WO2021027364A1 - Procédé et appareil d'authentification d'identité basée sur la reconnaissance de veine de doigt - Google Patents

Procédé et appareil d'authentification d'identité basée sur la reconnaissance de veine de doigt Download PDF

Info

Publication number
WO2021027364A1
WO2021027364A1 PCT/CN2020/093383 CN2020093383W WO2021027364A1 WO 2021027364 A1 WO2021027364 A1 WO 2021027364A1 CN 2020093383 W CN2020093383 W CN 2020093383W WO 2021027364 A1 WO2021027364 A1 WO 2021027364A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
finger vein
finger
target
target user
Prior art date
Application number
PCT/CN2020/093383
Other languages
English (en)
Chinese (zh)
Inventor
巢中迪
庄伯金
王少军
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2021027364A1 publication Critical patent/WO2021027364A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/14—Vascular patterns

Definitions

  • This application relates to the field of artificial intelligence, and in particular to an identity verification method and device based on finger vein recognition.
  • Finger vein recognition technology uses the texture of finger veins for identity verification, which is harmless to the human body and is not easy to be stolen or forged.
  • the identification technology can be widely used in access control systems in banking, finance, government, education and other fields.
  • embodiments of the present application provide an identity verification method and device based on finger vein recognition to solve the problem of low accuracy of identity verification based on finger vein images.
  • the embodiment of the application provides an identity verification method based on finger vein recognition, the method includes: obtaining a finger vein width parameter; constructing a target convolution kernel according to the finger vein width parameter; obtaining a finger vein image of a target user; The target convolution kernel filters the finger vein image of the target user; extracts the ROI area of the finger vein image of the target user after the filtering process to obtain a target area image; extracts finger veins from the target area image Feature vector to obtain the target finger vein feature vector; obtain all the finger vein feature vectors stored in a preset database, where the association relationship between the user and the finger vein feature vector is stored; calculate the target finger veins separately The similarity between the feature vector and all the finger vein feature vectors stored in the preset database to obtain multiple similarities; if at least one of the multiple similarities is greater than or equal to the preset similarity threshold, It is determined that the identity verification of the target user is successful; if the multiple similarities are all less than the preset similarity threshold, it is determined that the identity verification of the target user has failed.
  • An embodiment of the present application provides an identity verification device based on finger vein recognition.
  • the device includes: a first obtaining unit for obtaining a finger vein width parameter; a construction unit for constructing a target volume based on the finger vein width parameter.
  • the second acquisition unit is used to acquire the finger vein image of the target user; the filtering unit is used to filter the finger vein image of the target user using the target convolution kernel; the first extraction unit is used to extract After filtering, the ROI area of the finger vein image of the target user is obtained to obtain the target area image; the second extraction unit is configured to extract the finger vein feature vector from the target area image to obtain the target finger vein feature vector;
  • third The obtaining unit is used to obtain all the finger vein feature vectors stored in a preset database, where the association relationship between the user and the finger vein feature vectors is stored; the calculating unit is used to calculate the target finger veins respectively The similarity between the feature vector and all the finger vein feature vectors stored in the preset database to obtain multiple similarities; the first determining unit is configured to,
  • the target convolution kernel is constructed according to the characteristic parameters of the finger veins, and the target convolution kernel is used to filter the finger vein images, which effectively avoids the interference of epidermal texture and improves the accuracy of finger vein image matching, thereby solving The problem of low accuracy of identity verification based on finger vein images in the prior art is solved, and the effect of improving the accuracy of identity verification based on finger vein images is achieved.
  • Fig. 1 is a flowchart of an optional identity verification method based on finger vein recognition according to an embodiment of the present application
  • FIG. 2 is a schematic diagram of an optional identity verification device based on finger vein recognition according to an embodiment of the present application
  • Fig. 3 is a schematic diagram of a computer device provided by an embodiment of the present application.
  • the embodiment of the application provides an identity verification method based on finger vein recognition. As shown in FIG. 1, the method includes:
  • Step S100 Obtain a finger vein width parameter.
  • Step S101 construct a target convolution kernel according to the finger vein width parameter.
  • Step S102 acquiring a finger vein image of the target user.
  • Step S103 using the target convolution kernel to perform filtering processing on the finger vein image of the target user.
  • Step S104 Extract the ROI area of the finger vein image of the target user after the filtering process to obtain the target area image.
  • Step S105 Extract the finger vein feature vector from the target area image to obtain the target finger vein feature vector.
  • Step S106 Obtain all the finger vein feature vectors stored in the preset database, and the preset database stores the association relationship between the user and the finger vein feature vectors.
  • Step S107 Calculate the similarities between the target finger vein feature vector and all the finger vein feature vectors stored in the preset database, respectively, to obtain multiple similarities.
  • Step S108 if at least one of the multiple similarities is greater than or equal to the preset similarity threshold, it is determined that the identity verification of the target user is successful.
  • Step S109 If the multiple similarities are all less than the preset similarity threshold, it is determined that the identity verification of the target user has failed.
