Multi-identity information verification device

Abstract

The invention relates to the technical field of information verification, and discloses a multi-identity information verification device, which comprises an information collector, a central processing unit and a display, and is characterized in that the central processing unit comprises a data preprocessing module, a feature extraction module and an information verification module; according to the invention, identity card numbers, fingerprints, face images and iris information are covered; identity card number verification can screen out format errors, fingerprint noise reduction enables lines to be clear, face image graying retains key features and reduces data volume, and iris image enhancement facilitates subsequent operation. Key features are extracted by adopting corresponding effective methods for various types of information; the verification accuracy is ensured through a rigorous comparison means; whether notification is passed or not and specific inconsistent types are clearly determined, so that a user can conveniently know; and all links cooperate with one another, so that the reliability, accuracy and convenience of identity verification are effectively improved, and strict identity verification requirements in multiple scenes can be met.

Description

Technical Field

[0001] The present invention relates to the technical field of information verification, and in particular to a multiple identity information verification device. Background Art

[0002] In today's digital society, identity verification plays a key role in many fields, such as financial transactions, access control, airport security, etc. With the development of technology and the increase in security needs, the traditional single identity verification method can no longer meet the requirements of high accuracy and high security.

[0003] In the past, relying solely on ID card number verification could easily lead to misidentification due to problems such as number forgery and fraud, and it was impossible to accurately identify the authenticity of the identity. For simple fingerprint recognition, the accuracy rate will drop significantly when fingerprint collection is affected by the environment, such as when the finger is wet or stained; facial recognition may also make recognition errors if there are changes in lighting or significant changes in facial expressions; although iris recognition is accurate, the early equipment collection accuracy is insufficient and the processing algorithm is imperfect, so there is also a risk of misidentification.

[0004] In addition, various industries also have demands for faster and more convenient identity verification. For example, in public places with large traffic, lengthy and complicated verification processes will cause congestion and reduce traffic efficiency. Based on this, it is an inevitable trend to develop a multi-identity information verification device that integrates multiple identity information, has a rigorous processing process, and can both accurately verify and take into account efficiency, in order to cope with complex and changing real-world application scenarios and effectively ensure the reliability and security of identity verification. Summary of the invention

[0005] The purpose of the present invention is to provide a multiple identity information verification device to solve the technical problems raised in the background technology.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A multiple identity information verification device, comprising:

[0008] The information collector is used to collect different types of identity information; different types of identity information include: user's ID number, user's fingerprint information, user's facial image, user's iris information;

[0009] The central processing unit includes a data preprocessing module, a feature extraction module and an information verification module;

[0010] The data preprocessing module is used to perform standardization processing on different types of identity information, and the standardization processing includes ID card number verification, fingerprint information noise reduction processing, facial image grayscale processing, and iris information image enhancement processing;

[0011] The feature extraction module is used to extract identity card number features, fingerprint information features, facial image features, and iris information features;

[0012] The information verification module is used to perform identity card number feature comparison, fingerprint feature matching, facial feature comparison, and iris feature matching;

[0013] The display is used to display the identity information verification results to the user based on the verification results of the ID card number, fingerprint, facial image and iris information by the information verification module.

[0014] As a further solution of the present invention: the processing method of the data preprocessing module is as follows:

[0015] Step 1.1, ID card number verification:

[0016] First, mark the 18-digit ID number as ID = {a1, a2, ... a 18};

[0017] Then, AI is subjected to ID verification, and the ID card number is judged to meet the format requirements based on the ID verification result;

[0018] Step 1.2: Fingerprint information noise reduction processing:

[0019] Get the gray value of each pixel on the fingerprint image and mark it as F(x, y);

[0020] Wherein, (x, y) represents the coordinates of each pixel point on the fingerprint image, x = 1, 2, ... n, y = 1, 2, ... m, n represents the number of columns of pixels on the fingerprint image, and m represents the number of rows of pixels on the fingerprint image;

[0021] Then a p×q filter window is selected with (x, y) as the center;

[0022] Then pass:

[0023] Calculate the average grayscale value FP(x, y) of all pixels in the p×q filter window, and use the average grayscale value FP(x, y) as the grayscale value of the corresponding pixel after filtering;

[0024] Wherein, i = -p / 2, ... p / 2, j = -q / 2, ... q / 2;

[0025] Step 1.3: Grayscale processing of facial images:

[0026] Obtain a color facial image and mark the RGB value of each pixel as RGB(u,v)=[R(u,v),G(u,v),B(u,v)];

[0027] Wherein, (u, v) represents the coordinates of each pixel on the facial image, u = 1, 2, ... h, v = 1, 2, ... g, h represents the number of columns of pixels on the fingerprint image, k represents the number of rows of pixels on the fingerprint image, R represents the red channel value, G represents the green channel value, and B represents the blue channel value;

[0028] Then pass: H(u, v) = 0.299 × R (u, v) + 0.587 × G (u, v) + 0.114 × B (u, v)

[0029] Calculate the grayscale value H(u, v) of each pixel on the color facial image;

[0030] Step 1.4: Iris information image enhancement processing:

[0031] The histogram equalization algorithm is used to enhance the collected iris image. The algorithm is as follows:

[0032] In the formula, S k represents the new gray value corresponding to the pixel with gray level after histogram equalization; L represents the number of gray levels in the image; k is a variable that represents the gray level in the original iris image; H t is the grayscale histogram of the original iris image. Specifically, H j represents the number of pixels with gray value j in the original iris image; N represents the total number of pixels in the original iris image.

