CN112069482A - Identity authentication system based on footprint comparison algorithm - Google Patents
Identity authentication system based on footprint comparison algorithm Download PDFInfo
- Publication number
- CN112069482A CN112069482A CN202010912788.9A CN202010912788A CN112069482A CN 112069482 A CN112069482 A CN 112069482A CN 202010912788 A CN202010912788 A CN 202010912788A CN 112069482 A CN112069482 A CN 112069482A
- Authority
- CN
- China
- Prior art keywords
- footprint
- image
- information
- comparison
- area
- Prior art date
- Legal status (The legal status 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 status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 18
- 238000005457 optimization Methods 0.000 claims description 10
- 238000012795 verification Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 8
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000004891 communication Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 230000004913 activation Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/32—User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/36—User authentication by graphic or iconic representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C9/00—Individual registration on entry or exit
- G07C9/30—Individual registration on entry or exit not involving the use of a pass
- G07C9/32—Individual registration on entry or exit not involving the use of a pass in combination with an identity check
- G07C9/37—Individual registration on entry or exit not involving the use of a pass in combination with an identity check using biometric data, e.g. fingerprints, iris scans or voice recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Security & Cryptography (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Human Computer Interaction (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses an identity authentication system based on a footprint comparison algorithm, relates to the technical field of identity authentication, and particularly relates to an identity authentication system based on a footprint comparison algorithm, which can help a specific place of a warehouse, a bank and a prison to realize the automation of entrance and exit identity authentication; a footprint is one of the biological features that can characterize a person's identity information; the acquisition device is arranged at a necessary position of an entrance and an exit of a relevant place without being matched by an acquisition person intentionally in the acquisition process of the footprints, and the footprints can be acquired in a concealed manner by being disguised as normal ground; by adopting the identity authentication system based on the footprint comparison algorithm, internal personnel in related places only need to collect footprint data once when registering information, and identity authentication work is completed by the footprint identity authentication system laid at an entrance and an exit; the problem that the security protection capability of security systems in related places of warehouses, banks and prisons is not strong enough at present is solved.
Description
Technical Field
The invention relates to the technical field of identity authentication, in particular to an identity authentication system based on a footprint comparison algorithm.
Background
The storage of goods, the office area of a bank, prisons, and other scenarios are in principle not accessible to non-insider personnel, and therefore require security monitoring of these areas. According to the traditional security method, cameras or an access control system based on an IC card need to be arranged at an entrance and an exit, so that the workload of security workers is increased in the actual operation process, and troubles are brought to the workers needing to frequently come in and go out of the places. And the ways are not strong in concealment and are easily confused by lawless persons.
The invention provides an identity authentication system based on a footprint comparison algorithm, which can be used for constructing an identity authentication system and helping a specific place of a warehouse, a bank and a prison to realize the automation of entrance and exit identity authentication; a footprint is one of the biological features that can characterize a person's identity information; the footprints are not required to be matched by the collection personnel in the collection process, the collection device is arranged at the necessary part of the entrance and exit of the relevant places, and the collection device is disguised as the normal ground to carry out hidden footprint collection.
By adopting the identity authentication system based on the footprint comparison algorithm, internal personnel in related places only need to collect footprint data once during information registration, and then can freely come in and go out, and identity authentication work is completed by the footprint identity authentication system laid at an entrance and an exit.
Disclosure of Invention
Technical problem to be solved
Aiming at the defects of the prior art, the invention provides an identity authentication system based on a footprint comparison algorithm, and solves the problems that the security protection capability of security systems in related places of warehouses, banks and prisons is not strong enough and the operation process is relatively troublesome at present.
(II) technical scheme
In order to achieve the purpose, the invention provides the following technical scheme: an identity authentication system based on footprint comparison algorithm comprises a supervision area, a gate channel, a footprint collector and a footprint authentication system, wherein a gate channel is respectively arranged in an entrance channel and an exit channel of the supervision area, and the footprint collector is arranged at an entrance end of the gate channel; the gate channel and the footprint collector are in communication connection with the footprint authentication system.