  • ROI region of interest
  • the region of interest the region of interest.
  • the area to be processed is outlined in the form of boxes, circles, ellipses, and irregular polygons from the processed image, which is called the region of interest.
  • Finger veins refer to the veins in the fingers of the human body. Finger vein recognition uses the characteristics of the vascular structure to achieve identity authentication. Under visible light, finger veins are invisible and can only be obtained with a special collection device. It is medically proven that the vascular structure of human finger veins can penetrate bones and muscles under the irradiation of near-infrared light, and the hemoglobin flowing through the venous blood vessels can easily absorb the infrared light of this band to highlight the vein structure. Finger vein images can be captured by a special image acquisition device such as an infrared CCD camera, and then the finger vein images can be analyzed and processed to obtain the finger vein characteristics. The finger vein structure of different people is different, even the finger veins of twins are different, and the finger vein structure of adults does not change, that is, the finger vein is unique, which provides a scientific basis for finger vein recognition.
  • a special image acquisition device such as an infrared CCD camera
  • A represents a finger vein feature vector
  • B represents a finger vein feature vector
  • a i represents each component of the finger vein feature vector A
  • B i represents a finger vein feature vector
  • Each component of B, n represents the number of components contained in finger vein feature vector A and finger vein feature vector B (the number of components contained in finger vein feature vector A and finger vein feature vector B is equal)
  • Simi(A,B) represents The similarity between the finger vein feature vector A and the finger vein feature vector B.
  • the finger vein image in the embodiment of the present application may refer to a vein near infrared image.
  • Finger vein images can be captured by near-infrared CCD or CMOS image acquisition modules.
  • Convolution kernel In image processing, given an input image, each pixel in the output image is a weighted average of pixels in a small area in the input image, where the weight is defined by a function, which is called the convolution kernel.
  • One property of the convolution kernel is locality. That is, it only focuses on local features, and the degree of locality depends on the size of the convolution kernel.
  • the finger vein width parameter is a parameter used to indicate the width range of the finger vein. Determine the size of the target convolution kernel according to the finger vein width parameter. Because the convolution kernel has the property of locality, the degree of locality depends on the size of the convolution kernel. Therefore, if there are more pixels that need attention in the finger vein image, the size of the convolution kernel can be larger; if the finger vein image If there are fewer pixels to be concerned in, the size of the convolution kernel can be smaller. For example, assuming that the vein width parameter indicates that the width of the finger vein covers m pixels, the size of the target convolution kernel can be m ⁇ m.
  • the target convolution kernel is used to filter the finger vein image, that is, the target convolution kernel and the finger vein image are convolved.
  • the target convolution kernel is used to perform a series of operations on each pixel in the finger vein image.
  • the target convolution kernel is an m ⁇ m matrix containing m ⁇ m elements, and each element has a preset weight value.
  • the new pixel value is used to replace the original pixel value, and the filtered finger vein image is obtained.
  • the characteristics of the finger vein image can be extracted, and the enhancement effect of the finger vein image can be realized, effectively avoiding the interference of the epidermal texture.
  • the following takes a 3 ⁇ 3 target convolution kernel as an example to describe in detail the process of convolution operation between the target convolution kernel and the finger vein image.
  • the pixel values of all pixels in the finger vein image form a matrix, each element of the matrix corresponds to a pixel, and the value of the element is the pixel value of the pixel.
  • the matrix formed by the pixel values of all pixels in the finger vein image is a matrix M1.
  • the matrix M2 is obtained.
  • E2(P,Q) E1(P-1,Q-1) ⁇ W(1,1)+E1(P-1,Q) ⁇ W(1,2)+E1(P-1,Q+1) ) ⁇ W(1,3)+E1(P,Q-1) ⁇ W(2,1)+E1(P,Q) ⁇ W(2,2)+E1(P,Q+1) ⁇ W( 2,3)+E1(P+1,Q-1) ⁇ W(3,1)+E1(P+1,Q) ⁇ W(3,2)+E1(P+1,Q+1) ⁇ W(3,3).
  • E1 (P-1, Q-1) represents the value of the element in the P-1 row and Q-1 column of the matrix M1;
  • E1 (P-1, Q) represents the P-1 row, The value of the element in the Qth column;
  • E1(P-1, Q+1) represents the value of the element in the P-1th row and the Q+1th column of the matrix M1, and the rest is the same.
  • W(1,1) represents the weight value of the element in the first row and the first column of the target convolution kernel;
  • W(1,2) represents the weight value of the element in the first row and second column of the target convolution kernel;
  • W(1,3) represents the weight value of the element in the first row and the third column of the target convolution kernel, and the rest is the same.
  • E2(P, Q) represents the value of the element in the Pth row and Qth column of the matrix M2.
  • E2(P,Q) E1(P-1,Q-1) ⁇ 10+E1(P-1,Q) ⁇ 10+E1(P-1,Q+1) ⁇ 10+E1(P,Q- 1) ⁇ 10+E1(P,Q) ⁇ 10+E1(P,Q+1) ⁇ 10+E1(P+1,Q-1) ⁇ 10+E1(P+1,Q) ⁇ 10+E1 (P+1, Q+1) ⁇ 10
  • the sum of the values of the adjacent nine elements is 400, then after the convolution operation is completed, the corresponding position of matrix M2
  • the value of the element of is 4000; assuming a certain element in the matrix M1 (assuming element B) as the center, the sum of the values of the adjacent nine elements is 70, then after the convolution operation is completed, the element at the corresponding position of the matrix M2 The value is 700.
  • the convolution operation can strengthen the key area (for example, the area where the finger vein is located is the key area in the embodiment of this application), and increase the pixel difference between the finger vein and the epidermis. Thereby reducing the interference of epidermal texture.