[0033] As a further solution of the present invention: the ID verification process is as follows:

[0034] Step 1.1.1, calculate the sum of the products of the first 17 digits and the corresponding weighting coefficients;

[0035] Right now:

[0036] Where S represents the sum of the products of the first 17 digits and the corresponding weighting coefficients, i0 = 1, 2, ... 17, λ i0 is a corresponding preset weighting coefficient, wherein the specific value of each weighting coefficient is determined according to the identity card number verification rule;

[0037] Step 1.1.2, calculate the remainder when S is divided by 11;

[0038] That is, r = Smod11;

[0039] In the formula, r is the remainder obtained by dividing S by 11, and the meaning of mod in the formula is to find the remainder;

[0040] Step 1.1.3, extract the corresponding preset remainder-check code relationship table, and determine the check code M according to the remainder r 18 ;

[0041] When M 18 =a 18 If established, it is determined that the ID card number meets the format requirements;

[0042] When M 18 =a 18 If it is not true, it is determined that the ID card number does not meet the format requirements.

[0043] As a further solution of the present invention: the remainder-check code relationship table is as follows:

[0044] When the remainder is 0, the check code M 18 is 1;

[0045] When the remainder r is 1, the check code M 18 is 0;

[0046] When the remainder r is 2, the check code M 18 X, X refers to the Roman numeral 10;

[0047] When the remainder r is 3, the check code M 18 is 9;

[0048] When the remainder r is 4, the check code M 18 is 8;

[0049] When the remainder r is 5, the check code M 18 is 7;

[0050] When the remainder r is 6, the check code M 18 is 6;

[0051] When the remainder r is 7, the check code M 18 is 5;

[0052] When the remainder r is 8, the check code M 18 is 4;

[0053] When the remainder r is 9, the check code M 18 is 3;

[0054] When the remainder r is 10, the check code M 18 is 2.

[0055] As a further solution of the present invention: the extraction method of the feature extraction module is as follows:

[0056] Step 2.1, ID card number feature extraction:

[0057] Extract key features corresponding to date of birth and gender from the verified ID number;

[0058] Step 2.2, fingerprint information feature extraction:

[0059] The fingerprint image is divided into several small blocks, and the direction field is obtained by calculating the gradient direction of the fingerprint image block. The detail points are determined by analyzing the direction field and judging the threshold. The specific method of fingerprint information feature extraction is as follows:

[0060] Step 2.2.1, direction field estimation:

[0061] Step 2.2.1.1, first divide the fingerprint image into several small blocks of 8×8 pixels;

[0062] Step 2.2.1.2, in each small block, obtain the gray value of each pixel and mark it as F0(x0, y0);

[0063] Wherein, (x0, y0) represents the coordinates of each pixel point in the small block, x0=1, 2, ... n0, y0=1, 2, ... m0, n0 represents the number of columns of the pixel point in the small block, and m0 represents the number of rows of the pixel point in the small block;

[0064] Step 2.2.1.3. Label the horizontal gradient at (x0, y0) as G x0 (x0, y0) = F0(x0+1, y0) - F(x0-1, y0), the vertical gradient is marked as G y0 (x0, y0)=F0(x0, y0+1)-F(x0, y0-1);

[0065] Step 2.2.1.4, by:

[0066] Calculate the gradient magnitude G0(x, y) and gradient direction θ0(x, y) of the pixels in each small block;

[0067] Step 2.2.1.5: The gradient direction of each pixel in each small block is used to calculate the direction field of the small block:

[0068] The formula is:

[0069] Where θb is the direction field of the corresponding small block, i1∈b, j1∈b, and b represents the pixel set in the block;

[0070] Step 2.2.2, detail point extraction:

[0071] After obtaining the direction field, the minutiae of the fingerprint are extracted by analyzing the changes of the direction field in the local area. The minutiae include endpoints and bifurcation points.

[0072] Step 2.2.2.1, endpoint determination:

[0073] In a local area, if the direction field changes from having a value to 0, and the curvature of the area is lower than the predetermined curvature threshold range, it is judged as an endpoint;

[0074] Step 2.2.2.2, bifurcation point determination:

[0075] When in a local area, the direction field suddenly changes from one direction value to multiple different direction values, and the angles between the multiple different direction values ​​are within a preset angle range, and the curvature of the area is within a preset curvature threshold range, it is determined to be a bifurcation point;

[0076] Step 2.3, facial image feature extraction:

[0077] A shape-based feature extraction method is used to extract facial feature points, which include the positions and shapes of eyes, nose, and mouth.

[0078] Step 2.4, iris information feature extraction:

[0079] The iris image is filtered using Gabor filter to obtain iris texture features of different scales and directions.

[0080] As a further solution of the present invention: the date of birth is obtained by intercepting the character string corresponding to the 7th to 14th digits in the ID number, and the gender feature is determined based on the 17th digit in the ID number, where odd numbers represent males and even numbers represent females.

[0081] As a further solution of the present invention: the verification method of the information verification module is as follows:

[0082] Step 3.1, ID card number feature comparison:

[0083] Perform a string comparison between the features extracted from the input ID number and the pre-stored legitimate ID number features to check whether the features such as date of birth and gender are consistent. If they are exactly the same, the identity information corresponding to the ID number features is determined to be matched successfully. Otherwise, it is determined that the match fails;

[0084] Step 3.2, fingerprint feature matching:

[0085] Extract the minutiae point set of the fingerprint to be verified and the pre-stored valid fingerprint minutiae point set, and mark them as Z1=[(x c ,y c ,θ c )] and Z2=[(x s ,y s ,θ s)];

[0086] Among them, x c and c Indicates the horizontal and vertical coordinates of the detail point in the fingerprint image corresponding to the fingerprint to be verified, θ c Indicates the direction of the detail point in the fingerprint image corresponding to the fingerprint to be verified, x s and s Represents the horizontal and vertical coordinates of the minutiae point in the pre-stored legal fingerprint corresponding to the fingerprint image, θ s represents the direction of the minutiae points in the fingerprint image corresponding to the pre-stored legal fingerprint, c=1, 2, ... w, s=1, 2, ... f, w represents the number of minutiae points in the fingerprint image corresponding to the fingerprint to be verified, and f represents the number of minutiae points in the fingerprint image corresponding to the pre-stored legal fingerprint;