As optimization, the backbone network of the footprint authentication system is VGG 11; firstly, splicing two footprint images to be compared in the depth direction, inputting the two footprint images into VGG11 to extract comparison features, inputting a 1000-dimensional vector output from the last layer of VGG11 as a comparison feature vector into a Similarity Score Computing Network (SSCN) designed by the invention, and calculating the Similarity between the two footprints; if the similarity is larger than the set threshold value TH, the two footprints can be considered to belong to the same person, otherwise, the two footprints can be considered not to belong to the same person. In the footprint comparison network, the similarity fraction calculation network has two fully-connected layers, the fully-connected layer of the first layer receives a 1000-dimensional comparison feature vector output by VGG11, the dimension is reduced to 30 dimensions, a BN (batch normalization) layer and a ReLU activation function layer are connected behind the fully-connected layer of the first layer, the fully-connected layer of the second layer receives the 30-dimensional feature vector output by the first layer, a one-dimensional vector is output as a predicted similarity value, the similarity value cannot be directly used as a final similarity prediction result, and a Sigmoid activation function is required to be used for converting the similarity value into a value between 0 and 1 as a final prediction result.
A footprint alignment algorithm comprising the steps of:
step 1: identity acquisition: collecting information of workers, wherein three numbers 1, 2 and 3 represent identity information of three persons respectively, and when the information is registered, a plurality of footprint images of the three persons, and other information such as identity card numbers, job numbers and the like are collected, wherein the footprint images of each person and the respective personal information are corresponding;
step 2: acquiring and preprocessing the footprint image and archiving information: after the footprint image of the registrant is preprocessed, irrelevant information on the image is removed, the image resolution is reduced to the extent that the image resolution can be stored in a larger scale without influencing the identity authentication precision, and after the preprocessing, the image acquired during the registration is filed and stored in a database;
and step 3: and (3) information comparison: when people enter the field, the people will collect their footprint images, then the images of the people and the images in the database form a comparison image pair, and the comparison image pair is sent to a footprint authentication system, and the footprint authentication system outputs the similarity scores between the images through the comparison of the images; assuming that the threshold value TH set by the system is 0.7, the similarity between the footprints of the persons who perform identity authentication and the data of the footprints of the persons who perform identity authentication is 0.8 > TH, and the person is the person No. 1 in the database; the comparison is passed, and the comparison can be released; if the similarity between any footprint image and the image to be authenticated is not larger than the threshold value in the database, the footprint authentication system considers that the person does not carry out information registration and does not pass.
As an optimization, the step 2: acquiring and preprocessing a footprint image, archiving information, and when the footprint is acquired: the collected person walks through the footprint collector, and the image collected by the footprint collector is a gray image with 256 gray levels;
preprocessing a footprint image: the purpose of preprocessing the image is mainly to remove noise on the image and scale information carried by a collecting instrument so as to facilitate image comparison, and reduce the size of the image so as to facilitate archiving.
As an optimization, the step 2: the footprint image acquisition preprocessing and information archiving method comprises the following steps:
a. removing the scale information on the footprint image, wherein the image fixed area acquired by the footprint collector is provided with scale information for calculating the size of the footprint, and the information has little use for footprint comparison, so that the information needs to be removed, and the specific operation is to directly assign the gray value of the image fixed area to 255;
b. denoising and footprint positioning are carried out based on an Otsu algorithm, and a distinguishing threshold th is determined based on the Otsu algorithm because a footprint area and a background area of the collected footprint image are easy to distinguish; the pixel points with the gray scale smaller than th in the image are regarded as points of the footprint area, the other points are regarded as backgrounds, and the footprint image can be scanned from four directions, namely the upper direction, the lower direction, the left direction and the right direction after the th is determined, four boundaries of the footprint area are detected, and therefore footprint positioning is completed; after the footprint positioning is completed, because most of the noise is outside the footprint area, the denoising work is actually completed at the moment;
c. adjusting the size of the footprint image and carrying out gray inversion, determining the position of the footprint area in the step b, wherein the distance between the upper boundary and the lower boundary of the footprint is larger than the distance between the left boundary and the right boundary under the general condition, so that the system considers that the footprint image with the distance between the upper boundary and the lower boundary smaller than the distance between the left boundary and the right boundary is an abnormal footprint image, and an alarm is sent out in the specific implementation of the system; after the distance between the upper boundary and the lower boundary is determined, a square area with the pixel value of 255 is defined, and the side length is the distance between the upper boundary and the lower boundary; then directly copying the footprint area to the midpoint of the square area, and subtracting the pixel value of each point of the square area from 255 to perform gray level inversion at the pixel level; and finally, downsampling the obtained square area, adjusting the resolution and storing.
