300 according to an embodiment of the present invention includes a fingerprint sensor 310. The size of the detection area of the fingerprint sensor 310 may be smaller than the size of the user's fingerprint. The fingerprint recognition device 300 can acquire the input partial image 315 via the fingerprint sensor 310. The fingerprint recognition device 300 can acquire the registered partial images (321 to 323) from the database 320 provided in advance. The database 320 is stored in a memory (not shown) included in the fingerprint recognition device 300, or is connected to an external device (not shown) such as a server that can be connected to the fingerprint recognition device 300 by wired, wireless, or network. It may be stored. According to an embodiment of the present invention, the size and shape of the input partial image may be the same as the size and shape of the registered partial image. For example, if the input partial image 315 and the registered partial images (321 to 323) are all captured using the fingerprint sensor 310, the input partial image 315 has the same size and shape as the registered partial images (321 to 323). Can have.

  The fingerprint recognition apparatus 300 recognizes the user's fingerprint by comparing the input partial image 315 and the registered partial images (321 to 323). Referring to FIG. 4, the fingerprint recognition apparatus 300 can match the input partial image 315 and the registered partial image 323 in order to compare the input partial image 315 and the registered partial image 323. For example, the fingerprint recognition apparatus 300 can scale, rotate, and / or translate the input partial image 315 so that the common areas included in the input partial image 315 and the registered partial image 323 overlap. However, since the input partial image 315 and the registered partial image 323 all correspond to partial images, the size of the overlapping area may be considerably smaller than the size of each of the input partial image 315 and the registered partial image 323. For example, the overlapping area 325 is an area where the partial image 315 and the partial image 323 overlap. As shown in FIG. 4, the size of the overlapping region 325 can be significantly smaller when compared to the overall size of the partial image 315. In this case, matching the input partial image 315 and the registered partial image 323 may be inefficient.

  Further, the input partial image 315 may partially include deterioration due to various factors. For example, there is a possibility that deformation occurs in the fingerprint image due to the pressure of pressing the sensor. When the input partial image 315 is generated, the pressure applied to each part of the detection area of the fingerprint sensor may be different. Therefore, there is a possibility that at least a part of the input partial image 315 is deformed. The registered partial images (321 to 323) may also include deterioration due to various factors. In this case, comparing the input partial image 315 and the registered partial images (321 to 323) may reduce the reliability of fingerprint recognition.

  FIG. 5 is a diagram for explaining a fingerprint recognition method for dividing an input partial image into a plurality of blocks and comparing it with a registered partial image according to an embodiment of the present invention. Referring to FIG. 5, the fingerprint recognition apparatus 300 can divide the input partial image 315 into a plurality of blocks (511 to 514). The operation of dividing the input partial image 315 into a plurality of blocks (511 to 514) may be referred to as block division. The fingerprint recognition apparatus 300 may compare the plurality of blocks (511 to 514) and the registered partial images (321 to 323) instead of comparing the input partial image 315 and the registered partial images (321 to 323). . As shown in FIG. 5, the fingerprint recognition apparatus 300 can divide the input partial image into a plurality of blocks such that each block is smaller than the input partial image. In image division or block division according to an embodiment of the present invention, blocks may be generated such that a plurality of blocks overlap each other, or blocks may be generated so that a plurality of blocks do not overlap each other.

  The fingerprint recognition apparatus 300 can divide the input partial image 315 by various methods. As an example, the fingerprint recognition apparatus 300 may divide the input partial image 315 by a predetermined pattern. The predetermined pattern may be determined in advance according to the shape and size of the detection area of the fingerprint sensor, the shape and size of the registered partial image, and the like. In some cases, the predetermined pattern may be changed dynamically. In addition, the blocks may be divided so as not to overlap each other, or may be divided so as to overlap each other with regions having a certain size.

  The fingerprint recognition device 300 recognizes a fingerprint using block pattern matching. Block pattern matching may include pattern matching of partial fingerprint images. As will be described in detail below, the fingerprint recognition apparatus 300 divides the fingerprint image input via the fingerprint sensor into a plurality of blocks, performs matching based on the frequency, and aligns the matching score of each block to obtain the upper K Whether or not user authentication is possible can be determined using the feature values of the individual blocks. Here, the feature value may be a value indicating a feature regarding the top K blocks. For example, the feature value may include a statistical value such as an average. The fingerprint recognition device 300 can recognize a fingerprint regardless of the direction of the finger when a fingerprint image is detected.

