CN202472696U - Identity verification system based on second-generation identity card and face feature identification - Google Patents
Identity verification system based on second-generation identity card and face feature identification Download PDFInfo
- Publication number
- CN202472696U CN202472696U CN2011204616137U CN201120461613U CN202472696U CN 202472696 U CN202472696 U CN 202472696U CN 2011204616137 U CN2011204616137 U CN 2011204616137U CN 201120461613 U CN201120461613 U CN 201120461613U CN 202472696 U CN202472696 U CN 202472696U
- Authority
- CN
- China
- Prior art keywords
- unit
- identity
- identification
- verification system
- identity card
- 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.)
- Expired - Lifetime
Links
Images
Landscapes
- Collating Specific Patterns (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The utility model relates to an identity verification system based on second-generation identity card and face feature identification, comprising a camera, a computer host, and an identity card reader; wherein, the camera and the identity card reader are respectively connected with the computer host; and the computer host comprises a camera shooting unit, a face detection and location unit, an image standardization unit, a wavelet transformation unit, a feature vector extraction unit, and a support vector machine unit. The identity verification system of the utility model has the beneficial effects that: a dual identity verification comprising second-generation identity card verification and face feature comparison and identification is realized; by utilizing the face features to perform identity identification, there is no need for any other forms of passwords, therefore, the way for utilizing a face identification technology to perform identity identification is safer, more reliable, and more convenient; the identity verification system requires low hardware condition, and there is no overmuch complex computing and a large number of data storage; the identity verification system has a quick comparison speed, thus the identity verification system can be used for most embedded equipments; and the identity verification system also has a high identification rate which can reach more than 90% in a certain environment.
Description
Technical field
The utility model relates to a kind of authentication system based on China second-generation identity card and face characteristic identification, belongs to the identification electronic distinguishment technical field.
Background technology
At present, the China second-generation identity card verification facility have been widely used in social all trades and professions such as public security, finance, communication, the tax, logistics, examination, housekeeping, realize the identity veritification through the checking to Certification of Second Generation.This mode also becomes every profession and trade and carries out the important means that identity is veritified.
But, if Certification of Second Generation is lost or by being forged, authentication there is the field of strict demand for public security, finance, examination etc., only rely on I.D. identification identity, the behavior that other people I.D. carries out illegal activity inevitably can appear usurping.Chinese patent file CN 1936930 discloses a kind of method of discerning second generation resident identification card; Not enough below but this patent exists: 1) I.D. card reader, picture pick-up device, personal identification card document storehouse are common apparatus, only have these equipment to be not enough to constitute face identification system; 2) do not record the concrete grammar and the technology path of recognition of face device in this patent document.
Summary of the invention
To the deficiency of prior art, the utility model relates to a kind of identification authentication system based on China second-generation identity card and face characteristic.
The technical scheme of the utility model is following:
A kind of authentication system based on China second-generation identity card and face characteristic identification comprises video camera, main frame and I.D. card reader, and described video camera links to each other with main frame respectively with the I.D. card reader; Described main frame comprises image unit, people's face detection and location unit, image standardization unit, wavelet transform unit, proper vector extraction unit and SVMs unit; Said image unit, people's face detection and location unit, image standardization unit, wavelet transform unit, proper vector extraction unit link to each other with the SVMs sequence of unit; Said video camera links to each other with the image unit of main frame, and said I.D. card reader links to each other with people's face detection and location unit.