As an optimization, the step 3: the information comparison comprises the following steps:
step one, network training: dividing the collected footprint sample into two parts according to the collected object, wherein the footprint sample of one part of the object is used as a training set sample, and the other part of the object is used as a verification set sample;
if two footprint images input to the network belong to the same person, the two images constitute a positive sample X with a label Y of 1, whereas the two images constitute a negative sample X with a labelY is 0; randomly sampling from a training set each time to obtain a Batch of (Batch) positive samples or negative samples, inputting the Batch of (Batch) positive samples or negative samples into a network for similarity calculation, and making an MSE loss on an output value and a label, wherein the loss expression is as follows:
in the formula, B represents the batch size during training, FCN represents a footprint comparison network, and any sample X is input into the network to obtain a prediction label of the network;
during training, performing back propagation on the loss L of each batch of samples, and then optimizing the model by using a gradient descent method; when the training batches are enough, the network is equivalent to parameter optimization on all possible positive and negative sample spaces; through repeated random extraction of positive and negative samples for model optimization, the model is gradually converged finally;
secondly, selecting a model: the model to be finally stored in the neural network generally needs to be selected from the trained intermediate models by using a verification set, enough samples need to be randomly selected from the verification set during model selection, the samples are input into the network for similarity calculation, if the similarity is more than 0.5, the model is regarded as the footprint of the same person, if the similarity is less than 0.5, the model is not the footprint of the same person, and the intermediate model with the highest footprint comparison accuracy is taken as the model to be finally stored.
(III) advantageous effects
The invention provides an identity authentication system based on a footprint comparison algorithm. The method has the following beneficial effects:
the identity authentication system based on the footprint comparison algorithm can help the specific places of warehouses, banks and prisons to realize the automation of the identity authentication of the entrance and exit; a footprint is one of the biological features that can characterize a person's identity information; the acquisition device is arranged at a necessary position of an entrance and an exit of a relevant place without being matched by an acquisition person intentionally in the acquisition process of the footprints, and the footprints can be acquired in a concealed manner by being disguised as normal ground; by adopting the identity authentication system adopting the footprint comparison algorithm, internal personnel in related places only need to collect footprint data once during information registration, and then can freely come in and go out, and identity authentication work is completed by the footprint identity authentication system laid at an entrance and an exit; the problems that the safety protection capability of the security system in the related places of the warehouse, the bank and the prison is not strong enough and the operation process is troublesome at present are solved.
Drawings
FIG. 1 is a schematic structural diagram of the present invention.
FIG. 2 is a schematic view of the footprint image pre-processing process of the present invention.
FIG. 3 is a footprint comparison network diagram of the present invention.
FIG. 4 is a flowchart of footprint identity authentication of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example 1
As shown in fig. 1 to 4, the present invention provides a technical solution: an identity authentication system based on footprint comparison algorithm comprises a supervision area, a gate channel, a footprint collector and a footprint authentication system, wherein a gate channel is respectively arranged in an entrance channel and an exit channel of the supervision area, and the footprint collector is arranged at an entrance end of the gate channel; the gate channel and the footprint collector are in communication connection with the footprint authentication system.
In this embodiment, the backbone network of the footprint authentication system is VGG 11; firstly, splicing two footprint images to be compared in the depth direction, inputting the two footprint images into VGG11 to extract comparison features, inputting a 1000-dimensional vector output from the last layer of VGG11 as a comparison feature vector into a Similarity Score Computing Network (SSCN) designed by the invention, and calculating the Similarity between the two footprints; if the similarity is larger than the set threshold value TH, the two footprints can be considered to belong to the same person, otherwise, the two footprints can be considered not to belong to the same person. In the footprint comparison network, the similarity fraction calculation network has two fully-connected layers, the fully-connected layer of the first layer receives a 1000-dimensional comparison feature vector output by VGG11, the dimension is reduced to 30 dimensions, a BN (batch normalization) layer and a ReLU activation function layer are connected behind the fully-connected layer of the first layer, the fully-connected layer of the second layer receives the 30-dimensional feature vector output by the first layer, a one-dimensional vector is output as a predicted similarity value, the similarity value cannot be directly used as a final similarity prediction result, and a Sigmoid activation function is required to be used for converting the similarity value into a value between 0 and 1 as a final prediction result.