  The fingerprint recognition apparatus 300 can improve the efficiency of matching by using a plurality of blocks (511 to 514). Referring to FIG. 5, the overlapping area 523 is an area where the block 513 and the partial image 321 overlap each other, and the overlapping area 522 is an area where the block 512 and the partial image 322 overlap. When the input partial image 315 is input, since there are not many regions where the input partial image 315 and the registered partial images (321 to 323) overlap each other, it is effective to divide them into a plurality of blocks. As an example, the fingerprint recognition apparatus 300 may calculate the rotation angle and movement amount of each block with respect to a plurality of registered partial images (321 to 323) based on the frequency information of the image. In this case, the ratio of the area included in the block 513 in common with the block 513 and the registered partial image 321 (hereinafter, the overlapping area) is larger than the ratio of the overlapping area in the input partial image 315. For example, the size of the overlapping regions (522, 523) may be considerably smaller than the overall size of the partial images (321, 322). The ratio of the size of the overlapping region 523 to the size of the block 513 may be larger than the ratio of the size of the overlapping region 523 to the size of the partial image 315. Similarly, the ratio of the size of the overlapping region 522 to the size of the block 512 may be larger than the ratio of the size of the overlapping region 522 to the size of the partial image 315. Therefore, matching can be performed more efficiently.

  In addition, the fingerprint recognition apparatus 300 can perform a strong operation against deterioration included in the input partial image 315 or the registered partial images (321 to 323) by using a plurality of blocks (511 to 514). For example, the fingerprint recognition apparatus 300 may use only the blocks that are well matched with the registered partial images (321 to 323) among the plurality of blocks (511 to 514). The fingerprint recognition apparatus 300 can perform fingerprint recognition that is resistant to deterioration by eliminating the comparison result using the block including deterioration and using only the comparison result using the block including no deterioration.

  In the registration stage, the fingerprint recognition apparatus 300 stores only the registered partial image, and additional information such as information obtained by stitching the registered partial image and alignment information between the registered partial images may not be stored. Therefore, the embodiment of the present invention can provide a technique for efficiently using a memory with low calculation complexity when registering a registered partial image.

  The fingerprint recognition apparatus 300 can align the plurality of blocks (511 to 514) and the registered partial images (321 to 323) by various methods. As an example, the fingerprint recognition device 300 is based on a frequency-based matching method, and includes movement information, rotation information, scale information, or a plurality of blocks (511 to 514) and registered partial images (321 to 323). Various combinations of these may be determined. The matching method based on the frequency may be a method of performing matching in the frequency domain.

  The movement information between the specific block and the specific registered partial image includes an x-axis movement parameter Tx and a y-axis movement parameter Ty. Further, the rotation information between the specific block and the specific registered partial image includes the rotation parameter R, and the scale information includes the scale parameter S. Hereinafter, Tx and Ty are referred to as movement amounts, and R is referred to as a rotation angle. The fingerprint recognition apparatus 300 can calculate the rotation angle, the movement amount, and the scale parameter by comparing the plurality of blocks (511 to 514) and the registered partial images (321 to 323) in the frequency domain. A specific method for calculating the rotation angle, the movement amount, and the scale parameter in the frequency domain will be described later with reference to FIG.

  The fingerprint recognition apparatus 300 can move and rotate the block based on the movement information from which the alignment result is derived. The fingerprint recognition apparatus 300 can enlarge or reduce the block based on the scale information from which the matching result is derived. Since the movement information, rotation information, and scale information are relative between the block and the registered partial image, the fingerprint recognition device 300 moves and rotates the registered partial image instead of moving, rotating, or scaling the block. Or can be scaled.

  If the block and the registered partial image overlap each other due to movement, rotation, and scale, the fingerprint recognition apparatus 300 can calculate a matching score in the overlapping region. For example, the fingerprint recognition apparatus 300 may calculate the matching score based on a normalized correlation method based on the luminance value of the image. As can be seen in FIG. 5, the fingerprint recognition device 300 can accurately match an input fingerprint image in which rotation exists, and can accurately print a fingerprint even if the input fingerprint image is input at an arbitrary angle. Can be recognized.

  FIG. 6 is an operational flowchart illustrating a fingerprint recognition method according to an embodiment of the present invention. Referring to FIG. 6, a fingerprint recognition method according to an embodiment of the present invention includes a step of receiving an input partial image (S610), a step of dividing the input partial image into blocks (S620), a block and a registration part. Comparing with an image (S630) and recognizing a user's fingerprint (S640).

  Step S610 for receiving the input partial image includes a step of performing a preprocessing process. The pre-processing process may include a series of operations for improving the quality of the fingerprint image. The fingerprint image includes an input partial image or a registered partial image. For example, the preprocessing process includes an operation for removing noise contained in a fingerprint image, an operation for increasing the contrast of the fingerprint image, a blur removal (deblurring) operation for removing blur contained in the fingerprint image, and a fingerprint. A warping operation for correcting distortion included in the image may be included.