  • the target convolution kernel is constructed according to the characteristic parameters of the finger veins, and the target convolution kernel is used to filter the finger vein images, which effectively avoids the interference of epidermal texture and improves the accuracy of finger vein image matching, thereby solving The problem of low accuracy of identity verification based on finger vein images in the prior art is solved, and the effect of improving the accuracy of identity verification based on finger vein images is achieved.
  • the embodiment of the present application provides three methods for extracting the ROI area of the finger vein image of the target user to obtain the target area image.
  • the first method perform super-pixel segmentation on the finger vein image of the target user to obtain the first image; perform Sobel operator edge detection on the finger vein image of the target user to obtain the second image; perform the first image and the second image Compare to obtain a plurality of first pixels, where the first pixel is the pixel that overlaps the first image and the second image; the second pixel is determined according to the first pixel, and the second pixel is a plurality of first pixels One of the points; perform edge tracking on the first image according to the second pixel point to obtain a third image, which is a complete finger edge image; perform angle correction, height cropping, width cropping and normalization on the third image , Get the target area image.
  • image segmentation refers to the process of subdividing a digital image into multiple image sub-regions (collections of pixels) (also called superpixels).
  • a super pixel is a small area composed of a series of adjacent pixels with similar characteristics such as color, brightness, and texture. Most of these small areas retain effective information for further image segmentation, and generally do not destroy the boundary information of objects in the image.
  • the result of image segmentation is a collection of sub-regions on the image (the whole of these sub-regions covers the entire image), or a collection of contour lines extracted from the image (for example, edge detection).
  • Each pixel in a sub-region is similar under a certain characteristic measurement, such as color, brightness, and texture. Adjacent regions are very different under a certain characteristic measurement.
  • the process of determining the second pixel point according to the first pixel point is: randomly selecting one of the overlapping points of one-third to one-half the length of the finger in the first image and the second image as the second pixel point.
  • the edge tracking of the first image according to the second pixel points includes: in the first image, the second pixel points of the upper and lower finger edges are respectively tracked to the fingertip and finger root direction of the finger.
  • Perform angle correction, height cropping, and width cropping on the third image including: performing angle correction on the third image according to the midline of the finger to obtain a third image after angle correction; and performing angle correction on the third image according to the projection value of the finger joints Perform height cropping; perform width cropping on the third image after height cropping according to the inner tangent line of the contour of the finger edge.
  • the third image must be normalized, including size normalization and gray-scale normalization.
  • size normalization are: 1. For different fingers, different sizes have no effect on the matching results of veins, that is, it will not cause misunderstanding; but if it is the same finger, if the size is different, it is easy to cause misunderstanding , That is, I cannot recognize my own situation. 2.
  • Image size normalization is essentially a geometric transformation of an image, which is generally achieved by mapping from the opposite direction of the target image. Reverse mapping is to scan each pixel of the target image and determine the original pixel corresponding to the target pixel according to a given transformation formula. Using this method to calculate the target image can ensure that the entire target image has no empty pixels, that is, each pixel of the obtained target image has a corresponding gray value. Gray normalization is mainly to increase the brightness of the image, make the details of the image clearer, and reduce the influence of light and light intensity.
  • the complete finger edge is tracked through the use of ultra-high pixel segmentation and Sobel operator edge detection, and then the finger vein map with blurred background is obtained, and then angle correction, height cropping, and width cropping are performed. Due to the characteristics of ultra-high pixels This method not only can segment finger vein images correctly, but also enhances the robustness of the image against background noise.
  • the second method extract a complete finger image from the finger vein image of the target user; extract the largest rectangular area from the complete finger image according to the boundary information of the finger contour; perform gradient operations on the largest rectangular area to generate a gradient image;
  • the gradient image locates the joint position; the target area image is determined according to the joint position.
  • the specific steps are: use the maximum between-class variance method to binarize the finger vein image to obtain a binarized image; use the Sobel edge detection operator to extract the finger contour image ; Subtract the corresponding pixels of the binarized image and the finger contour image to obtain the difference image; extract the maximum connected area of the difference image; multiply the maximum connected area with the finger image to remove the invalid information around the finger, and finally extract the complete Finger image.
  • the maximum between-class variance method is an adaptive threshold determination method, abbreviated as OTSU, which is a global binarization algorithm. It divides the image into two parts, the foreground and the background, according to the gray characteristics of the image. When the optimal threshold is taken, the difference between the two parts should be the largest.
  • the standard used in the OTSU algorithm to measure the difference is the more common maximum between-class variance. If the inter-class variance between the foreground and the background is larger, it means that the difference between the two parts of the image is greater. When part of the target is mistakenly classified as the background or part of the background is mistakenly classified as the target, it will cause the difference between the two parts. Decrease, when the threshold segmentation makes the variance between classes the largest, it means that the probability of misclassification is the smallest.