A footprint alignment algorithm comprising the steps of:
step 1: identity acquisition: collecting information of workers, wherein three numbers 1, 2 and 3 represent identity information of three persons respectively, and when the information is registered, a plurality of footprint images of the three persons, and other information such as identity card numbers, job numbers and the like are collected, wherein the footprint images of each person and the respective personal information are corresponding;
step 2: acquiring and preprocessing the footprint image and archiving information: after the footprint image of the registrant is preprocessed, irrelevant information on the image is removed, the image resolution is reduced to the extent that the image resolution can be stored in a larger scale without influencing the identity authentication precision, and after the preprocessing, the image acquired during the registration is filed and stored in a database;
and step 3: and (3) information comparison: when people enter the field, the people will collect their footprint images, then the images of the people and the images in the database form a comparison image pair, and the comparison image pair is sent to a footprint authentication system, and the footprint authentication system outputs the similarity scores between the images through the comparison of the images; assuming that the threshold value TH set by the system is 0.7, the similarity between the footprints of the persons who perform identity authentication and the data of the footprints of the persons who perform identity authentication is 0.8 > TH, and the person is the person No. 1 in the database; the comparison is passed, and the comparison can be released; if the similarity between any footprint image and the image to be authenticated is not larger than the threshold value in the database, the footprint authentication system considers that the person does not carry out information registration and does not pass.
In this embodiment, the step 2: acquiring and preprocessing a footprint image, archiving information, and when the footprint is acquired: the collected person walks through the footprint collector, and the image collected by the footprint collector is a gray image with 256 gray levels;
preprocessing a footprint image: the purpose of preprocessing the image is mainly to remove noise on the image and scale information carried by a collecting instrument so as to facilitate image comparison, and reduce the size of the image so as to facilitate archiving.
In this embodiment, the step 2: the footprint image acquisition preprocessing and information archiving method comprises the following steps:
a. removing the scale information on the footprint image, wherein the image fixed area acquired by the footprint collector is provided with scale information for calculating the size of the footprint, and the information has little use for footprint comparison, so that the information needs to be removed, and the specific operation is to directly assign the gray value of the image fixed area to 255;
b. denoising and footprint positioning are carried out based on an Otsu algorithm, and a distinguishing threshold th is determined based on the Otsu algorithm because a footprint area and a background area of the collected footprint image are easy to distinguish; the pixel points with the gray scale smaller than th in the image are regarded as points of the footprint area, the other points are regarded as backgrounds, and the footprint image can be scanned from four directions, namely the upper direction, the lower direction, the left direction and the right direction after the th is determined, four boundaries of the footprint area are detected, and therefore footprint positioning is completed; after the footprint positioning is completed, because most of the noise is outside the footprint area, the denoising work is actually completed at the moment;
c. adjusting the size of the footprint image and carrying out gray inversion, determining the position of the footprint area in the step b, wherein the distance between the upper boundary and the lower boundary of the footprint is larger than the distance between the left boundary and the right boundary under the general condition, so that the system considers that the footprint image with the distance between the upper boundary and the lower boundary smaller than the distance between the left boundary and the right boundary is an abnormal footprint image, and an alarm is sent out in the specific implementation of the system; after the distance between the upper boundary and the lower boundary is determined, a square area with the pixel value of 255 is defined, and the side length is the distance between the upper boundary and the lower boundary; then directly copying the footprint area to the midpoint of the square area, and subtracting the pixel value of each point of the square area from 255 to perform gray level inversion at the pixel level; and finally, downsampling the obtained square area, adjusting the resolution and storing.