Said image unit, to the video camera transmitting control commands, the control video camera grasps the holder scene photograph and preserves;
Said people's face detection and location unit detects, locatees human face region respectively from identity card picture and scene photograph;
The image standardization unit; Pupil of both eyes in the human face region is positioned; With the distance setting between the pupil of both eyes in the human face region is 64 pixels; In identity card picture and scene photograph, be picture centre respectively with pupil of both eyes line center, the human face photo of intercepting 240x320 pixel respectively, the standardization human face photo, the standardization human face photo of scene photograph of I.D.; It is to adopt the image standardization unit to search for pupil location or the artificial pupil location of clicking automatically that pupil of both eyes in the human face region is positioned;
Wavelet transform unit; The standardization human face photo is carried out two-dimensional discrete wavelet conversion: the standardization human face photo is four sub-frequency bands by wavelet decomposition; Obtain ground floor wavelet decomposition frequency band figure, the LL sub-band is a low frequency sub-band, and it is that the low dimension of standardized human face photo is approximate; The LH subband has been described the horizontal direction high frequency characteristics of standardized human face photo; The HL subband has been described the vertical direction high frequency characteristics of standardized human face photo; The HH subband has been described the characteristic of standardized human face photo diagonal; Wherein low frequency sub-band LL continuation is four sub-frequency bands by wavelet decomposition, obtains second layer wavelet decomposition frequency band figure;
The proper vector extraction unit is used to extract the proper vector of standardization human face photo; After 2 wavelet decomposition of standardization human face photo for the 240x320 pixel, each frequency band size is 240/2
k* 320/2
k, cutting apart each sub-band with the grid regions of m * m is many sub-block, calculates 6 proper vectors of every sub-block then, is the proper vector of standardization human face photo:
The horizontal direction variable:
The vertical direction variable:
45 ° of direction variables:
135 ° of direction variables:
The variance of every sub-block:
The mean value of every sub-block and its transposition: V
r=1/m * m (E (AA
T))
Wherein, (x is that (x, the y) wavelet coefficient of position, A represent sub-piece wavelet coefficient matrix, and m is the size of grid, preferred m=5, V in every sub-block y) to coef
h, V
v, V
45, V
135Be respectively level, vertically, 45 °, 135 ° direction vector, V
bBe variance, V
rBe the mean value of the transposition product of A and A, total proper vector dimension is 6/m
2* 240/2
k* 320/2
k
The SVMs unit, with the proper vector of identity card picture that holder is held as positive sample, with the proper vector of scene photograph as sample to be tested.In addition; In the SVMs unit, store the identity card picture that has ID card No. of some; Store the identity card picture more than 5000 or 5000 in the SVMs unit of the utility model; Therefrom the photo 50 or 50 or more of picked at random except that holding identity card picture is as negative sample (through checking ID card No., holding identity card picture is got rid of from negative sample, held identity card picture to some extent to guarantee not contain in the negative sample).The classification lineoid is sought automatically according to positive sample and negative sample in the SVMs unit, when the proper vector of scene photograph is in classification lineoid negative sample side, judges that then holder identity and ID card information are inconsistent; When the proper vector of scene photograph is in the positive sample side of classification lineoid, judge that then the holder identity is consistent with ID card information.
The method of work of the utility model is following:
1) control I.D. card reader reading identity card information transfers to main frame;
2) scene photograph of control video camera extracting holder transfers to computing machine;
3) locate the human face region of identity card picture and scene photograph respectively;
4) obtain the standardization human face photo of identity card picture and scene photograph respectively;
5) proper vector of the standardization human face photo of extraction identity card picture; Extract the proper vector of the standardization human face photo of scene photograph;
6) with the proper vector of identity card picture as positive sample; With the proper vector of scene photograph as sample to be tested; In the SVMs unit, store negative sample; The classification lineoid is sought automatically according to positive sample and negative sample in the SVMs unit, when the proper vector of scene photograph is in classification lineoid negative sample side, judges that then holder identity and ID card information are inconsistent; When the proper vector of scene photograph is in the positive sample side of classification lineoid, judge that then the holder identity is consistent with ID card information;
7) if the judged result of approval step 6) gained then finishes this authentication; If do not admit the judged result of step 6) gained, then return step 2).
The advantage of the utility model is:
1. dual veritification: the utility model combines China second-generation identity card verification technology and face recognition technology, has realized that the dual identity of Certification of Second Generation checking and face characteristic comparison identification is veritified.
2. safe ready: individual face characteristic be have uniqueness, constant stability in the regular period, extremely difficult the forgery and personation.Utilize face characteristic to carry out identification, need not the password of any other form, so, utilize face recognition technology to carry out the identity identification, safer, reliable, convenient.
3. the utility model is low to hardware requirement, does not have the storage of too much complex calculation and lot of data; Versus speed is fast, can on most embedded devices, use; Discrimination is high, and discrimination can reach more than 90% under certain environment.