  The preprocessing process may include an operation for evaluating the quality of the fingerprint image. If the quality of the fingerprint image is less than the threshold quality, the preprocessing process may include an operation of discarding the acquired fingerprint image and inputting a new fingerprint image from the user. According to an embodiment of the present invention, the threshold quality may be selected according to user preference or may be selected in the manufacturing process of the fingerprint recognition device 300.

  The items described above with reference to FIGS. 1 to 5 may be applied to the steps shown in FIG. 6 as they are. Hereinafter, step S630 and step S640 will be described in detail with reference to FIGS.

  FIG. 7 is an operational flowchart illustrating a user authentication method according to an embodiment of the present invention. Referring to FIG. 7, the fingerprint recognition apparatus divides the input partial image into N blocks in step S620. Here, N is a positive integer of 2 or more. In step S710, the fingerprint recognition apparatus acquires L registered partial images from the database 715. Here, L is a positive integer of 1 or more. In step S710, the fingerprint recognition apparatus matches the currently processed block with the L registered partial images out of the N blocks. As an example, the fingerprint recognition apparatus may match a block and a registered partial image using a matching method based on frequency.

  FIG. 8 is an operation flowchart illustrating a frequency-based matching method according to an embodiment of the present invention. Referring to FIG. 8, in step S1011, the time domain information included in the block is converted into frequency domain information through fast Fourier transform (FFT). In step S1012, fast Fourier transform is applied to the registered partial image. The frequency domain information may be based on an orthogonal coordinate system that expresses information using (x, y) coordinates.

  In step S1021, the coordinate system of the frequency domain information included in the block is converted to a polar coordinate system through Log-Polar conversion. As an example, the Log-Polar transform may be performed on the magnitude value of the pixel in the FFT image derived through the fast Fourier transform. The polar coordinate system may represent information using a radius, an angle, or a combination of a radius and an angle. In step S1022, Log-Polar conversion is applied to frequency domain information included in the registered partial image.

  FIG. 9 is a diagram illustrating Log-Polar conversion according to an embodiment of the present invention. Referring to FIG. 9, concentric circles are set based on the center region 1110 in the orthogonal coordinate system. The concentric circles may be divided into a plurality of regions by a radius, an angle, or a combination of a radius and an angle. As an example, the Log-Polar transformation may map a plurality of regions on the orthogonal coordinate system to regions on the polar coordinate system (radius, angle). In this case, the central region 1110 of the orthogonal coordinate system can be mapped to the (0, 0 °) region 1115 of the polar coordinate system. Further, the region 1120, the region 1130, the region 1140, and the region 1150 of the orthogonal coordinate system are mapped to the region 1125, the region 1135, the region 1145, and the region 1155 of the polar coordinate system, respectively.

  Although not shown in the figure, the Log-Polar transformation may map a plurality of regions on the orthogonal coordinate system to regions on the (radius) polar coordinate system. In this case, the region 1120 of the orthogonal coordinate system is mapped to the (0 °) region of the polar coordinate system, the region 1130 and the region 1140 of the orthogonal coordinate system are mapped to the (36 °) region of the polar coordinate system, and Region 1150 can be mapped to a (324 °) region of the polar coordinate system.

  Referring to FIG. 8 again, in step S1031, fast Fourier transform is applied to the image of the block's Log-Polar transform. In step S1032, fast Fourier transform is applied to the Log-Polar transform image of the registered partial image. In step S1040, phase correlation is performed. As a result of the phase correlation, a peak is detected, and the position of the detected peak can indicate rotation information (θ) between the block and the registered partial image.

  According to another embodiment of the present invention, the position of the detected peak may indicate scale information between the block and the registered partial image. For example, one axis of the Log-Polar conversion image corresponds to an angle, and the other axis corresponds to a radius. In this case, the position of the peak detected by the phase correlation may be expressed as (axis coordinate corresponding to an angle, axis coordinate corresponding to a radius). The coordinate of the axis corresponding to the angle may indicate rotation information, and the coordinate of the axis corresponding to the radius may indicate scale information.

  Usually, since the fingerprint image has substantially no scale change, the radius is fixed to a predetermined value (for example, 1). In this case, the position of the peak detected by the phase correlation may be expressed as the coordinate of the axis corresponding to the angle. The coordinates of the axis corresponding to the angle may indicate rotation information.

  In step S1050, the block is rotated based on the rotation information (θ). In step S1060, a fast Fourier transform is applied to the rotated block, and in step S1070 phase correlation is performed. As a result of performing the phase correlation, the position of the detected peak can indicate movement information (Tx, Ty) between the block and the registered partial image. In step S1080, the rotated block moves based on the movement information (Tx, Ty).