[0087] Then through:

[0088] Calculate the matching score PS between the fingerprint to be verified and the pre-stored legal fingerprint;

[0089] Where, d cs For detail point pair (x c ,y c ) and (x s ,y s ), μ and η are weight coefficients used to adjust the importance of distance and direction differences in the matching score, and γ is the matching weight of the detail point pair;

[0090] Then the matching score PS is compared with the preset matching score threshold PSy:

[0091] If PS>PSy, the fingerprint matching is determined to be successful; otherwise, the fingerprint matching is determined to be unsuccessful;

[0092] Step 3.3, facial feature comparison:

[0093] Calculate the sum of the Euclidean distances between the facial feature points to be verified and the pre-stored legal facial feature points as the matching error. When the matching error is less than a pre-set error threshold, the facial feature matching is determined to be successful. Otherwise, the facial feature matching is determined to be unsuccessful.

[0094] Step 3.4, iris feature matching:

[0095] The iris texture features of different scales and directions in the iris to be verified are defined as the texture feature vector AT k1 ;

[0096] The iris texture features of different scales and directions in the pre-stored legal iris are defined as the texture feature vector BT k1 ;

[0097] Wherein, k1=1, 2, ..., M0, M0 represents the number of vector elements in the texture feature vector;

[0098] Then through:

[0099] Calculate the similarity SIM between the iris to be verified and the legal iris;

[0100] Then the similarity SIM is compared with the preset similarity threshold SIMy:

[0101] When SIM>SIMy, it is determined that the iris feature matching is successful; otherwise, it is determined that the iris feature matching fails.

[0102] As a further solution of the present invention: in step 3.3, the matching error is calculated as follows:

[0103] Extract the coordinates of the feature points of the face to be verified and mark them (Px i2 , Py i2 ), and extract the coordinates of the feature points of the legal face and mark them (Qx i2 , Qy i2 ), where i2 represents different feature points, i2=1, 2, ... t0, t0 represents the number of feature points;

[0104] Then through:

[0105] Calculate the sum of the distances of each feature point, WE, which is the matching error between the face to be verified and the legitimate face;

[0106] Where D0 i2 is the distance between each feature point.

[0107] As a further solution of the present invention: when the verification result is displayed:

[0108] If all types of identity information are matched successfully, "Identity information verification passed" will be output;

[0109] If one type of identity information fails to match, the output is "Identity information verification failed", and it is clearly stated which type of identity information failed to match.

[0110] Beneficial effects of the present invention:

[0111] The present invention can collect various types of identity information, including the user's ID number, fingerprint information, facial image, and iris information. By comprehensively verifying various identity information, the comprehensiveness and accuracy of identity verification are greatly improved, avoiding possible misjudgments in single information verification, and can adapt to the needs of accurate identity confirmation in various application scenarios.

[0112] The present invention, through a rigorous ID verification process, can accurately determine whether the identity card number meets the format requirements according to the preset weighted coefficient calculation, remainder judgment and the combination of the remainder-verification code relationship table and other steps, effectively exclude numbers with incorrect formats from entering the subsequent verification link, and ensure the accuracy of the basic verification data.

[0113] The present invention adopts a method of selecting a filter window with a pixel point as the center and calculating the average grayscale value of the pixel points in the window to perform noise reduction, which can reduce noise interference in the fingerprint image, make the fingerprint features extracted subsequently more reliable, and help to improve the accuracy of fingerprint feature matching.

[0114] The present invention converts a color facial image into a grayscale image through a specific formula, simplifies the image data while retaining key facial feature information, provides convenience for subsequent operations such as shape-based extraction of facial feature points, and improves the efficiency and accuracy of facial feature extraction and comparison.

[0115] The present invention utilizes a histogram equalization algorithm to enhance the iris image, which can effectively improve the quality of the iris image, highlight key features such as iris texture, and facilitate subsequent more accurate extraction and matching of iris information features.

[0116] The present invention can accurately extract key features such as date of birth and gender from the verified ID card number, providing clear comparison elements for comparison with pre-stored legal ID card number features, which helps to quickly and accurately determine whether the identity information corresponding to the ID card number matches.

[0117] The present invention, through a series of operations such as dividing into small blocks, performing direction field estimation and detail point extraction, first scientifically analyzes the direction field of the fingerprint image, and then accurately extracts detail points based on the judgment conditions of detail points such as endpoints and bifurcation points, so that the fingerprint feature extraction is more detailed and accurate, laying a good foundation for fingerprint feature matching.

[0118] The present invention adopts a shape-based feature extraction method, focusing on the position and shape features corresponding to key facial parts such as eyes, nose, and mouth. This targeted extraction method can effectively capture the unique features of the face, facilitating subsequent accurate comparison with legitimate facial features.

[0119] The present invention uses Gabor filter to filter iris images to obtain iris texture features of different scales and directions, which can comprehensively and accurately mine the feature information contained in the iris image and help improve the success rate of iris feature matching.

[0120] The present invention can intuitively and accurately determine whether the input ID number is consistent with the pre-stored legal number features by comparing key features such as date of birth and gender through character strings, thereby quickly determining the identity information verification result corresponding to the ID number.

[0121] The present invention reasonably defines the detail point set of the fingerprint to be verified and the pre-stored legal fingerprint, calculates the matching score according to a specific formula, and compares it with a preset threshold to determine the matching situation, so that the fingerprint matching process is scientific and quantitative, and the reliability of the fingerprint feature matching result is guaranteed.