4. the utility model is treated to the standardization human face photo that pixel is 240x320 to identity card picture and scene photograph at first respectively, the also standardization of people's eye coordinates, thus underwriter's face geometric properties can effectively align.Utilize wavelet transformation to separate half-tone information and profile information then, and be converted into 180 dimensional feature vectors through proper transformation.The benefit of wavelet transformation is, separates half-tone information, can further eliminate the adverse effect of illumination.
5. the utility model utilizes algorithm of support vector machine, according to photo characteristic of correspondence vector, obtain these photos 180 dimensional feature space corresponding characteristic area.If identity card picture characteristic of correspondence vector drops in this zone, promptly the decidable holder is consistent with I.D..
Description of drawings
Fig. 1 is that the hardware of the utility model connects synoptic diagram; Wherein, 1, video camera; 3, I.D. card reader; 4, image unit; 5, people's face detection and location unit; 6, image standardization unit; 7, wavelet transform unit; 8, proper vector extraction unit; 9, SVMs unit.
Fig. 2 is the workflow diagram of the utility model;
Fig. 3 is the wavelet transformation synoptic diagram;
Fig. 4 is the wavelet reconstruction synoptic diagram;
Fig. 5 is the wavelet decomposition frequency band synoptic diagram;
Fig. 6 is the proper vector synoptic diagram of a frequency band in the standardization human face photo;
Fig. 7 judges the schematic diagram that holder identity and ID card information be whether consistent for utilizing the classification lineoid.
Embodiment
Below in conjunction with embodiment and Figure of description the utility model is done detailed explanation, but be not limited thereto.
A kind of authentication system based on China second-generation identity card and face characteristic identification comprises video camera 1, main frame and I.D. card reader 3, and described video camera 1 links to each other with main frame respectively with I.D. card reader 3; Described main frame comprises image unit 4, people's face detection and location unit 5, image standardization unit 6, wavelet transform unit 7, proper vector extraction unit 8 and SVMs unit 9; Said image unit 4, people's face detection and location unit 5, image standardization unit 6; Wavelet transform unit 7, proper vector extraction unit 8 and SVMs unit 9 orders link to each other; Said video camera 1 links to each other with the image unit 4 of main frame, and said I.D. card reader 3 links to each other with people's face detection and location unit 5.
I.D. card reader reading identity card information also transmits it to main frame, comprises identity card picture in the said ID card information;
Said image unit, to the video camera transmitting control commands, the control video camera grasps the holder scene photograph and preserves;
Said people's face detection and location unit detects, locatees human face region respectively from identity card picture and scene photograph;
Said image standardization unit; Pupil of both eyes in the human face region is positioned; With the distance setting between the pupil of both eyes in the human face region is 64 pixels; In identity card picture and scene photograph, be picture centre respectively with pupil of both eyes line center, the human face photo of intercepting 240x320 pixel respectively, the standardization human face photo, the standardization human face photo of scene photograph of I.D.; It is to adopt the image standardization unit to search for pupil location or the artificial pupil location of clicking automatically that pupil of both eyes in the human face region is positioned;
For simplifying subsequent calculations all standardization of the size of detected human face region, brightness, contrast, step is following:
1, confirm the pupil of both eyes centre coordinate, the image standardization unit is searched for the pupil location automatically
(1) neural network training detects the approximate range of eyes in identification identity card picture and the scene photograph;
(2) utilize the image integration algorithm to confirm the pupil of both eyes accurate coordinates;
2, face-image intercepting
(1) distance is 64 pixels between the setting pupil of both eyes, calculates the scaling ratio of identity card picture and scene photograph;
(2) respectively in identity card picture and scene photograph, be standardization human face photo center with pupil of both eyes line center, the human face photo of intercepting 240x320 pixel;
(3), make the photo grey scaleization to get the standardization human face photo to the human face photo luminance histogram equalizationization.
For the detection test shows of front face, detect accuracy rate under the conventional situation and be higher than 98%.