  • Positioning the joint position according to the gradient image includes: performing row pixel summation operation on the gradient image to obtain the gray distribution curve of the row pixel sum of the gradient image; taking the position of the crest point of the gray distribution curve as the joint position.
  • the gradient value will form a local maximum near the interphalangeal joint.
  • the finger can be located by locating the local maximum. Joint position.
  • the gray distribution curve of the sum of the pixels of the gradient image can be obtained.
  • the gray distribution curve of the pixel sum of the gradient map presents a peak state at the interphalangeal joint position, and the joint position of the finger vein image can be located by locating the peak position of the curve.
  • the third method filter the finger vein infrared sample image to filter out image noise; use image binarization to extract the finger mask from the finger vein infrared image sample after the image noise is filtered out; reverse the finger vein sample image Mask processing to obtain a region of interest (ROI).
  • ROI region of interest
  • the finger vein feature vector is extracted from the image of the target area to obtain the target finger vein feature vector, which can be realized by a convolutional neural network.
  • the convolutional neural network needs to be trained.
  • the training process is as follows: annotate the finger vein sample images as the training set, distinguish background pixels and finger vein pixels, and several background pixels and finger vein pixels form a training sample set; input the training sample set into the convolutional neural network , Each pixel in the training sample set will generate a feature vector, all feature vectors form a feature vector set, the feature vector set is 256 dimensions, that is, a 16*16 matrix set; the two similar pictures in the feature vector set and A different type of picture composes triplets (triplets), using triplets loss as the loss function for training, specifically, the triplets are input to the convolutional neural network to be trained for training, using the stochastic gradient descent algorithm, and the volume is adjusted according to the value of the loss function
  • the product neural network performs parameter optimization, and when the value of the loss function
  • composition of triplets includes three types of samples: anchor, positive, and negative.
  • anchor is a reference sample
  • positive is a sample of the same type that belongs to the same type as the anchor
  • negative is a heterogeneous sample that belongs to a different type from the anchor.
  • the calculation formula of the loss function is:
  • Eigenvectors of reference samples Represents the feature vector of the i-th sample of the same kind, Represents the feature vector of the i-th heterogeneous sample, ⁇ represents the inter-class interval parameter, i represents the serial number of the current comparison, and N represents the number of triplets used for comparison.
  • the ROI area is extracted from the collected near-infrared sample images of finger veins
  • the finger vein feature vector is extracted from the ROI area using a convolutional neural network to obtain the target finger vein Feature vector, respectively calculate the similarity between the target finger vein feature vector and all the finger vein feature vectors stored in the preset database to obtain multiple similarities, and judge whether the identity verification is passed according to the similarity, avoiding the use of minutia-based
  • Traditional algorithms with features such as texture and statistical characteristics have led to the problem of low identity authentication and recognition efficiency, which has achieved the effect of improving the efficiency of identity authentication and recognition.
  • the embodiment of the present application also provides an identity verification device based on finger vein recognition.
  • the device is used to perform the above-mentioned identity verification method based on finger vein recognition.
  • the device includes: a first obtaining unit 10, Unit 11, second acquisition unit 12, filtering unit 13, first extraction unit 14, second extraction unit 15, third acquisition unit 16, calculation unit 17, first determination unit 18, and second determination unit 19.
  • the first obtaining unit 10 is configured to obtain a finger vein width parameter.
  • the construction unit 11 is used to construct the target convolution kernel according to the finger vein width parameter.
  • the second acquiring unit 12 is used to acquire a finger vein image of the target user.
  • the filtering unit 13 is used for filtering the finger vein image of the target user by using the target convolution kernel.
  • the first extraction unit 14 is configured to extract the ROI area of the finger vein image of the target user after the filtering process to obtain the target area image.
  • the second extraction unit 15 is used to extract the finger vein feature vector from the target area image to obtain the target finger vein feature vector.
  • the third acquiring unit 16 is configured to acquire all finger vein feature vectors stored in a preset database, and the preset database stores the association relationship between the user and the finger vein feature vectors.
  • the calculation unit 17 is configured to calculate the similarity between the target finger vein feature vector and all the finger vein feature vectors stored in the preset database to obtain multiple similarities.
  • the first determining unit 18 is configured to determine that the identity verification of the target user is successful if at least one of the multiple similarities is greater than or equal to the preset similarity threshold.
  • the second determining unit 19 is configured to determine that the identity verification of the target user fails if the multiple similarities are all less than the preset similarity threshold.
  • the first extraction unit 14 includes: a segmentation subunit, an edge detection subunit, a comparison subunit, a first determination subunit, an edge tracking subunit, and a processing subunit.
  • the segmentation subunit is used to perform super pixel segmentation on the finger vein image of the target user to obtain the first image.
  • the edge detection subunit is used to perform Sobel operator edge detection on the finger vein image of the target user to obtain the second image.
  • the comparison subunit is used to compare the first image with the second image to obtain a plurality of first pixel points, where the first pixel points are pixels overlapping the first image and the second image.