[0122] The present invention calculates the sum of the Euclidean distances between the facial feature points to be verified and the legitimate facial feature points as the matching error, and determines the matching result based on the error threshold. This distance measurement-based method can objectively reflect the matching degree of facial features and improve the accuracy of facial feature comparison.

[0123] The present invention defines the iris texture features to be verified and pre-stored as texture feature vectors respectively, determines the matching situation by calculating the similarity and comparing it with a preset threshold, effectively measures the consistency of the iris features, and ensures the accuracy of iris feature matching.

[0124] The present invention displays the verification results clearly and concisely. According to the matching status of all identity information, it clearly outputs "identity information verification passed" or "identity information verification failed" and points out the specific type of identity information that failed to match, so that users can quickly and intuitively understand the verification results and facilitate subsequent further operations and processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0125] The present invention will be further described below in conjunction with the accompanying drawings.

[0126] Figure 1 It is a system block diagram of a multiple identity information verification device of the present invention. DETAILED DESCRIPTION

[0127] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0128] Embodiment 1

[0129] See also Figure 1 As shown, the present invention is a multiple identity information verification device, comprising:

[0130] An information collector, used to collect different types of identity information;

[0131] Different types of identity information include:

[0132] The user manually inputs the ID number through the keyboard equipped on the device; the device is equipped with a fingerprint collector, which collects the user's fingerprint information through contact with the user's finger; the built-in camera of the device takes the user's facial image; and the iris scanner collects the user's iris information;

[0133] The central processing unit is used to extract ID card number features, fingerprint information features, facial image features, and iris information features through a feature extraction module set therein; and also to perform ID card number feature comparison, fingerprint feature matching, facial feature comparison, and iris feature matching through an information verification module set therein;

[0134] The feature extraction module is extracted as follows:

[0135] Step 2.1, ID card number feature extraction:

[0136] Extract key features corresponding to date of birth and gender from the verified ID number;

[0137] The date of birth is obtained by intercepting the string corresponding to the 7th to 14th digits of the ID number, and the gender feature is determined by the 17th digit of the ID number, where odd numbers represent males and even numbers represent females;

[0138] Step 2.2, fingerprint information feature extraction:

[0139] Step 2.2.1, direction field estimation:

[0140] Step 2.2.1.1, first divide the fingerprint image into several small blocks of 8×8 pixels;

[0141] Step 2.2.1.2, in each small block, obtain the gray value of each pixel and mark it as F0(x0, y0);

[0142] Wherein, (x0, y0) represents the coordinates of each pixel point in the small block, x0=1, 2, ... n0, y0=1, 2, ... m0, n0 represents the number of columns of the pixel point in the small block, and m0 represents the number of rows of the pixel point in the small block;

[0143] Step 2.2.1.3. Label the horizontal gradient at (x0, y0) as Gx0 (x0, y0) = F0(x0+1, y0) - F(x0-1, y0), the vertical gradient is marked as G y0 (x0, y0)=F0(x0, y0+1)-F(x0, y0-1);

[0144] Step 2.2.1.4, by:

[0145] Calculate the gradient magnitude G0(x, y) and gradient direction θ0(x, y) of the pixels in each small block;

[0146] Step 2.2.1.5: The gradient direction of each pixel in each small block is used to calculate the direction field of the small block:

[0147] The formula is:

[0148] Where θb is the direction field of the corresponding small block, i1∈b, j1∈b, and b represents the pixel set in the block;

[0149] Step 2.2.2, detail point extraction:

[0150] After obtaining the direction field, the minutiae of the fingerprint are extracted by analyzing the changes of the direction field in the local area. The minutiae include endpoints and bifurcation points.

[0151] Step 2.2.2.1, endpoint determination:

[0152] In a local area, if the direction field changes from having a value to 0, and the curvature of the area is lower than the predetermined curvature threshold range, it is judged as an endpoint;

[0153] Step 2.2.2.2, bifurcation point determination:

[0154] When in a local area, the direction field suddenly changes from one direction value to multiple different direction values, and the angles between the multiple different direction values ​​are within a preset angle range, and the curvature of the area is within a preset curvature threshold range, it is determined to be a bifurcation point;

[0155] Step 2.3, facial image feature extraction:

[0156] A shape-based feature extraction method is used to extract facial feature points, which include the positions and shapes of eyes, nose, and mouth.

[0157] In this embodiment, the method of extracting facial feature points is a prior art, so it is not described in detail. Among them, the position of the eyes can be located by finding the area with large color and grayscale changes in the facial image, and then the shape of the eyes can be further determined based on the texture and shape characteristics around the eyes; other feature points such as the nose and mouth can also be determined by a similar method based on image geometric features and grayscale changes; the feature vector of the face is formed by calculating the geometric relationship between various facial organs, such as the distance between the eyes, the distance from the eyes to the nose, etc.;

[0158] Step 2.4, iris information feature extraction:

[0159] The iris image is filtered using Gabor filter to obtain iris texture features of different scales and directions;

[0160] In this embodiment, the Gabor filter is a linear filter whose impulse response is modulated by a sinusoidal plane wave and a Gaussian kernel function; it has good localization characteristics in both the spatial domain and the frequency domain, and can analyze the signal in both space and frequency. It is a prior art, so it will not be described in detail.