Said wavelet transform unit; As shown in Figure 5, the standardization human face photo is carried out two-dimensional discrete wavelet conversion: the standardization human face photo is four sub-frequency bands by wavelet decomposition, obtains ground floor wavelet decomposition frequency band figure; The LL sub-band is a low frequency sub-band, and it is that the low dimension of standardized human face photo is approximate; The LH sub-band has been described the horizontal direction high frequency characteristics of standardized human face photo; The HL subband has been described the vertical direction high frequency characteristics of standardized human face photo; The HH subband has been described the characteristic of standardized human face photo diagonal; Wherein low frequency sub-band LL continuation is four sub-frequency bands by wavelet decomposition, obtains second layer wavelet decomposition frequency band figure;
Utilize wavelet transformation that the standardization human face photo is carried out pre-service, make face characteristic more outstanding.The utility model carries out two-dimensional discrete wavelet conversion to the standardization human face photo, and 2-d discrete wavelet function and scaling function obtain through tensor transformation through one dimension wavelet function and scaling function.The basic decomposition and reconstruction step of 2-d wavelet can utilize Fig. 3 and Fig. 4 to represent respectively.
Said proper vector extraction unit is used to extract the proper vector of standardization human face photo; After 2 wavelet decomposition of standardization human face photo for M * N pixel, each frequency band size is M/2
k* N/2
k, M=240 wherein, N=320, cutting apart each sub-band with the grid regions of m * m is many sub-block, m=5 wherein calculates 6 proper vectors of every sub-block then, is the proper vector of standardization human face photo:
The horizontal direction variable:
The vertical direction variable:
45 ° of direction variables:
135 ° of direction variables:
The variance of every sub-block:
The mean value of every sub-block and its transposition: V
r=1/m * m (E (AA
T))
Wherein, (x is that (x, the y) wavelet coefficient of position, A represent sub-piece wavelet coefficient matrix, and m is the size of grid, V in every sub-block y) to coef
h, V
v, V
45, V
135Be respectively level, vertically, 45 °, 135 ° direction vector, V
bBe variance, V
rBe the mean value of the transposition product of A and A, total proper vector dimension is 6/m
2* M/2
k* N/2
kBe illustrated in figure 6 as the proper vector figure of a frequency band, wherein horizontal ordinate is the dimension sequence number that characteristic is rung amount, and ordinate is the component value of proper vector in this dimension.
The SVMs unit; With the proper vector of identity card picture that holder is held as positive sample; The proper vector of scene photograph as sample to be tested, is stored negative sample in the SVMs unit, the classification lineoid is sought automatically according to positive sample and negative sample in the SVMs unit; When the proper vector of scene photograph is in classification lineoid negative sample side, judge that then holder identity and ID card information are inconsistent; When the proper vector of scene photograph is in the positive sample side of classification lineoid, judge that then the holder identity is consistent with ID card information.
Set up face characteristic data training set, make up supporting vector machine model
The utility model adopts SVMs, and (Support Vector Machine, SVM) method is carried out final consistance differentiation.SVM requires the training sample set of certain scale to support.The utility model adopts the positive sample of proper vector conduct of the holder scene photograph of shooting, in the SVMs unit, stores negative sample.
SVMs (Support Vector Machine; SVM) be a kind of learning method that supervision is arranged; SVMs is regarded positive sample and negative sample as two set at 2 dimension spaces; Automatically seek the classification lineoid then this 2 dimension space is divided into two parts, make positive sample and negative sample drop on respectively in two different semispaces, guarantee that simultaneously two intervals between the set are maximum.
So-called maximum at interval is meant that the distance of gathering between two tangent parallel lineoid with positive and negative samples is maximum.As shown in Figure 7, lineoid A corresponding intervals 2da is just big than lineoid B corresponding intervals 2db, and wherein lineoid A is exactly the classification lineoid that searches out automatically.
Represent two components of proper vector respectively by horizontal ordinate among Fig. 7 and ordinate, visible, have only those decisions a few data at interval to what the classification lineoid played a decisive role, claim that these determinative data are support vector.These vectors have determined size at interval, have also determined final classification lineoid.For the sample to be tested of new input, promptly the proper vector of the standardization human face photo of scene photograph is in the positive sample side or the negative sample side of classification lineoid through judging it, and then judges that sample to be tested is positive sample or negative sample.