  • the first determining subunit is used to determine the second pixel according to the first pixel, and the second pixel is one of the multiple first pixels.
  • the edge tracking subunit is used to perform edge tracking on the first image according to the second pixel to obtain a third image, which is a complete finger edge image.
  • the processing subunit is used to perform angle correction, height cropping, width cropping and normalization on the third image to obtain the target area image.
  • the processing subunit includes: an angle correction module, a height cropping module, and a width cropping module.
  • the angle correction module is used to perform angle correction on the third image according to the center line of the finger to obtain the third image after angle correction.
  • the height cropping module is used to perform height cropping on the third image after angle correction according to the projection values of the finger joints.
  • the width cropping module is used to perform width cropping on the third image after height cropping according to the inner tangent line of the contour of the finger edge.
  • the first extraction unit 14 includes: an extraction subunit, an interception subunit, a gradient operation subunit, a positioning subunit, and a second determination subunit.
  • the extraction subunit is used to extract a complete finger image from the finger vein image of the target user.
  • the interception subunit is used to intercept the largest rectangular area from the complete finger image according to the boundary information of the finger contour.
  • the gradient operation subunit is used to perform gradient operation on the largest rectangular area to generate a gradient image.
  • the positioning subunit is used to locate the joint position according to the gradient image.
  • the second determining subunit is used to determine the target area image according to the joint position.
  • the positioning subunit includes: a summation module and a determination module.
  • the summation module is used to perform row pixel summation operation on the gradient image to obtain the gray distribution curve of the row pixel sum of the gradient image.
  • the determining module is used to take the position of the crest point of the gray distribution curve as the joint position.
  • the embodiment of the present application provides a computer-readable storage medium.
  • the computer-readable storage medium may be non-volatile or volatile.
  • the computer-readable storage medium includes a stored program, wherein the computer is controlled when the program is running.
  • the device where the readable storage medium is located executes the identity verification method based on finger vein recognition described in any of the above embodiments.
  • An embodiment of the application provides a computer device, including a memory and a processor, the memory is used to store information including program instructions, the processor is used to control the execution of the program instructions, and the program instructions are loaded and executed by the processor to implement any of the above The identity verification method based on finger vein recognition described in the embodiment.
  • Fig. 3 is a schematic diagram of a computer device provided by an embodiment of the present application.
  • the computer device 50 of this embodiment includes: a processor 51, a memory 52, and a computer program 53 stored in the memory 52 and running on the processor 51.
  • the computer program 53 is executed by the processor 51, To implement the identity verification method based on finger vein recognition in the embodiment, in order to avoid repetition, it will not be repeated here.
  • the computer program is executed by the processor 51, the function of each model/unit in the identity verification device based on finger vein recognition in the embodiment is realized. To avoid repetition, it will not be repeated here.
  • the computer device 50 may be a computing device such as a desktop computer, a notebook, a palmtop computer, and a cloud server.
  • the computer device may include, but is not limited to, a processor 51 and a memory 52.
  • FIG. 3 is only an example of the computer device 50, and does not constitute a limitation on the computer device 50. It may include more or less components than those shown in the figure, or a combination of certain components, or different components.
  • computer equipment may also include input and output devices, network access devices, buses, and so on.
  • the so-called processor 51 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor or the like.
  • the memory 52 may be an internal storage unit of the computer device 50, such as a hard disk or memory of the computer device 50.
  • the memory 52 may also be an external storage device of the computer device 50, such as a plug-in hard disk equipped on the computer device 50, a smart memory card (Smart Media Card, SMC), a Secure Digital (SD) card, and a flash memory card (Flash). Card) and so on.
  • the memory 52 may also include both an internal storage unit of the computer device 50 and an external storage device.
  • the memory 52 is used to store computer programs and other programs and data required by the computer equipment.
  • the memory 52 can also be used to temporarily store data that has been output or will be output.
  • the disclosed system, device, and method may be implemented in other ways.
  • the device embodiments described above are merely illustrative, for example, the division of the units is only a logical function division, and there may be other divisions in actual implementation, for example, multiple units or components may be combined Or it can be integrated into another system, or some features can be ignored or not implemented.
  • the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • each unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
  • the above-mentioned integrated unit may be implemented in the form of hardware, or may be implemented in the form of hardware plus software functional units.
  • the above-mentioned integrated unit implemented in the form of a software functional unit may be stored in a computer readable storage medium.
  • the above-mentioned software functional unit is stored in a storage medium and includes several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor (Processor) execute the method described in each embodiment of the present application Part of the steps.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disk and other media that can store program code .