[0161] The verification method of the information verification module is as follows:

[0162] Step 3.1, ID card number feature comparison:

[0163] Perform a string comparison between the features extracted from the input ID number and the pre-stored legitimate ID number features to check whether the features such as date of birth and gender are consistent. If they are exactly the same, the identity information corresponding to the ID number features is determined to be matched successfully. Otherwise, it is determined that the match fails;

[0164] Step 3.2, fingerprint feature matching:

[0165] Extract the minutiae point set of the fingerprint to be verified and the pre-stored valid fingerprint minutiae point set, and mark them as Z1=[(x c ,y c ,θ c )] and Z2=[(x s ,y s ,θ s )];

[0166] Among them, x c and c Indicates the horizontal and vertical coordinates of the detail point in the fingerprint image corresponding to the fingerprint to be verified, θ c Indicates the direction of the detail point in the fingerprint image corresponding to the fingerprint to be verified, x s and sRepresents the horizontal and vertical coordinates of the minutiae point in the pre-stored legal fingerprint corresponding to the fingerprint image, θ s represents the direction of the minutiae points in the fingerprint image corresponding to the pre-stored legal fingerprint, c=1, 2, ... w, s=1, 2, ... f, w represents the number of minutiae points in the fingerprint image corresponding to the fingerprint to be verified, and f represents the number of minutiae points in the fingerprint image corresponding to the pre-stored legal fingerprint;

[0167] Then through:

[0168] Calculate the matching score PS between the fingerprint to be verified and the pre-stored legal fingerprint;

[0169] Where, d cs For detail point pair (x c ,y c ) and (x s ,y s ), μ and η are weight coefficients used to adjust the importance of distance and direction differences in the matching score, and γ is the matching weight of the detail point pair;

[0170] Then the matching score PS is compared with the preset matching score threshold PSy:

[0171] If PS>PSy, the fingerprint matching is determined to be successful; otherwise, the fingerprint matching is determined to be unsuccessful;

[0172] Step 3.3, facial feature comparison:

[0173] Calculate the sum of the Euclidean distances between the facial feature points to be verified and the pre-stored legal facial feature points as the matching error. When the matching error is less than a pre-set error threshold, the facial feature matching is determined to be successful. Otherwise, the facial feature matching is determined to be unsuccessful.

[0174] The matching error is calculated as follows:

[0175] Extract the coordinates of the feature points of the face to be verified and mark them (Px i2 , Py i2 ), and extract the coordinates of the feature points of the legal face and mark them (Qx i2 , Qy i2 ), where i2 represents different feature points, i2=1, 2, ... t0, t0 represents the number of feature points;

[0176] In this embodiment, when i2=1, it indicates the left eye, when i2=2, it indicates the right eye, and when i2=3, it indicates the nose;

[0177] Then through:

[0178] Calculate the sum of the distances of each feature point, WE, which is the matching error between the face to be verified and the legitimate face;

[0179] Where D0 i2 is the distance between each feature point;

[0180] Step 3.4, iris feature matching:

[0181] The iris texture features of different scales and directions in the iris to be verified are defined as the texture feature vector AT k1 ;

[0182] The iris texture features of different scales and directions in the pre-stored legal iris are defined as the texture feature vector BT k1 ;

[0183] Wherein, k1=1, 2, ..., M0, M0 represents the number of vector elements in the texture feature vector;

[0184] Then through:

[0185] Calculate the similarity SIM between the iris to be verified and the legal iris;

[0186] Then the similarity SIM is compared with the preset similarity threshold SIMy:

[0187] When SIM>SIMy, it is determined that the iris feature matching is successful; otherwise, it is determined that the iris feature matching fails.

[0188] This embodiment can fully obtain identity information through a variety of information collection methods, such as inputting ID number through the keyboard, collecting fingerprints with a fingerprint collector, taking facial images with a built-in camera, and collecting iris information with an iris scanner; the feature extraction module adopts corresponding effective methods for different types of information, such as accurately extracting date of birth and gender from the ID number, extracting fingerprint detail points according to specific steps, extracting facial feature points based on shape, and extracting iris texture features with a Gabor filter, laying the foundation for subsequent verification; the information verification module uses scientific comparison methods, such as ID number string comparison, fingerprint matching score calculation, facial feature point distance sum judgment, and iris texture feature similarity comparison, to ensure verification accuracy.

[0189] Embodiment 2

[0190] As the second embodiment of the present invention, when the present application is implemented, compared with the first embodiment, the technical solution of the present embodiment is different from the first embodiment only in that in the present embodiment, the central processing unit is also used to perform standardization processing on different types of identity information through a data preprocessing module set up inside it, and the standardization processing includes identity card number verification, fingerprint information noise reduction processing, facial image grayscale processing, and iris information image enhancement processing;

[0191] The data preprocessing module is processed as follows:

[0192] Step 1.1, ID card number verification:

[0193] First, mark the 18-digit ID number as ID = {a1, a2, ... a 18};

[0194] Then, AI is subjected to ID verification, and the ID card number is judged to meet the format requirements based on the ID verification result;

[0195] The ID verification process is as follows:

[0196] Step 1.1.1, calculate the sum of the products of the first 17 digits and the corresponding weighting coefficients;

[0197] Right now:

[0198] Where S represents the sum of the products of the first 17 digits and the corresponding weighting coefficients, i0 = 1, 2, ... 17, λ i0 is a corresponding preset weighting coefficient, wherein the specific value of each weighting coefficient is determined according to the identity card number verification rule;

[0199] Step 1.1.2, calculate the remainder when S is divided by 11;

[0200] That is, r = Smod11;

[0201] In the formula, r is the remainder obtained by dividing S by 11, and the meaning of mod in the formula is to find the remainder;

[0202] Step 1.1.3, extract the corresponding preset remainder-check code relationship table, and determine the check code M according to the remainder r 18 ;

[0203] When M 18 =a 18 If established, it is determined that the ID card number meets the format requirements;

[0204] When M 18 =a 18 If it is not established, it is determined that the ID card number does not meet the format requirements;

[0205] The remainder-check code relationship table is as follows:

[0206] When the remainder is 0, the check code M 18 is 1;

[0207] When the remainder r is 1, the check code M 18 is 0;