  Referring to FIG. 7 again, in step S710, the fingerprint recognition apparatus calculates L block scores indicating the degree to which the block currently being processed is matched with L registered partial images based on the matching result. As described above in FIG. 8, the fingerprint recognition device is currently processing so that the block currently being processed and the first registered partial image overlap based on the {movement information, rotation information} pair for the first registered partial image. Can be rotated and moved.

  The fingerprint recognition device can calculate a block score in the overlapping region. The fingerprint recognition device may calculate the block score by various methods. For example, the fingerprint recognition device may calculate the block score based on a normalized correlation method based on the luminance value of the image. For example, the fingerprint recognition apparatus can calculate a block score using Equation (1).

Here, W is shown a region overlapping between the images I ₁ and the image I _2. Image I ₁ is a block that is rotated, the image I ₂ is a registered partial image. i indicates the x-axis coordinate of the pixel in the overlapping region, and j indicates the y-axis coordinate of the pixel in the overlapping region. x indicates movement information (Tx) in the x-axis direction, and y indicates movement information (Ty) in the y-axis direction. I ₁ (i, j) indicates a pixel value (for example, a luminance value of the pixel) in the (i, j) coordinate of the image I ₁ , and I ₂ (x + i, y + j) indicates the (x + i, y + j) coordinate of the image I ₂ Indicates the pixel value (for example, the luminance value of the pixel).

  FIG. 10 is a diagram illustrating an operation for processing a block of a registered image and an input image in order to calculate a block score according to an embodiment of the present invention. Referring to FIG. 10, a registered image 1210 is converted into a first LPT image 1220 through a fast Fourier transform and a log-polar transform. The input image block 1215 is transformed into a second LPT image 1225 via a fast Fourier transform and a Log-Polar transform.

  The rotation information (θ) between the registered image 1210 and the block 1215 is determined by the phase correlation 1230 between the first LPT image 1220 and the second LPT image 1225. The block 1215 is rotated according to the rotation information (θ). Movement information (Tx, Ty) between the registered image 1210 and the block 1215 is determined by a phase correlation 1250 between the FFT image of the registered image 1210 and the rotated FFT image of the block 1245.

  Based on the rotation information (θ) and the movement information (Tx, Ty), the registered image 1210 and the block 1215 are aligned. A score 1270 is calculated for the region where the registered image 1210 and the rotated block 1245 overlap in the aligned image 1260. The score 1270 may be called a block score, a matching score, or the like.

  Referring to FIG. 7 again, the fingerprint recognition apparatus can also calculate block scores for the second registered partial image to the Lth registered partial image in the same manner. Hereinafter, the block score may be called a matching score, a matching score, or the like.

  In step S720, the fingerprint recognition apparatus checks whether the block just processed is the last block among the N blocks. If the block just processed is not the last block among the N blocks, the fingerprint recognition apparatus performs step S710 on the block that has not been processed yet. The fingerprint recognition apparatus can calculate N × L block scores by repeating step S720 and step S730 N times.

  In step S730, the fingerprint recognition apparatus selects the top K block scores from the N × L block scores. As described above with reference to FIG. 4, the input partial image 315 and the registered partial image 323 partially overlap each other. In this case, among the blocks of the input partial image 315, a block located in an overlapping area derives a meaningful block score, whereas a block located in a non-overlapping area derives a meaningless block score. The fingerprint recognition apparatus can eliminate the meaningless block score by selecting the top K block scores from the N × L block scores. K may be determined in the range of 1 to N × L.

  The fingerprint recognition device can calculate the feature value Val based on the top K block scores. For example, the fingerprint recognition apparatus can calculate the sum of the top K block scores using Equation (2).

Here, ↓ Socre (B _i ) is the i-th block score arranged in descending order. The method of calculating the feature value Val may be variously modified such as obtaining the average of the top K block scores. Further, the weight value of each block score for calculating the feature value may be applied differently depending on the calculated optimum rotation angle (for example, the highest block score).

  In step S740, the fingerprint recognition device determines whether the user can be authenticated by comparing the feature value Val with a threshold value. If the feature value Val is larger than the threshold value, the fingerprint recognition apparatus determines that the authentication is successful in step S750. If the feature value Val is equal to or smaller than the threshold value, the fingerprint recognition device determines that the authentication has failed in step S760. Here, the threshold value may be determined by a method for calculating the feature value Val, the number of registered partial images, the quality of the registered partial images, the quality of the input partial images, and the like.

  FIG. 11 is an operation flowchart illustrating a user authentication method according to another embodiment of the present invention.

  The fingerprint recognition device can calculate the rotation angle for each block of the registered partial image, and can calculate the optimum rotation angle for the registered partial image. The fingerprint recognition device applies the optimal rotation angle corresponding to each registered partial image to all the blocks, and then derives the matching score by calculating the amount of movement between the rotated block and the registered partial image. be able to.