In this embodiment, the step 3: the information comparison comprises the following steps:
step one, network training: dividing the collected footprint sample into two parts according to the collected object, wherein the footprint sample of one part of the object is used as a training set sample, and the other part of the object is used as a verification set sample;
if two footprint images input into the network belong to the same person, the two images form a positive sample X with a label Y of 1, otherwise, the two images form a negative sample X with a label Y of 0; randomly sampling from a training set each time to obtain a Batch of (Batch) positive samples or negative samples, inputting the Batch of (Batch) positive samples or negative samples into a network for similarity calculation, and making an MSE loss on an output value and a label, wherein the loss expression is as follows:
in the formula, B represents the batch size during training, FCN represents a footprint comparison network, and any sample X is input into the network to obtain a prediction label of the network;
during training, performing back propagation on the loss L of each batch of samples, and then optimizing the model by using a gradient descent method; when the training batches are enough, the network is equivalent to parameter optimization on all possible positive and negative sample spaces; through repeated random extraction of positive and negative samples for model optimization, the model is gradually converged finally;
secondly, selecting a model: the model to be finally stored in the neural network generally needs to be selected from the trained intermediate models by using a verification set, enough samples need to be randomly selected from the verification set during model selection, the samples are input into the network for similarity calculation, if the similarity is more than 0.5, the model is regarded as the footprint of the same person, if the similarity is less than 0.5, the model is not the footprint of the same person, and the intermediate model with the highest footprint comparison accuracy is taken as the model to be finally stored.
The working principle is as follows:
when the method is used, firstly, information is collected for workers, the collected people walk through a footprint collector, an image collected by the footprint collector is a gray image with 256 gray levels, a fixed image area collected by the footprint collector is provided with scale information for calculating the size of a footprint, the information has little use for footprint comparison, and therefore needs to be removed, the specific operation is that the gray value of the fixed image area is directly assigned to be 255, and the footprint area and a background area of the collected footprint image are easily distinguished, so that a distinguishing threshold th is determined based on an Otsu algorithm; the pixel points with the gray scale smaller than th in the image are regarded as points of the footprint area, the other points are regarded as backgrounds, and the footprint image can be scanned from four directions, namely the upper direction, the lower direction, the left direction and the right direction after the th is determined, four boundaries of the footprint area are detected, and therefore footprint positioning is completed; after the footprint positioning is completed, because most of the noise is outside the footprint area, the denoising work is actually completed at the moment; adjusting the size of the footprint image, carrying out gray inversion, and defining a square area with the pixel value of 255 after determining the distance between the upper boundary and the lower boundary, wherein the side length is the distance between the upper boundary and the lower boundary; then directly copying the footprint area to the midpoint of the square area, and subtracting the pixel value of each point of the square area from 255 to perform gray level inversion at the pixel level; finally, downsampling the obtained square area, adjusting the resolution ratio, storing the square area, collecting footprint images of personnel when the personnel enter the field in the using process, forming comparison image pairs by the images of the personnel and the images in the database, sending the comparison image pairs into a footprint authentication system, and outputting similarity scores between the images by the footprint authentication system through comparison of the images; assuming that the threshold value TH set by the system is 0.7, the similarity between the footprints of the persons who perform identity authentication and the data of the footprints of the persons who perform identity authentication is 0.8 > TH, and the person is the person No. 1 in the database; the comparison is passed, and the comparison can be released; if the similarity between any footprint image and the image to be authenticated is not larger than the threshold value in the database, the footprint authentication system considers that the person does not carry out information registration and does not pass.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (5)
1. An identity authentication system based on footprint comparison algorithm is characterized in that: the identity authentication system based on the footprint comparison algorithm comprises a supervision area, gate channels, footprint collectors and a footprint authentication system, wherein one gate channel is respectively arranged in an entrance channel and an exit channel of the supervision area, and the footprint collectors are arranged at entrance ends of the gate channels; the gate channel and the footprint collector are in communication connection with the footprint authentication system.
2. A footprint comparison algorithm, comprising: the footprint comparison algorithm comprises the following steps:
step 1: identity acquisition: collecting information of workers, wherein three numbers 1, 2 and 3 represent identity information of three persons respectively, and when the information is registered, a plurality of footprint images of the three persons, and other information such as identity card numbers, job numbers and the like are collected, wherein the footprint images of each person and the respective personal information are corresponding;
step 2: acquiring and preprocessing the footprint image and archiving information: after the footprint image of the registrant is preprocessed, irrelevant information on the image is removed, the image resolution is reduced to the extent that the image resolution can be stored in a larger scale without influencing the identity authentication precision, and after the preprocessing, the image acquired during the registration is filed and stored in a database;
and step 3: and (3) information comparison: when people enter the field, the people will collect their footprint images, then the images of the people and the images in the database form a comparison image pair, and the comparison image pair is sent to a footprint authentication system, and the footprint authentication system outputs the similarity scores between the images through the comparison of the images; assuming that the threshold value TH set by the system is 0.7, the similarity between the footprints of the persons who perform identity authentication and the data of the footprints of the persons who perform identity authentication is 0.8 > TH, and the person is the person No. 1 in the database; the comparison is passed, and the comparison can be released; if the similarity between any footprint image and the image to be authenticated is not larger than the threshold value in the database, the footprint authentication system considers that the person does not carry out information registration and does not pass.