[0208] When the remainder r is 2, the check code M 18 X, X refers to the Roman numeral 10;

[0209] When the remainder r is 3, the check code M 18 is 9;

[0210] When the remainder r is 4, the check code M 18 is 8;

[0211] When the remainder r is 5, the check code M 18 is 7;

[0212] When the remainder r is 6, the check code M 18 is 6;

[0213] When the remainder r is 7, the check code M 18 is 5;

[0214] When the remainder r is 8, the check code M 18 is 4;

[0215] When the remainder r is 9, the check code M 18 is 3;

[0216] When the remainder r is 10, the check code M 18 is 2;

[0217] In this embodiment, this step can preliminarily determine whether the ID card number meets the format requirements to avoid entering the subsequent processing flow if an incorrect number is entered;

[0218] Step 1.2: Fingerprint information noise reduction processing:

[0219] Get the gray value of each pixel on the fingerprint image and mark it as F(x, y);

[0220] Wherein, (x, y) represents the coordinates of each pixel point on the fingerprint image, x = 1, 2, ... n, y = 1, 2, ... m, n represents the number of columns of pixels on the fingerprint image, and m represents the number of rows of pixels on the fingerprint image;

[0221] Then a p×q filter window is selected with (x, y) as the center;

[0222] Then pass:

[0223] Calculate the average grayscale value FP(x, y) of all pixels in the p×q filter window, and use the average grayscale value FP(x, y) as the grayscale value of the corresponding pixel after filtering;

[0224] Wherein, i = -p / 2, ... p / 2, j = -q / 2, ... q / 2;

[0225] In this embodiment, this step can remove noise points in the fingerprint image to make the fingerprint pattern clearer;

[0226] Step 1.3: Grayscale processing of facial images:

[0227] Obtain a color facial image and mark the RGB value of each pixel as RGB(u,v)=[R(u,v),G(u,v),B(u,v)];

[0228] Wherein, (u, v) represents the coordinates of each pixel on the facial image, u = 1, 2, ... h, v = 1, 2, ... g, h represents the number of columns of pixels on the fingerprint image, k represents the number of rows of pixels on the fingerprint image, R represents the red channel value, G represents the green channel value, and B represents the blue channel value;

[0229] Then pass: H(u, v) = 0.299 × R (u, v) + 0.587 × G (u, v) + 0.114 × B (u, v)

[0230] Calculate the grayscale value H(u, v) of each pixel on the color facial image;

[0231] In this embodiment, this step converts the color image into a grayscale image according to the sensitivity of the human eye to different colors, thereby reducing the amount of data for subsequent processing while retaining the main feature information of the face;

[0232] Step 1.4: Iris information image enhancement processing:

[0233] The histogram equalization algorithm is used to enhance the collected iris image. The algorithm is as follows:

[0234] In the formula, S k represents the new gray value corresponding to the pixel with gray level after histogram equalization; L represents the number of gray levels in the image; k is a variable that represents the gray level in the original iris image; H t is the grayscale histogram of the original iris image. Specifically, H j represents the number of pixels with gray value j in the original iris image; N represents the total number of pixels in the original iris image;

[0235] In this embodiment, S k It is the core output result of the histogram equalization algorithm. It enhances the image contrast by remapping the grayscale value of the original image. In common 8-bit grayscale images, L = 256, because 8 bits can represent 2 8 = 256 different grayscale values, ranging from 0 to 255, which determines the upper limit of the grayscale value range after histogram equalization; the value range of k is usually from 0 to 255. For example, in an 8-bit grayscale image, the value range of is 0 to 255. When calculating the histogram equalization, each grayscale level needs to be processed separately; H t By counting the pixels of the entire image, the number of pixels corresponding to each gray value can be obtained to form a grayscale histogram; N is the normalization factor used to calculate the grayscale histogram. When calculating the cumulative distribution function, the number of pixels at each gray level needs to be divided by the total number of pixels to obtain the proportion of the gray level pixels in the image;

[0236] In this embodiment, this step can readjust the grayscale distribution of the iris image, so that the image with uneven grayscale distribution becomes more uniform, thereby enhancing detail information such as iris texture, facilitating subsequent feature extraction and recognition operations;

[0237] In this embodiment, based on the first embodiment, the central processing unit adds a data preprocessing module. The ID card number verification follows a rigorous process to eliminate incorrectly formatted numbers, the fingerprint information is denoised to make the texture clearer, the facial image is grayed out to reduce the amount of data and retain key features, and the iris information image is enhanced to highlight the details. These preprocessings provide better quality data for feature extraction and verification, and improve the overall verification accuracy.

[0238] Embodiment 3

[0239] As the third embodiment of the present invention, when the present application is implemented, compared with the first and second embodiments, the technical solution of this embodiment is to combine the solutions of the first and second embodiments, and the difference between the technical solution of this embodiment and the first and second embodiments is that this embodiment also includes:

[0240] A display, used to display the identity information verification result to the user according to the verification result of the identity card number, fingerprint, facial image and iris information by the information verification module;

[0241] If all types of identity information are matched successfully, "Identity information verification passed" will be output;

[0242] If one type of identity information fails to match, the output is "Identity information verification failed", and it clearly indicates which type of identity information failed to match;

[0243] For example, if the fingerprint does not match, the output is "Identity information verification failed, fingerprint information does not match";

[0244] This embodiment combines the first and second embodiments, and adds a display to intuitively present the verification results, clearly informing the user that "the identity information verification has passed" or pointing out the specific type of information that has failed, such as "the identity information verification has failed, the fingerprint information does not match", so that the user can quickly know the verification status and improve the user experience.

[0245] Embodiment 4

[0246] As the fourth embodiment of the present invention, when the present application is specifically implemented, compared with the first, second and third embodiments, the technical solution of this embodiment is to combine and implement the solutions of the above-mentioned first, second and third embodiments.