Landscapes

  • Engineering & Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Collating Specific Patterns (AREA)

Abstract

Des modes de réalisation, la présente invention concerne un procédé et un appareil d'authentification d'identité basée sur la reconnaissance de veine de doigt. La présente invention se rapporte au domaine de l'intelligence artificielle. Le procédé consiste à : obtenir une image de veine de doigt d'un utilisateur cible ; filtrer l'image de veine de doigt de l'utilisateur cible à l'aide d'un noyau de convolution cible ; extraire une ROI de l'image de veine de doigt filtrée de l'utilisateur cible pour obtenir une image de région cible ; extraire un vecteur de caractéristique de veine de doigt de l'image de région cible pour obtenir un vecteur de caractéristique de veine de doigt cible ; obtenir tous les vecteurs de caractéristique de veine de doigt stockés dans une base de données prédéfinie ; calculer séparément la similarité entre le vecteur de caractéristique de veine de doigt cible et chacun des vecteurs de caractéristique de veine de doigt stockés dans la base de données prédéfinie pour obtenir une pluralité de similarités ; et si au moins une similarité de la pluralité de similarités est supérieure ou égale à un seuil de similarité prédéfini, déterminer que l'authentification d'identité pour l'utilisateur cible a réussi. La solution technique proposée par les modes de réalisation de la présente invention résout le problème selon lequel l'authentification d'identité basée sur une image de veine de doigt a une faible précision.
PCT/CN2020/093383 2019-08-13 2020-05-29 Procédé et appareil d'authentification d'identité basée sur la reconnaissance de veine de doigt WO2021027364A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910744258.5A CN110717372A (zh) 2019-08-13 2019-08-13 基于指静脉识别的身份验证方法和装置
CN201910744258.5 2019-08-13