3. The footprint comparison algorithm according to claim 2, wherein: the step 2: acquiring and preprocessing a footprint image, archiving information, and when the footprint is acquired: the collected person walks through the footprint collector, and the image collected by the footprint collector is a gray image with 256 gray levels;
preprocessing a footprint image: the purpose of preprocessing the image is mainly to remove noise on the image and scale information carried by a collecting instrument so as to facilitate image comparison, and reduce the size of the image so as to facilitate archiving.
4. The footprint comparison algorithm according to claim 2, wherein: the step 2: the footprint image acquisition preprocessing and information archiving method comprises the following steps:
a. removing the scale information on the footprint image, wherein the image fixed area acquired by the footprint collector is provided with scale information for calculating the size of the footprint, and the information has little use for footprint comparison, so that the information needs to be removed, and the specific operation is to directly assign the gray value of the image fixed area to 255;
b. denoising and footprint positioning are carried out based on an Otsu algorithm, and a distinguishing threshold th is determined based on the Otsu algorithm because a footprint area and a background area of the collected footprint image are easy to distinguish; the pixel points with the gray scale smaller than th in the image are regarded as points of the footprint area, the other points are regarded as backgrounds, and the footprint image can be scanned from four directions, namely the upper direction, the lower direction, the left direction and the right direction after the th is determined, four boundaries of the footprint area are detected, and therefore footprint positioning is completed; after the footprint positioning is completed, because most of the noise is outside the footprint area, the denoising work is actually completed at the moment;
c. adjusting the size of the footprint image and carrying out gray inversion, determining the position of the footprint area in the step b, wherein the distance between the upper boundary and the lower boundary of the footprint is larger than the distance between the left boundary and the right boundary under the general condition, so that the system considers that the footprint image with the distance between the upper boundary and the lower boundary smaller than the distance between the left boundary and the right boundary is an abnormal footprint image, and an alarm is sent out in the specific implementation of the system; after the distance between the upper boundary and the lower boundary is determined, a square area with the pixel value of 255 is defined, and the side length is the distance between the upper boundary and the lower boundary; then directly copying the footprint area to the midpoint of the square area, and subtracting the pixel value of each point of the square area from 255 to perform gray level inversion at the pixel level; and finally, downsampling the obtained square area, adjusting the resolution and storing.
5. The footprint comparison algorithm according to claim 2, wherein: the step 3: the information comparison comprises the following steps:
step one, network training: dividing the collected footprint sample into two parts according to the collected object, wherein the footprint sample of one part of the object is used as a training set sample, and the other part of the object is used as a verification set sample;
if two footprint images input into the network belong to the same person, the two images form a positive sample X with a label Y of 1, otherwise, the two images form a negative sample X with a label Y of 0; randomly sampling from a training set each time to obtain a Batch of (Batch) positive samples or negative samples, inputting the Batch of (Batch) positive samples or negative samples into a network for similarity calculation, and making an MSE loss on an output value and a label, wherein the loss expression is as follows:
in the formula, B represents the batch size during training, FCN represents a footprint comparison network, and any sample X is input into the network to obtain a prediction label of the network;
during training, performing back propagation on the loss L of each batch of samples, and then optimizing the model by using a gradient descent method; when the training batches are enough, the network is equivalent to parameter optimization on all possible positive and negative sample spaces; through repeated random extraction of positive and negative samples for model optimization, the model is gradually converged finally;