  The operation of the fingerprint recognition device can be divided into two steps. For example, in the first step, one registered partial image and N blocks of the input partial image are matched (for example, using the matching method shown in FIG. 8), and N matched scores are acquired. Can do. The fingerprint recognition device may apply the rotation angle of a block having a high matching score (for example, having the highest similarity score) out of the N matching scores acquired in the matching result to all the blocks. Good. In the second step, the fingerprint recognition apparatus can calculate the movement amount for each block.

  In addition, as will be described in detail below, the fingerprint recognition apparatus selects the top M registered partial images based on the matching score in the first step, and only the selected registered partial images in the second step. It is possible to effectively improve the speed performed by calculating the movement amount using.

  Referring to FIG. 11, the fingerprint recognition apparatus acquires L registered partial images from the database 815 in step S810. In step S810, the fingerprint recognition apparatus matches the N blocks and the L registered partial images. The alignment operation in step S810 is performed for each of the L registered partial images. For example, for the first registered partial image among the L registered partial images, the fingerprint recognition apparatus aligns the N blocks of the first registered partial image and the input partial image (as shown in FIG. 8 as an example). N primary block scores for the first registered image can be obtained according to the matching result. In step S810, the same operation may be performed on the remaining registered partial images among the L registered images. As a result, the fingerprint recognition apparatus can calculate N × L primary block scores based on the matching result. In step S810, the matters described above in steps S710 and S720 of FIG. 7 are applied as they are.

  In step S820, the fingerprint recognition apparatus determines an optimum rotation angle R for each registered partial image. The optimum rotation angle for each registered partial image is an angle for rotating the block corresponding to the registered partial image. For example, the fingerprint recognition device can select the highest block score among N primary block scores related to the first registered partial image. The fingerprint recognition device can extract the rotation angle from the matching information for which the selected highest block score is calculated. For example, the fingerprint recognition apparatus in step S820 can extract the rotation angle R with respect to the first registered partial image. The extracted rotation angle R may be the rotation angle of the block having the highest block score among the matching performed on the first registered image in step S810.

  The matching information may be a {rotation angle, movement amount} pair. The fingerprint recognition apparatus may determine the extracted rotation angle R as an optimal rotation angle for rotating all N blocks related to the first registered partial image. The fingerprint recognition apparatus in step S820 may determine an optimum rotation angle for rotating the block in the same manner corresponding to each of the remaining registered partial images among the L registered partial images. Accordingly, the fingerprint recognition apparatus in step S820 can determine L optimal rotation angles for L registered images.

  In step S830, the fingerprint recognition apparatus rotates N blocks based on the optimum rotation angle R for each registered partial image. For example, the fingerprint recognition apparatus can rotate all N blocks by an optimum rotation angle with respect to the first registered partial image in correspondence with the first registered partial image. Also, the fingerprint recognition apparatus can rotate all N blocks by an optimum rotation angle with respect to the registered partial image in correspondence with the other registered partial images.

  In step S850, the fingerprint recognition apparatus compares the rotated block corresponding to the registered partial image with the registered partial image. For example, the fingerprint recognition device can realign the registered partial image with the rotated block corresponding to each of the registered partial images based on a frequency-based matching method. According to one embodiment of the present invention, the realignment in step S850 includes the steps (S1060 to S1080) of FIG. 8 and may not include the steps (S1011 to S1050).

  Here, the fingerprint recognition apparatus may determine the movement amount of the block while maintaining the rotation angle of the block. The fingerprint recognition device may calculate a secondary block score based on the realignment result. According to an embodiment of the present invention, the realignment in step S850 may be performed on each of the L registered partial images.

  According to an embodiment of the present invention, the steps (S820, S830, S850) may be performed as described above for all L registered images. However, as will be described in more detail below, according to one embodiment of the present invention, in addition, a ranking operation is performed in step S840, and steps (S820, S830, S850) include L registration portions. It may be performed smaller than the image.

  For example, the fingerprint recognition apparatus may not use all L registered partial images when calculating the secondary block score. In step S840, the fingerprint recognition apparatus aligns the L registered partial images based on the primary block score. For example, the fingerprint recognition device may align the registered partial images such that the ranking of the registered partial images associated with a high primary block score increases. For example, the fingerprint recognition apparatus may rank the L registered partial images in descending order based on the primary block score determined in step S810. For example, the fingerprint recognition apparatus may rank the L registered partial images in descending order based on the block score average of the L registered partial images. The block score average may be an average of the N block scores determined in step S810. As another example, the fingerprint recognition apparatus may rank the L registered partial images in descending order based on the maximum block score of the L registered partial images. The maximum block score may be the maximum score among the N block scores determined in step S810.