Publications (1)

Publication Number Publication Date
WO2021027364A1 true WO2021027364A1 (fr) 2021-02-18

Family

ID=69210157

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/093383 WO2021027364A1 (fr) 2019-08-13 2020-05-29 Procédé et appareil d'authentification d'identité basée sur la reconnaissance de veine de doigt

Country Status (2)

Country Link
CN (1) CN110717372A (fr)
WO (1) WO2021027364A1 (fr)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112883356A (zh) * 2021-03-31 2021-06-01 中国工商银行股份有限公司 一种身份认证方法、装置及设备
CN113505716A (zh) * 2021-07-16 2021-10-15 重庆工商大学 静脉识别模型的训练方法、静脉图像的识别方法及装置
CN115063375A (zh) * 2022-02-18 2022-09-16 厦门中翎易优创科技有限公司 一种对排卵试纸检测结果进行自动分析的图像识别方法
CN116226822A (zh) * 2023-05-05 2023-06-06 深圳市魔样科技有限公司 一种智能戒指身份数据采集方法及系统

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110717372A (zh) * 2019-08-13 2020-01-21 平安科技（深圳）有限公司 基于指静脉识别的身份验证方法和装置
CN111914616B (zh) * 2020-06-05 2024-04-16 华南理工大学 指静脉识别与防伪一体化方法、装置、存储介质和设备
CN112560710B (zh) * 2020-12-18 2024-03-01 北京曙光易通技术有限公司 一种用于构建指静脉识别系统的方法及指静脉识别系统
CN112711748B (zh) * 2021-01-15 2023-12-05 电子科技大学中山学院 一种指静脉身份认证方法、装置、电子设备及存储介质
CN112801031A (zh) * 2021-02-08 2021-05-14 电子科技大学中山学院 静脉图像识别方法、装置、电子设备及可读存储介质
CN113516096B (zh) * 2021-07-29 2022-07-19 中国工商银行股份有限公司 指静脉roi区域提取方法及装置

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105975974A (zh) * 2016-05-10 2016-09-28 深圳市金脉智能识别科技有限公司 一种手指静脉识别中提取roi图像的方法
CN106529468A (zh) * 2016-11-07 2017-03-22 重庆工商大学 一种基于卷积神经网络的手指静脉识别方法及系统
CN108830158A (zh) * 2018-05-16 2018-11-16 天津大学 手指轮廓和梯度分布相融合的静脉感兴趣区域提取方法
CN109271966A (zh) * 2018-10-15 2019-01-25 广州广电运通金融电子股份有限公司 一种基于指静脉的身份认证方法、装置及设备
KR20190014912A (ko) * 2017-08-04 2019-02-13 동국대학교 산학협력단 지정맥 인식 장치 및 방법
CN109934102A (zh) * 2019-01-28 2019-06-25 浙江理工大学 一种基于图像超分辨率的指静脉识别方法
CN110717372A (zh) * 2019-08-13 2020-01-21 平安科技（深圳）有限公司 基于指静脉识别的身份验证方法和装置