[0247] This embodiment integrates the previous three embodiments, and has the advantages of comprehensive information collection, precise feature extraction, scientific verification, and clear result display. It fully guarantees the efficiency, accuracy, and convenience of multiple identity information verification, and can meet the needs of strict identity verification in various scenarios.

[0248] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters and thresholds in the formula are set by technicians in this field according to actual conditions.

[0249] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A multiple identity information verification device, characterized in that: include: The information collector is used to collect different types of identity information; different types of identity information include: user's ID number, user's fingerprint information, user's facial image, user's iris information; The central processing unit includes a data preprocessing module, a feature extraction module and an information verification module; The data preprocessing module is used to perform standardization processing on different types of identity information, and the standardization processing includes ID card number verification, fingerprint information noise reduction processing, facial image grayscale processing, and iris information image enhancement processing; The feature extraction module is used to extract identity card number features, fingerprint information features, facial image features, and iris information features; The information verification module is used to perform identity card number feature comparison, fingerprint feature matching, facial feature comparison, and iris feature matching; The display is used to display the identity information verification results to the user based on the verification results of the ID card number, fingerprint, facial image and iris information by the information verification module.

2. A multiple identity information verification device according to claim 1, characterized in that: The processing method of the data preprocessing module is as follows: Step 1.1, ID card number verification: First, mark the 18-digit ID number as ID = {a1, a2, ... a 18 }; Then, AI is subjected to ID verification, and the ID card number is judged to meet the format requirements based on the ID verification result; Step 1.2: Fingerprint information noise reduction processing: Get the gray value of each pixel on the fingerprint image and mark it as F(x, y); Wherein, (x, y) represents the coordinates of each pixel point on the fingerprint image, x = 1, 2, ... n, y = 1, 2, ... m, n represents the number of columns of pixels on the fingerprint image, and m represents the number of rows of pixels on the fingerprint image; Then a p×q filter window is selected with (x, y) as the center; Then pass: Calculate the average grayscale value FP(x, y) of all pixels in the p×q filter window, and use the average grayscale value FP(x, y) as the grayscale value of the corresponding pixel after filtering; Wherein, i = -p / 2, ... p / 2, j = -q / 2, ... q / 2; Step 1.3: Grayscale processing of facial images: Obtain a color facial image and mark the RGB value of each pixel as RGB(u,v)=[R(u,v),G(u,v),B(u,v)]; Wherein, (u, v) represents the coordinates of each pixel on the facial image, u = 1, 2, ... h, v = 1, 2, ... g, h represents the number of columns of pixels on the fingerprint image, k represents the number of rows of pixels on the fingerprint image, R represents the red channel value, G represents the green channel value, and B represents the blue channel value; Then pass: H(u, v) = 0.299 × R (u, v) + 0.587 × G (u, v) + 0.114 × B (u, v) Calculate the grayscale value H(u, v) of each pixel on the color facial image; Step 1.4: Iris information image enhancement processing: The histogram equalization algorithm is used to enhance the collected iris image. The algorithm is as follows: In the formula, S k represents the new gray value corresponding to the pixel with gray level after histogram equalization; L represents the number of gray levels in the image; k is a variable that represents the gray level in the original iris image; H t is the grayscale histogram of the original iris image. Specifically, H j represents the number of pixels with gray value j in the original iris image; N represents the total number of pixels in the original iris image.

3. A multiple identity information verification device according to claim 2, characterized in that: In step 1.1, the ID verification process is as follows: Step 1.1.1, calculate the sum of the products of the first 17 digits and the corresponding weighting coefficients; Right now: Where S represents the sum of the products of the first 17 digits and the corresponding weighting coefficients, i0 = 1, 2, ... 17, λ i0 is a corresponding preset weighting coefficient, wherein the specific value of each weighting coefficient is determined according to the identity card number verification rule; Step 1.1.2, calculate the remainder when S is divided by 11; That is, r = Smod11; In the formula, r is the remainder obtained by dividing S by 11, and the meaning of mod in the formula is to find the remainder; Step 1.1.3, extract the corresponding preset remainder-check code relationship table, and determine the check code M according to the remainder r 18 ; When M 18 =a 18 If established, it is determined that the ID card number meets the format requirements; When M 18 =a 18 If it is not true, it is determined that the ID card number does not meet the format requirements.

4. A multiple identity information verification device according to claim 3, characterized in that: The remainder-check code relationship table is as follows: When the remainder is 0, the check code M 18 is 1; When the remainder r is 1, the check code M 18 is 0; When the remainder r is 2, the check code M 18 X, X refers to the Roman numeral 10; When the remainder r is 3, the check code M 18 is 9; When the remainder r is 4, the check code M 18 is 8; When the remainder r is 5, the check code M 18 is 7; When the remainder r is 6, the check code M 18 is 6; When the remainder r is 7, the check code M 18 is 5; When the remainder r is 8, the check code M 18 is 4; When the remainder r is 9, the check code M 18 is 3; When the remainder r is 10, the check code M 18 is 2.

5. A multiple identity information verification device according to claim 1, characterized in that: The extraction method of the feature extraction module is as follows: Step 2.1, ID card number feature extraction: Extract key features corresponding to date of birth and gender from the verified ID number; Step 2.2, fingerprint information feature extraction: The fingerprint image is divided into several small blocks, and the direction field is obtained by calculating the gradient direction of the fingerprint image block. The detail points are determined by analyzing the direction field and judging the threshold value. Step 2.3, facial image feature extraction: A shape-based feature extraction method is used to extract facial feature points; Step 2.4, iris information feature extraction: The iris image is filtered using Gabor filter to obtain iris texture features of different scales and directions.