  The fingerprint recognition device can select M registered partial images according to the arranged order. For example, the fingerprint recognition apparatus may select M maximum number of registered partial images from among the L registered partial images. Here, M is a positive integer smaller than L. By comparing only the M registered partial images of the L registered partial images with the input partial image, the processing speed of fingerprint recognition can be improved.

  In this case, in step S820, the fingerprint recognition apparatus calculates an optimum rotation angle for the M registered partial images. In step S830, the fingerprint recognition apparatus rotates the block based on the optimum rotation angle for each of the M registered partial images. In step S850, the fingerprint recognition apparatus calculates N × M secondary block scores.

  In step S860, the fingerprint recognition apparatus selects the top K block scores from the N × M block scores. The fingerprint recognition device may calculate the feature value Val based on the top K block scores. In step S870, the fingerprint recognition apparatus determines whether or not the user can be authenticated by comparing the feature value Val with a threshold value. If the feature value Val is greater than the threshold value, the fingerprint recognition device determines that the authentication has succeeded in step S880, and if the feature value Val is equal to or less than the threshold value, determines that the authentication has failed in step S890.

  Although the method for comparing with a plurality of registered images has been described above, the same authentication method may be applied even when there is one registered image (L = 1). In this case, K can be determined in the range of 1 to N. Further, although the case where the input and registered images are partial fingerprint images has been described, the described authentication method may be applied even when the input and registered images are all fingerprint images.

  FIG. 12 is a block diagram illustrating an electronic system according to an embodiment of the present invention. Referring to FIG. 12, the electronic system includes a sensor 920, a processor 910, and a memory 930. Sensor 920, processor 910, and memory 930 communicate via bus 940.

  The sensor 920 may be the fingerprint sensor 310 shown in FIG. The sensor 920 can capture a fingerprint image by a known method (for example, a method of converting an optical image into the signal). The image is output from the processor 910.

  The processor 910 may include at least one device as described above in FIGS. 1-11, or may perform at least one method as described above in FIGS. For example, the processor 910 may include the fingerprint recognition device 300 of FIG. The memory 930 stores a registered partial image captured and registered by the sensor 920, an input partial image captured by the sensor 920, a matching result processed by the processor 910, and / or a block score calculated by the processor 910, and the like. To do. Memory 930 may be volatile memory or non-volatile memory.

  The processor 910 executes a program and controls the electronic system. Program code executed by processor 910 may be stored in memory 930. The electronic system may be connected to an external device (for example, a personal computer or a network) via an input / output device (not shown) to exchange data.

  Electronic systems include mobile devices such as mobile phones, smartphones, PDAs, tablet PCs, laptop computers, computer devices such as personal computers, tablet PCs and netbooks, or televisions, smart TVs, security devices for gate control, etc. Various electronic systems, such as electronic products, may be included.

  In the above, the case where the user is recognized using a part of the user's fingerprint has been described, but the embodiment of the present invention may be extended to the case where the user is recognized using a part of the user's biometric data. Good. Here, the biometric data may include information about the user's fingerprint, information about veins, information about irises, and the like. In this case, the processor 910 receives input partial data corresponding to a part of the user's biometric data from the sensor 920, divides the input partial data into blocks, and the partial data of the biometric data already registered with the block. The registered partial data corresponding to the above can be compared, and the user can be recognized based on the comparison result.

  As an example, the sensor 920 may include a sensor that recognizes a user's vein pattern. The sensor 920 can extract a vein pattern from the back of the user's hand. The sensor 920 can acquire an image including a vein pattern after using infrared illumination and a filter to maximize the luminance of the blood vessel relative to the skin. Here, the sensor 920 may acquire a partial image corresponding to a part of the vein pattern. In this case, the processor 910 can recognize the user by comparing a partial image corresponding to a part of the vein pattern with a partial image of the already registered vein pattern.

  As another example, the sensor 920 may include a sensor that recognizes a user's iris pattern. Sensor 920 can scan or capture an iris pattern between the user's pupil and sclera (white eye region). Here, the sensor 920 may acquire a partial image corresponding to a part of the iris pattern. In this case, the processor 910 can recognize the user by comparing a partial image corresponding to a part of the iris pattern with a partial image of the already registered iris pattern.

  The above-described embodiments are implemented by hardware components, software components, or a combination of hardware components and software components. For example, the apparatus and components described in the present embodiment include, for example, a processor, a controller, an ALU (arithmetic logic unit), a digital signal processor (digital signal processor), a microcomputer, an FPA (field programmable array), and a PLU (programmable logarithm). It may be implemented using one or more general purpose or special purpose computers, such as units, microprocessors, or different devices that execute and respond to instructions. The processing device executes an operating system (OS) and one or more software applications running on the operating system. The processing device also accesses, stores, manipulates, processes, and generates data in response to software execution. For convenience of understanding, one processing device may be described as being used, but those having ordinary knowledge in the art may recognize that the processing device has multiple processing elements and / or multiple types of processing. It can be seen that it contains elements. For example, the processing device includes a plurality of processors or one processor and one controller. Other processing configurations such as parallel processors are also possible.

  The software includes computer programs, code, instructions, or a combination of one or more of these, configures the processing device to operate as desired, and instructs the processing device independently or in combination. Software and / or data is interpreted by the processing device and any type of machine, component, physical device, virtual device, computer storage medium or device for providing instructions or data to the processing device, or transmitted signal It can be realized permanently or temporarily via waves. The software is distributed on computer systems connected to the network and stored or executed in a distributed manner. Software and data are stored on one or more computer-readable recording media.

   The method according to the present embodiment is embodied in the form of program instructions executed through various computer means, and is recorded on a computer-readable recording medium. The recording medium includes program instructions, data files, data structures, etc. alone or in combination. The recording medium and program instructions may be specially designed and configured for the purposes of the present invention, and may be known and usable by those skilled in the art of computer software technology. Good. Examples of computer-readable recording media include magnetic media such as hard disks, floppy (registered trademark) disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magnetic-lights such as floppy disks. The medium and hardware devices specially configured to store and execute program instructions such as ROM, RAM, flash memory, etc. may be included. Examples of program instructions include not only machine language code generated by a compiler but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

  As described above, the embodiments have been described with reference to the limited drawings. However, various technical modifications and variations can be applied to those having ordinary knowledge in the art based on the above description. . By way of example, the described techniques may be performed in a different order than the described methods and / or components of the described system, structure, apparatus, circuit, etc. may be combined or combined in a different form than the described methods. Appropriate results may be achieved even if replaced or replaced by other components or equivalents. Accordingly, the scope of the present invention is not limited to the disclosed embodiments, but is defined by the claims and the equivalents of the claims.

100, 200: Fingerprints 110-170, 210-295, 315, 321-323: Partial image 300: Fingerprint recognition device 310: Fingerprint sensors 320, 715, 815: Databases 511-514, 1215: Block 1110: Central region 1120, 1130, 1140, 1150: area 1210: registered image 1220: first LPT image 1225: second LPT image 1230, 1250: phase correlation 1245: rotated block 1260: aligned image 1270: score 910: processor 920: sensor 930: Memory 940: Bus

Claims (28)

Receiving an input image corresponding to the user's biometric information;
Dividing the input image into blocks;
Comparing the one or more blocks with one or more registered partial images corresponding to already registered biometric information;
Recognizing, identifying, and / or authenticating the user based on the result of the comparison;
A biological image recognition method comprising:   The biological image recognition method according to claim 1, wherein the receiving step includes a step of detecting a partial region of the user's biological information via a detection region smaller than the user's biological information in the input image.   The method further comprises the step of generating the registered partial image by repeatedly detecting a partial area of the biometric information of the already registered user through a detection area smaller than the biometric information of the already registered user. 3. The biological image recognition method according to 1 or 2.   The biological image recognition method according to any one of claims 1 to 3, wherein the dividing step includes a step of dividing the input image into the blocks based on a predetermined pattern.   The biological image recognition method according to claim 1, wherein the comparing includes calculating a score indicating a degree to which each of the blocks is matched with the registered partial image. The comparing step includes:
Aligning the block and the registered partial image;
Comparing the block with the registered partial image based on the matching result;
The biological image recognition method according to any one of claims 1 to 5, further comprising:   The matching step includes determining at least one of movement information, rotation information, and scale information between the block and the registered partial image based on a frequency-based matching method. 7. The biological image recognition method according to 6. The rotation information is a first rotation angle with respect to the registered partial image,
The comparing step compares the block rotated based on the first rotation angle with the registered partial image.
The biological image recognition method according to claim 7, comprising: The first rotation angle is determined based on a score based on a second rotation angle between one or a plurality of blocks different from the block and the registered partial image based on a matching method based on frequency. ,
The biological image recognition method according to claim 8. Recognizing, identifying and / or authenticating comprises:
Selecting a predetermined number of {block, registered partial image} pairs based on the result of the comparison;
Recognizing, identifying, and / or authenticating the user based on the selected pair;
The biological image recognition method according to any one of claims 1 to 9, further comprising:   The biological body according to claim 10, wherein the selecting step includes a step of selecting the predetermined number of pairs based on a score between a block included in each of a plurality of pairs and a registered partial image. Image recognition method. The input image is
The biological image recognition method according to any one of claims 1 to 11, comprising at least one of a fingerprint image, a vein image, and an iris image.   A computer program that, when executed by a processor, causes the processor to execute the method according to any one of claims 1 to 12. A fingerprint sensor that receives an input image corresponding to the user's fingerprint;
Dividing the input image into a plurality of blocks, comparing the plurality of blocks with at least one registered image corresponding to an already registered fingerprint, and at least recognizing the user's fingerprint based on the result of the comparison One processor,
A fingerprint recognition device. The input image corresponds to a partial image of the user's fingerprint;
The fingerprint recognition apparatus according to claim 14, wherein the at least one registered image corresponds to at least one partial image of the already registered fingerprint. The at least one processor comprises:
Aligning the plurality of blocks with the at least one registered image;
The fingerprint recognition device according to claim 14 or 15, wherein a score indicating a degree to which the plurality of blocks are matched with the at least one registered image is calculated based on the matching result. The at least one processor comprises:
The fingerprint of claim 16, wherein at least one of movement information, rotation information, and scale information between the plurality of blocks and the at least one registered image is determined based on a frequency-based matching scheme. Recognition device. The at least one processor comprises:
As the rotation information, a first rotation angle for the at least one registered image is determined,
The fingerprint recognition apparatus according to claim 17, wherein the plurality of blocks rotated based on the determined first rotation angle are compared with the at least one registered image. The at least one processor comprises:
Determining a rotation angle between the plurality of blocks and the at least one based on a frequency-based matching scheme;
The fingerprint recognition device according to claim 18, wherein an optimum rotation angle is determined based on a score based on the rotation angle. The at least one processor comprises:
20. The fingerprint recognition device according to any one of claims 14 to 19, wherein the user is authenticated and / or identified based on a result of the comparison. The at least one processor comprises:
21. The user selects a predetermined number of {block, registered image} pairs based on the comparison result, and recognizes the user's fingerprint based on the selected pair. The fingerprint recognition device described in 1. The at least one processor comprises:
The fingerprint recognition device according to claim 21, wherein the predetermined number of pairs is selected based on a score between a block included in each of a plurality of pairs and a registered image. A fingerprint recognition method,
Receiving an input fingerprint image captured by an image sensor;
Generating a plurality of blocks, wherein the plurality of blocks are image blocks of the input fingerprint image; and
Selecting a first rotation angle for each of one or more registered fingerprint images by performing a first alignment operation based on the registered fingerprint image and the plurality of blocks;
Rotating each of the plurality of blocks by the first rotation angle;
Performing a second alignment operation based on the registered fingerprint image and the plurality of blocks rotated by the first rotation angle;
Determining the presence or absence of authentication of the input fingerprint image based on a second alignment operation performed on each of the one or more registered fingerprint images;
A fingerprint recognition method including: The step of selecting is for a first registered image of the one or more registered fingerprint images,
Determining a plurality of rotation angles corresponding to the plurality of blocks by performing the first alignment operation based on the plurality of blocks and the first registered image;
Rotating the plurality of blocks by the plurality of rotation angles corresponding to the plurality of blocks;
Generating a plurality of first block scores corresponding to the plurality of blocks based on the input fingerprint image and the plurality of blocks rotated by the plurality of rotation angles;
The fingerprint recognition method according to claim 23, further comprising: selecting a first rotation angle among the plurality of rotation angles based on the plurality of first block scores. The first alignment operation is:
Generating information of a first frequency domain based on the first registered image;
Generating, for each block of the plurality of blocks, second frequency domain information based on the block;
Performing a phase correlation operation based on the information of the first frequency domain and the information of the second frequency domain;
Determining the determined rotation angle corresponding to the block among the plurality of determined rotation angles based on the phase correlation operation;
The recognition method of Claim 24 containing these. The step of determining includes
For each registered fingerprint image of the one or more registered fingerprint images,
Generating a plurality of second block scores corresponding to the plurality of blocks based on the registered fingerprint image and the plurality of blocks rotated by the first rotation angle selected for the registered fingerprint image; When the plurality of blocks are rotated by the first rotation angle, the plurality of blocks corresponding to the plurality of second block scores are aligned with the registered fingerprint image. Instructing the degree to which
Determining the presence or absence of authentication of the input fingerprint image based on the plurality of second block scores generated for the one or more registered fingerprint images;
The fingerprint recognition method according to claim 24 or 25, comprising:   The input fingerprint image is a partial image of a fingerprint captured by an image sensor, each of the one or more registered fingerprint images is a partial image of a fingerprint, and each of the one or more registered fingerprint images The fingerprint recognition method according to any one of claims 23 to 26, wherein is stored in a memory.   The fingerprint according to any one of claims 23 to 27, wherein generating the plurality of blocks includes generating the plurality of blocks by dividing the input fingerprint image into the plurality of blocks. Recognition method.

Images