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101477622A (zh) * 2009-01-15 2009-07-08 山东大学 手指静脉识别系统
CN107729820B (zh) * 2017-09-27 2020-07-21 五邑大学 一种基于多尺度hog的手指静脉识别方法
CN108520211A (zh) * 2018-03-26 2018-09-11 天津大学 基于手指折痕的手指静脉图像特征的提取方法
CN108615002A (zh) * 2018-04-22 2018-10-02 广州麦仑信息科技有限公司 一种基于卷积神经网络的手掌静脉认证方法
CN108805023B (zh) * 2018-04-28 2023-12-19 平安科技（深圳）有限公司 一种图像检测方法、装置、计算机设备及存储介质
CN109815869A (zh) * 2019-01-16 2019-05-28 浙江理工大学 一种基于fcn全卷积网络的指静脉识别方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105975974A (zh) * 2016-05-10 2016-09-28 深圳市金脉智能识别科技有限公司 一种手指静脉识别中提取roi图像的方法
CN106529468A (zh) * 2016-11-07 2017-03-22 重庆工商大学 一种基于卷积神经网络的手指静脉识别方法及系统
KR20190014912A (ko) * 2017-08-04 2019-02-13 동국대학교 산학협력단 지정맥 인식 장치 및 방법
CN108830158A (zh) * 2018-05-16 2018-11-16 天津大学 手指轮廓和梯度分布相融合的静脉感兴趣区域提取方法
CN109271966A (zh) * 2018-10-15 2019-01-25 广州广电运通金融电子股份有限公司 一种基于指静脉的身份认证方法、装置及设备
CN109934102A (zh) * 2019-01-28 2019-06-25 浙江理工大学 一种基于图像超分辨率的指静脉识别方法
CN110717372A (zh) * 2019-08-13 2020-01-21 平安科技（深圳）有限公司 基于指静脉识别的身份验证方法和装置

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112883356A (zh) * 2021-03-31 2021-06-01 中国工商银行股份有限公司 一种身份认证方法、装置及设备
CN112883356B (zh) * 2021-03-31 2024-04-23 中国工商银行股份有限公司 一种身份认证方法、装置及设备
CN113505716A (zh) * 2021-07-16 2021-10-15 重庆工商大学 静脉识别模型的训练方法、静脉图像的识别方法及装置
CN113505716B (zh) * 2021-07-16 2022-07-01 重庆工商大学 静脉识别模型的训练方法、静脉图像的识别方法及装置
CN115063375A (zh) * 2022-02-18 2022-09-16 厦门中翎易优创科技有限公司 一种对排卵试纸检测结果进行自动分析的图像识别方法
CN115063375B (zh) * 2022-02-18 2024-06-04 厦门中翎易优创科技有限公司 一种对排卵试纸检测结果进行自动分析的图像识别方法
CN116226822A (zh) * 2023-05-05 2023-06-06 深圳市魔样科技有限公司 一种智能戒指身份数据采集方法及系统
CN116226822B (zh) * 2023-05-05 2023-07-14 深圳市魔样科技有限公司 一种智能戒指身份数据采集方法及系统

Also Published As

Publication number Publication date
CN110717372A (zh) 2020-01-21

Similar Documents

Publication Publication Date Title
WO2021027364A1 (fr) Procédé et appareil d'authentification d'identité basée sur la reconnaissance de veine de doigt
US10789465B2 (en) Feature extraction and matching for biometric authentication
Qiu et al. Finger-vein recognition based on dual-sliding window localization and pseudo-elliptical transformer
Qin et al. Finger-vein verification based on the curvature in Radon space
US10748017B2 (en) Palm vein identification method and device
Kang et al. Contactless palm vein recognition using a mutual foreground-based local binary pattern
Yang et al. Comparative competitive coding for personal identification by using finger vein and finger dorsal texture fusion
Matsuda et al. Finger-vein authentication based on deformation-tolerant feature-point matching
Han et al. Palm vein recognition using adaptive Gabor filter
WO2017059591A1 (fr) Procédé et dispositif d'identification de veine de doigt
CN107980140B (zh) 一种掌静脉的识别方法及装置
CN102542281B (zh) 非接触式生物特征识别方法和系统
CN107729820B (zh) 一种基于多尺度hog的手指静脉识别方法
TWI434220B (zh) 利用手掌靜脈之生物特徵辨識使用者之方法
CN100373397C (zh) 一种虹膜图像预处理方法
Lu et al. Finger vein identification using polydirectional local line binary pattern
Huang et al. A novel iris segmentation using radial-suppression edge detection
Jaswal et al. Multiple feature fusion for unconstrained palm print authentication
WO2021027155A1 (fr) Procédé et appareil de vérification basés sur une image de veine de doigt, et support d'informations et dispositif informatique
Oldal et al. Biometric Authentication System based on Hand Geometry and Palmprint Features.
Arakala et al. The palm vein graph feature extraction and matching
Zhang et al. Reflection-type finger vein recognition for mobile applications
Yang et al. Finger vein verification based on neighbor pattern coding
CN112801034A (zh) 一种指静脉识别装置
Sontakke et al. Automatic ROI extraction and vein pattern imaging of dorsal hand vein images

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20852160

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20852160

Country of ref document: EP

Kind code of ref document: A1