Claims (1)
1. the authentication system based on China second-generation identity card and face characteristic identification is characterized in that this system comprises video camera, main frame and I.D. card reader, and described video camera links to each other with main frame respectively with the I.D. card reader; Described main frame comprises image unit, people's face detection and location unit, image standardization unit, wavelet transform unit, proper vector extraction unit and SVMs unit; Said image unit, people's face detection and location unit, image standardization unit; Wavelet transform unit, proper vector extraction unit and SVMs sequence of unit link to each other; Said video camera links to each other with the image unit of main frame, and said I.D. card reader links to each other with people's face detection and location unit.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011204616137U CN202472696U (en) | 2011-11-18 | 2011-11-18 | Identity verification system based on second-generation identity card and face feature identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011204616137U CN202472696U (en) | 2011-11-18 | 2011-11-18 | Identity verification system based on second-generation identity card and face feature identification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN202472696U true CN202472696U (en) | 2012-10-03 |
Family
ID=46920840
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011204616137U Expired - Lifetime CN202472696U (en) | 2011-11-18 | 2011-11-18 | Identity verification system based on second-generation identity card and face feature identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN202472696U (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104517137A (en) * | 2013-10-06 | 2015-04-15 | 李策 | Method for preventing stealing of identification card information |
| CN109564620A (en) * | 2016-06-03 | 2019-04-02 | 奇跃公司 | Augmented reality authentication |
-
2011
- 2011-11-18 CN CN2011204616137U patent/CN202472696U/en not_active Expired - Lifetime
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104517137A (en) * | 2013-10-06 | 2015-04-15 | 李策 | Method for preventing stealing of identification card information |
| CN109564620A (en) * | 2016-06-03 | 2019-04-02 | 奇跃公司 | Augmented reality authentication |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102509138B (en) | Authentication system based on second-generation ID card and human face feature recognition and working method therefor | |
| CN107292287B (en) | Face identification method, device, electronic equipment and storage medium | |
| WO2020151489A1 (en) | Living body detection method based on facial recognition, and electronic device and storage medium | |
| CN101558431B (en) | Face authentication device | |
| CN103116763B (en) | A kind of living body faces detection method based on hsv color Spatial Statistical Character | |
| WO2019033572A1 (en) | Method for detecting whether face is blocked, device and storage medium | |
| CN102254188B (en) | Palmprint recognizing method and device | |
| CN102629320B (en) | Ordinal measurement statistical description face recognition method based on feature level | |
| CN106250821A (en) | The face identification method that a kind of cluster is classified again | |
| CN105956578A (en) | Face verification method based on identity document information | |
| CN103605958A (en) | Living body human face detection method based on gray scale symbiosis matrixes and wavelet analysis | |
| CN104700094B (en) | A kind of face identification method for intelligent robot and system | |
| CN103761507B (en) | Local multi-value pattern face recognition method based on Weber law | |
| CN106446772A (en) | Cheating-prevention method in face recognition system | |
| CN103530648A (en) | Face recognition method based on multi-frame images | |
| CN108564040B (en) | Fingerprint activity detection method based on deep convolution characteristics | |
| CN109858375A (en) | Living body faces detection method, terminal and computer readable storage medium | |
| CN111695099A (en) | Security system and corresponding access method, machine readable storage device | |
| CN104751143A (en) | Person and credential verification system and method based on deep learning | |
| CN107273783A (en) | Face identification system and its method | |
| CN113111880B (en) | Certificate image correction method, device, electronic equipment and storage medium | |
| CN104346883A (en) | Point of sale (POS) device capable of detecting customer | |
| CN106682589A (en) | Face recognition and prison roll call system | |
| CN106709458A (en) | Human face living body detection method and device | |
| CN202472696U (en) | Identity verification system based on second-generation identity card and face feature identification |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C56 | Change in the name or address of the patentee | ||
| CP01 | Change in the name or title of a patent holder |
Address after: Shun high tech Zone of Ji'nan City, Shandong province 250101 China West Road No. 699 Patentee after: SYNTHESIS ELECTRONIC TECHNOLOGY CO., LTD. Address before: Shun high tech Zone of Ji'nan City, Shandong province 250101 China West Road No. 699 Patentee before: Shandong Synthesis Electronic Technology Co., Ltd. |
|
| CX01 | Expiry of patent term | ||
| CX01 | Expiry of patent term |
Granted publication date: 20121003 |