6. A multiple identity information verification device according to claim 5, characterized in that: The specific method of fingerprint information feature extraction is as follows: Step 2.2.1, direction field estimation: Step 2.2.1.1, first divide the fingerprint image into several small blocks of 8×8 pixels; Step 2.2.1.2, in each small block, obtain the gray value of each pixel and mark it as F0(x0, y0); Wherein, (x0, y0) represents the coordinates of each pixel point in the small block, x0=1, 2, ... n0, y0=1, 2, ... m0, n0 represents the number of columns of the pixel point in the small block, and m0 represents the number of rows of the pixel point in the small block; Step 2.2.1.

3. Label the horizontal gradient at (x0, y0) as G x0 (x0, y0) = F0(x0+1, y0) - F(x0-1, y0), the vertical gradient is marked as G y0 (x0, y0)=F0(x0, y0+1)-F(x0, y0-1); Step 2.2.1.4, by: Calculate the gradient magnitude G0(x, y) and gradient direction θ0(x, y) of the pixels in each small block; Step 2.2.1.5: The gradient direction of each pixel in each small block is used to calculate the direction field of the small block: The formula is: Where θb is the direction field of the corresponding small block, i1∈b, j1∈b, and b represents the pixel set in the block; Step 2.2.2, detail point extraction: After obtaining the direction field, the minutiae of the fingerprint are extracted by analyzing the changes of the direction field in the local area. The minutiae include endpoints and bifurcation points. Step 2.2.2.1, endpoint determination: In a local area, if the direction field changes from having a value to 0, and the curvature of the area is lower than the predetermined curvature threshold range, it is judged as an endpoint; Step 2.2.2.2, bifurcation point determination: When in a local area, the direction field suddenly changes from one direction value to multiple different direction values, and the angles between the multiple different direction values ​​are within a preset angle range, and the curvature of the area is within a preset curvature threshold range, it is judged as a bifurcation point.

7. A multiple identity information verification device according to claim 5, characterized in that: The date of birth is obtained by intercepting the character string corresponding to the 7th to 14th digits of the ID number, and the gender feature is determined based on the 17th digit of the ID number, where odd numbers represent males and even numbers represent females.

8. A multiple identity information verification device according to claim 5, characterized in that: The verification method of the information verification module is as follows: Step 3.1, ID card number feature comparison: Perform a string comparison between the features extracted from the input ID number and the pre-stored legitimate ID number features to check whether the features such as date of birth and gender are consistent. If they are exactly the same, the identity information corresponding to the ID number features is determined to be matched successfully. Otherwise, it is determined that the match fails; Step 3.2, fingerprint feature matching: Extract the minutiae point set of the fingerprint to be verified and the pre-stored valid fingerprint minutiae point set, and mark them as Z1=[(x c ,y c ,θ c )] and Z2=[(x s ,y s ,θ s )]; Among them, x c and c Indicates the horizontal and vertical coordinates of the detail point in the fingerprint image corresponding to the fingerprint to be verified, θ c Indicates the direction of the detail point in the fingerprint image corresponding to the fingerprint to be verified, x s and s Represents the horizontal and vertical coordinates of the minutiae point in the pre-stored legal fingerprint corresponding to the fingerprint image, θ s represents the direction of the minutiae points in the fingerprint image corresponding to the pre-stored legal fingerprint, c=1, 2, ... w, s=1, 2, ... f, w represents the number of minutiae points in the fingerprint image corresponding to the fingerprint to be verified, and f represents the number of minutiae points in the fingerprint image corresponding to the pre-stored legal fingerprint; Then through: Calculate the matching score PS between the fingerprint to be verified and the pre-stored legal fingerprint; Where, d cs For detail point pair (x c ,y c ) and (x s ,y s ), μ and η are weight coefficients used to adjust the importance of distance and direction differences in the matching score, and γ is the matching weight of the detail point pair; Then the matching score PS is compared with the preset matching score threshold PSy: If PS>PSy, the fingerprint matching is determined to be successful; otherwise, the fingerprint matching is determined to be unsuccessful; Step 3.3, facial feature comparison: Calculate the sum of the Euclidean distances between the facial feature points to be verified and the pre-stored legal facial feature points as the matching error. When the matching error is less than a pre-set error threshold, the facial feature matching is determined to be successful. Otherwise, the facial feature matching is determined to be unsuccessful. Step 3.4, iris feature matching: The iris texture features of different scales and directions in the iris to be verified are defined as the texture feature vector AT k1 ; The iris texture features of different scales and directions in the pre-stored legal iris are defined as the texture feature vector BT k1 ; Wherein, k1=1, 2, ..., M0, M0 represents the number of vector elements in the texture feature vector; Then through: Calculate the similarity SIM between the iris to be verified and the legal iris; Then the similarity SIM is compared with the preset similarity threshold SIMy: When SIM>SIMy, it is determined that the iris feature matching is successful; otherwise, it is determined that the iris feature matching fails.

9. A multiple identity information verification device according to claim 8, characterized in that: In step 3.3, the matching error is calculated as follows: Extract the coordinates of the feature points of the face to be verified and mark them (Px i2 , Py i2 ), and extract the coordinates of the feature points of the legal face and mark them (Qx i2 , Qy i2 ), where i2 represents different feature points, i2=1, 2, ... t0, t0 represents the number of feature points; Then through: Calculate the sum of the distances of each feature point, WE, which is the matching error between the face to be verified and the legitimate face; Where D0 i2 is the distance between each feature point.

10. A multiple identity information verification device according to claim 1, characterized in that: When the verification result is displayed: If all types of identity information are matched successfully, "Identity information verification passed" will be output; If one type of identity information fails to match, "Identity information verification failed" will be output, and it will clearly indicate which type of identity information failed to match.

Images