secondly, selecting a model: the model to be finally stored in the neural network generally needs to be selected from the trained intermediate models by using a verification set, enough samples need to be randomly selected from the verification set during model selection, the samples are input into the network for similarity calculation, if the similarity is more than 0.5, the model is regarded as the footprint of the same person, if the similarity is less than 0.5, the model is not the footprint of the same person, and the intermediate model with the highest footprint comparison accuracy is taken as the model to be finally stored.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010912788.9A CN112069482B (en) | 2020-09-03 | 2020-09-03 | Identity authentication system based on footprint comparison method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010912788.9A CN112069482B (en) | 2020-09-03 | 2020-09-03 | Identity authentication system based on footprint comparison method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112069482A true CN112069482A (en) | 2020-12-11 |
| CN112069482B CN112069482B (en) | 2023-04-07 |
Family
ID=73666392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010912788.9A Active CN112069482B (en) | 2020-09-03 | 2020-09-03 | Identity authentication system based on footprint comparison method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112069482B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010261159A (en) * | 2009-04-30 | 2010-11-18 | Yamatake Corp | System and method for room entry-exit control |
| CN103985175A (en) * | 2014-05-21 | 2014-08-13 | 清华大学 | Control method and system of home entrance guard |
| CN205091816U (en) * | 2015-11-06 | 2016-03-16 | 天津大学 | Personnel management system based on identity information |
| CN111382629A (en) * | 2018-12-28 | 2020-07-07 | 中国科学院半导体研究所 | Footprint identification and information mining method and system based on convolutional neural network |
| CN111402202A (en) * | 2020-02-24 | 2020-07-10 | 安徽大学 | Similarity evaluation method for plantar pressure images |
-
2020
- 2020-09-03 CN CN202010912788.9A patent/CN112069482B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010261159A (en) * | 2009-04-30 | 2010-11-18 | Yamatake Corp | System and method for room entry-exit control |
| CN103985175A (en) * | 2014-05-21 | 2014-08-13 | 清华大学 | Control method and system of home entrance guard |
| CN205091816U (en) * | 2015-11-06 | 2016-03-16 | 天津大学 | Personnel management system based on identity information |
| CN111382629A (en) * | 2018-12-28 | 2020-07-07 | 中国科学院半导体研究所 | Footprint identification and information mining method and system based on convolutional neural network |
| CN111402202A (en) * | 2020-02-24 | 2020-07-10 | 安徽大学 | Similarity evaluation method for plantar pressure images |
Non-Patent Citations (1)
| Title |
|---|
| 陈扬等: "一种基于CNN的足迹图像检索与匹配方法", 《南京师范大学学报(工程技术版)》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112069482B (en) | 2023-04-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102629319B (en) | Robust iris region segmentation method based on specific boundary detectors | |
| CN100382093C (en) | Registration method for biometrics authentication system, biometrics authentication system, and program for same | |
| US20180089187A1 (en) | Image recognition method and image recognition apparatus | |
| CN104700094B (en) | A kind of face identification method for intelligent robot and system | |
| CN104992148A (en) | ATM terminal human face key points partially shielding detection method based on random forest | |
| CN103049736A (en) | Face identification method based on maximum stable extremum area | |
| CN113269070B (en) | Pedestrian re-identification method fusing global and local features, memory and processor | |
| CN113221086B (en) | Offline face authentication method and device, electronic equipment and storage medium | |
| US9378406B2 (en) | System for estimating gender from fingerprints | |
| Samuel et al. | Novel feature extraction technique for off-line signature verification system | |
| Hendryli et al. | Automatic license plate recognition for parking system using convolutional neural networks | |
| CN114037838A (en) | Neural network training method, electronic device and computer program product | |
| Ojo et al. | Development of an improved convolutional neural network for an automated face based university attendance system | |
| CN111488798B (en) | Fingerprint identification method, fingerprint identification device, electronic equipment and storage medium | |
| Gavisiddappa et al. | Multimodal biometric authentication system using modified ReliefF feature selection and multi support vector machine | |
| CN105184236A (en) | Robot-based face identification system | |
| CN114840748A (en) | Information pushing method, device and equipment based on face recognition and storage medium | |
| CN112069482B (en) | Identity authentication system based on footprint comparison method | |
| Kumar | Signature verification using neural network | |
| WO2020232889A1 (en) | Check encashment method, apparatus and device, and computer-readable storage medium | |
| Verma et al. | Static Signature Recognition System for User Authentication Based Two Level Cog, Hough Tranform and Neural Network | |
| CN114445862A (en) | Attendance detection method and system based on offline classroom monitoring | |
| CN113591619A (en) | Face recognition verification device based on video and verification method thereof | |
| CN113689613A (en) | Access control system, access control method, and storage medium | |
| CN112182537A (en) | Monitoring method, device, server, system and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |