CN109409314A - A kind of finger vein identification method and system based on enhancing network - Google Patents
A kind of finger vein identification method and system based on enhancing network Download PDFInfo
- Publication number
- CN109409314A CN109409314A CN201811318592.6A CN201811318592A CN109409314A CN 109409314 A CN109409314 A CN 109409314A CN 201811318592 A CN201811318592 A CN 201811318592A CN 109409314 A CN109409314 A CN 109409314A
- Authority
- CN
- China
- Prior art keywords
- image
- network
- quality
- identification
- finger vein
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
Landscapes
- Engineering & Computer Science (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a kind of finger vein identification methods and system based on enhancing network, belong to biometrics identification technology field.Finger vein identification method based on enhancing network of the invention, picture quality is evaluated, separate high quality graphic and low-quality image, low-quality image is enhanced using enhancing network, low-quality image is converted into high quality graphic, identification is carried out by high quality graphic of the identification network to output.The finger vein identification method based on enhancing network of the invention improves picture quality while identifying image, to help to improve accuracy of identification, has good application value.
Description
Technical field
The present invention relates to biometrics identification technology fields, specifically provide a kind of finger vena identification based on enhancing network
Method and system.
Background technique
With the development of the society, every field technology has significant progress.Biometrics identification technology is to utilize people
Body biological characteristic or behavioural characteristic carry out human body authentication, wherein human body biological characteristics mainly include two major classes: outside is raw
Object feature and Internal biological feature.External biological characteristic such as fingerprint, iris view and shape of face etc..Internal biological feature such as finger is quiet
Arteries and veins, hand back vein and palm vein etc..In external biological characteristic, due to fingerprint recognition have very strong uniqueness, stability,
The features such as ease for use, application are extremely wide.But in fingerprint recognition, must be requested that user keeps hand in typing fingerprint
Finger is clean, smooth, because any dirt being present on fingerprint or stain can bring difficulty to identification, and fingerprint is easy to
It forges, causes the safety coefficient of fingerprint identification technology relatively low, compared with external biological characteristic, Internal biological feature is difficult to steal
And forgery, therefore there is higher security performance, in internal biological characteristic, because finger vein features are with very strong universal
Property and uniqueness, therefore finger vena identification becomes biological identification technology is opened up in recent years frontier.
Finger vena identification usually obtains finger venous image using transmitted light or reflected light two ways, then from finger
Vein pattern is extracted in vein image, carries out characteristic matching, to realize the identification of finger venous image, confirms user's body
Part.Since its vivo identification, internal identification, acquisition equipment cost be low etc., advantages have become a kind of great hair for finger vena identification
Open up the biometrics identification technology of potentiality.However, due to finger venous image acquisition when, due to illumination, finger rotate etc. because
Element influences, and can reduce finger vein image quality, thus causes to extract vein pattern progress feature from finger venous image
It goes wrong during matching, brings difficulty to the process of confirmation user's identity.
Summary of the invention
Technical assignment of the invention is in view of the above problems, to provide one kind and improve image while identifying image
Quality, to help to improve the finger vein identification method based on enhancing network of accuracy of identification.
Another technical assignment of the invention is to provide a kind of finger vein recognition system based on enhancing network.
To achieve the above object, the present invention provides the following technical scheme that
A kind of finger vein identification method based on enhancing network, finger vein image quality are evaluated, and high quality is separated
Image and low-quality image enhance low-quality image using enhancing network, low-quality image are converted into high-quality
Picture carries out identification by high quality graphic of the identification network to output.
For low-quality finger venous image, it is not easy to be identified.It is quiet by the finger based on enhancing network
Arteries and veins recognition methods finger vein image is evaluated, so that finger venous image to be identified is separated high quality graphic and low
Two class of quality image directly passes through high quality graphic on identification network and carries out identification.It is first for low-quality image
Enhancing network is first passed through, high quality graphic is converted into, identification is then carried out by identification network again.
The finger vein identification method based on enhancing network is compared to traditional finger vein identification method, is identifying
While improve finger vein image quality, the image of high quality facilitates the raising of accuracy of identification, to facilitate identity
Identification.
Preferably, this method includes training stage and cognitive phase, detailed process is as follows:
S1: acquisition finger venous image constructs training set；
S2: training image quality testing network is trained on training set, and finger vein image quality is detected, if
Step S4 is then carried out for high quality graphic, otherwise carries out step S3；
S3: low-quality image is converted into high quality graphic by building image enhancement network；
S4: establishing identification network, carries out identification to high quality graphic.
In step S2, personalized sample weights are introduced in picture quality detection network, for easily identifying mistake
Finger venous image sample introduces higher weight, and finger venous image sample easy to identify introduces lower weight, to disappear
Influence except user's otherness to recognition result.
In step S4, identification network is personalized identification network.
Training stage in this method are as follows:
1) finger venous image is collected, training set is constructed, quality status stamp is carried out for training set finger venous image, is divided into high-quality
Spirogram picture and low-quality image.
2) training image quality testing network is trained, so as to finger vein image on existing training set
Quality is detected.
3) image enhancement network is constructed.In general, in finger vena identification, compare low-quality image, high quality
Image is easier to be identified, and therefore, improves finger vein image quality to improve the accuracy of finger vena identification.In this method
The characteristics of by study high quality graphic, the finger venous image of high quality can be generated based on low-quality finger venous image.
4) personalized identification network is established.By the processing of quality testing network and image enhancement network, so that carrying out
Before identification, all images are all high quality graphics.In order to reduce the otherness between different user, picture quality is detected in network
Personalized sample weights are introduced, the finger venous image sample for easily identifying mistake introduces higher weight, easy to identify
Finger venous image sample introduce lower weight, to eliminate influence of user's otherness to recognition result.
Cognitive phase are as follows:
1) finger venous image of collecting test user, is input in network, first passes around quality testing network to test user
Finger vein image quality detected.
2) result detected is then identified to obtain identification knot using personalized identification network if it is high quality graphic
Fruit generates corresponding high quality graphic, is then input to personalization again conversely, finger venous image is input to enhancing network
Identification network, obtain final recognition result, carry out identification.
Preferably, detecting network as picture quality using Resnet in step S2.
Preferably, image enhancement network utilizes GAN frame in step S3, low-quality image is turned using TP-GAN technology
Change high quality graphic into.
A kind of finger vein recognition system based on enhancing network, including picture quality detect network module, image enhancement
Network module and identification network module, picture quality detect network module and carry out quality testing for finger vein image, point
High quality graphic and low-quality image out, image enhancement module are used to for low-quality image to be converted into high quality graphic, identify net
Network module is used to carry out identification to high quality graphic.
Network module finger vein image quality is detected by picture quality to be detected.In finger vena identification,
Compare low-quality image, and high quality graphic is easier to be identified, and therefore, it is quiet that raising finger vein image quality can improve finger
The accuracy of arteries and veins identification.Low-quality finger venous image can be generated to the finger vena of high quality by enhancing network module
Image.By the processing of quality testing network module and image enhancement network module, so that all images are all before being identified
It is high quality graphic.In order to reduce the otherness between different user, personalized sample is introduced in picture quality detection network
Weight, the finger venous image sample for easily identifying mistake introduce higher weight, finger venous image sample easy to identify
The lower weight of this introducing, to eliminate influence of user's otherness to recognition result.Finally by identification network module to height
Quality image is identified, final recognition result is obtained, and completes identification.
Preferably, the system further includes image input module, acquires finger venous image and construct training set, pass through figure
Network module is detected as finger venous image is input to picture quality by input module, picture quality detects network module in training
Picture quality is detected on collection.
Finger venous image is collected, training set is constructed, picture quality detects network module on training set to picture quality
It is detected, separates high quality graphic and low-quality image.
Preferably, detecting net as picture quality using using Resnet in described image quality testing network module
Network.
Preferably, described image enhancing network module utilizes GAN frame, low-quality image is carried out using TP-GAN technology
It is converted into high quality graphic.
Compared with prior art, the finger vein identification method of the invention based on enhancing network has with following prominent
Beneficial effect: the finger vein identification method based on enhancing network improves image while identifying finger venous image
Quality facilitates the raising of accuracy of identification, preferably carries out identification, has good application value.
Detailed description of the invention
Fig. 1 is the flow chart of the finger vein identification method of the present invention based on enhancing network；
Fig. 2 is the structural block diagram of the finger vein recognition system of the present invention based on enhancing network.
Specific embodiment
Below in conjunction with drawings and examples, to the finger vein identification method and system of the invention based on enhancing network
It is described in further detail.
Embodiment
As shown in Figure 1, it is of the invention based on enhancing network finger vein identification method, finger vein image quality into
Row evaluation, separates high quality graphic and low-quality image, is enhanced using enhancing network low-quality image, by low-quality spirogram
As being converted into high quality graphic, identification is carried out by high quality graphic of the identification network to output.Including the training stage and
Cognitive phase, detailed process is as follows:
S1: acquisition finger venous image constructs training set.
S2: training image quality testing network detects network as picture quality using Resnet, carries out on training set
Training, finger vein image quality are detected, and then carry out step S4 if high quality graphic, otherwise carry out step S3.
S3: low-quality image is converted into high quality graphic by building image enhancement network.
S4: establishing identification network, carries out identification to high quality graphic.Image enhancement network utilizes GAN frame, uses
Low-quality image is converted into high quality graphic by TP-GAN technology.
Specifically, the training stage are as follows:
Finger venous image is collected, training set is constructed, quality status stamp is carried out for training set image, is divided into high quality graphic and low
Quality image；
Training image quality testing network is trained on existing training set, so as to finger vein image quality into
Row detection；
Construct image enhancement network.In general, in finger vena identification, compare low-quality image, and high quality graphic is more
It is easily identified, therefore, improves finger vein image quality to improve the accuracy of finger vena identification.Pass through in this method
The characteristics of practising high quality graphic can generate the finger venous image of high quality based on low-quality finger venous image；
Establish personalized identification network.By the processing of quality testing network and image enhancement network, so that before being identified,
All images are all high quality graphics.In order to reduce the otherness between different user, picture quality detects introducing in network
Property sample weights, the finger venous image sample for easily identifying mistake introduces higher weight, finger easy to identify
Vein image sample introduces lower weight, to eliminate influence of user's otherness to recognition result.
Cognitive phase are as follows:
The finger venous image of collecting test user, is input in network, first passes around quality testing network to test user's
Finger vein image quality is detected；
The result of detection then is identified to obtain recognition result, instead if it is high quality graphic using personalized identification network
It, is input to enhancing network for finger venous image, generates corresponding high quality graphic, is then input to personalized identification again
Network obtains final recognition result, carries out identification.
Finger vein recognition system based on enhancing network of the invention, including the detection of image input module, picture quality
Network module, image enhancement network module and identification network module.
Acquisition finger venous image simultaneously constructs training set, and finger venous image is input to image by image input module
Quality testing network module, picture quality detection network module detect picture quality on training set.In finger vena
In identification, compare low-quality image, and high quality graphic is easier to be identified, and finger can be improved by improving finger vein image quality
The accuracy of hand vein recognition.The finger that low-quality finger venous image can be generated high quality by enhancing network module is quiet
Arteries and veins image.By the processing of quality testing network module and image enhancement network module, so that before being identified, all images
It is all high quality graphic.Personalized sample weights are introduced in picture quality detection network, for easily identifying the finger of mistake
Vein image sample introduces higher weight, and finger venous image sample easy to identify introduces lower weight, to eliminate use
Influence of the family otherness to recognition result.High quality graphic is identified finally by identification network module, identity is completed and knows
Not.
Embodiment described above, the only present invention more preferably specific embodiment, those skilled in the art is at this
The usual variations and alternatives carried out within the scope of inventive technique scheme should be all included within the scope of the present invention.
Claims (8)
1. a kind of finger vein identification method based on enhancing network, it is characterised in that: finger vein image quality is commented
Valence separates high quality graphic and low-quality image, is enhanced using enhancing network low-quality image, low-quality image is turned
It changes high quality graphic into, identification is carried out by high quality graphic of the identification network to output.
2. the finger vein identification method according to claim 1 based on enhancing network, it is characterised in that: this method includes
Training stage and cognitive phase, detailed process is as follows:
S1: acquisition finger venous image constructs training set；
S2: training image quality testing network is trained on training set, and finger vein image quality is detected, if
Step S4 is then carried out for high quality graphic, otherwise carries out step S3；
S3: low-quality image is converted into high quality graphic by building image enhancement network；
S4: establishing identification network, carries out identification to high quality graphic.
3. the finger vein identification method according to claim 2 based on enhancing network, it is characterised in that: adopted in step S2
Resnet is used to detect network as picture quality.
4. the finger vein identification method according to claim 3 based on enhancing network, it is characterised in that: in step S3
Image enhancement network utilizes GAN frame, and low-quality image is converted into high quality graphic using TP-GAN technology.
5. it is a kind of based on enhancing network finger vein recognition system, it is characterised in that: including picture quality detection network module,
Image enhancement network module and identification network module, picture quality detect network module and carry out quality for finger vein image
Detection separates high quality graphic and low-quality image, and image enhancement module is used to low-quality image being converted into high quality graphic,
Identify that network module is used to carry out identification to high quality graphic.
6. the finger vein recognition system according to claim 5 based on enhancing network, it is characterised in that: further include image
Input module acquires finger venous image and constructs training set, finger venous image is input to figure by image input module
As quality testing network module, picture quality detection network module detects picture quality on training set.
7. the finger vein recognition system according to claim 6 based on enhancing network, it is characterised in that: described image matter
Network is detected as picture quality using using Resnet in amount detection network module.
8. the finger vein recognition system according to claim 7 based on enhancing network, it is characterised in that: described image increases
Strong network module utilizes GAN frame, carries out low-quality image using TP-GAN technology and is converted into high quality graphic.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811318592.6A CN109409314A (en) | 2018-11-07 | 2018-11-07 | A kind of finger vein identification method and system based on enhancing network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811318592.6A CN109409314A (en) | 2018-11-07 | 2018-11-07 | A kind of finger vein identification method and system based on enhancing network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109409314A true CN109409314A (en) | 2019-03-01 |
Family
ID=65472110
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811318592.6A Pending CN109409314A (en) | 2018-11-07 | 2018-11-07 | A kind of finger vein identification method and system based on enhancing network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109409314A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079624A (en) * | 2019-12-11 | 2020-04-28 | 北京金山云网络技术有限公司 | Method, device, electronic equipment and medium for collecting sample information |
| CN111428718A (en) * | 2020-03-30 | 2020-07-17 | 南京大学 | Natural scene text recognition method based on image enhancement |
| CN112258428A (en) * | 2020-12-21 | 2021-01-22 | 四川圣点世纪科技有限公司 | Finger vein enhancement method and device based on cycleGAN |
| CN117058727A (en) * | 2023-07-18 | 2023-11-14 | 广州脉泽科技有限公司 | Image enhancement-based hand vein image recognition method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102222216A (en) * | 2011-06-02 | 2011-10-19 | 天津理工大学 | Identification system based on biological characteristics of fingerprints |
| CN106326886A (en) * | 2016-11-07 | 2017-01-11 | 重庆工商大学 | Finger-vein image quality evaluation method and system based on convolutional neural network |
| CN106709450A (en) * | 2016-12-23 | 2017-05-24 | 上海斐讯数据通信技术有限公司 | Recognition method and system for fingerprint images |
| CN107590786A (en) * | 2017-09-08 | 2018-01-16 | 深圳市唯特视科技有限公司 | A kind of image enchancing method based on confrontation learning network |
| US20180165508A1 (en) * | 2016-12-08 | 2018-06-14 | Veridium Ip Limited | Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices |
-
2018
- 2018-11-07 CN CN201811318592.6A patent/CN109409314A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102222216A (en) * | 2011-06-02 | 2011-10-19 | 天津理工大学 | Identification system based on biological characteristics of fingerprints |
| CN106326886A (en) * | 2016-11-07 | 2017-01-11 | 重庆工商大学 | Finger-vein image quality evaluation method and system based on convolutional neural network |
| US20180165508A1 (en) * | 2016-12-08 | 2018-06-14 | Veridium Ip Limited | Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices |
| CN106709450A (en) * | 2016-12-23 | 2017-05-24 | 上海斐讯数据通信技术有限公司 | Recognition method and system for fingerprint images |
| CN107590786A (en) * | 2017-09-08 | 2018-01-16 | 深圳市唯特视科技有限公司 | A kind of image enchancing method based on confrontation learning network |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079624A (en) * | 2019-12-11 | 2020-04-28 | 北京金山云网络技术有限公司 | Method, device, electronic equipment and medium for collecting sample information |
| CN111079624B (en) * | 2019-12-11 | 2023-09-01 | 北京金山云网络技术有限公司 | Sample information acquisition method and device, electronic equipment and medium |
| CN111428718A (en) * | 2020-03-30 | 2020-07-17 | 南京大学 | Natural scene text recognition method based on image enhancement |
| CN112258428A (en) * | 2020-12-21 | 2021-01-22 | 四川圣点世纪科技有限公司 | Finger vein enhancement method and device based on cycleGAN |
| CN117058727A (en) * | 2023-07-18 | 2023-11-14 | 广州脉泽科技有限公司 | Image enhancement-based hand vein image recognition method and device |
| CN117058727B (en) * | 2023-07-18 | 2024-04-02 | 广州脉泽科技有限公司 | Image enhancement-based hand vein image recognition method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109409314A (en) | A kind of finger vein identification method and system based on enhancing network | |
| CN104795067B (en) | Voice interactive method and device | |
| CN105320248B (en) | Aerial gesture input method and device | |
| CN102542281B (en) | Non-contact biometric feature identification method and system | |
| CN111931758B (en) | Face recognition method and device combining facial veins | |
| CN102467663B (en) | Biometrics authentication device | |
| CN100514352C (en) | Vena characteristic extracting method of finger vena identification system | |
| CN105868613A (en) | Biometric feature recognition method, biometric feature recognition device and mobile terminal | |
| CN105184254B (en) | A kind of identity identifying method and system | |
| CN105999670A (en) | Shadow-boxing movement judging and guiding system based on kinect and guiding method adopted by same | |
| CN106411952B (en) | One kind is every lost motion state gesture user identity identifying method and device | |
| CN103793690A (en) | Human body biotic living body detection method based on subcutaneous bloodstream detection and application | |
| CN111462379A (en) | Access control management method, system and medium containing palm vein and face recognition | |
| CN103761465A (en) | Method and device for identity authentication | |
| CN101681497A (en) | Vein pattern management system, vein pattern registration device, vein pattern authentication device, vein pattern registration method, vein pattern authentication method, program, and vein data struc | |
| CN102542242B (en) | The biological characteristic area positioning method and device of contactless collection image | |
| CN108021892A (en) | A kind of human face in-vivo detection method based on extremely short video | |
| CN110287918A (en) | Vivo identification method and Related product | |
| CN105760841A (en) | Identify recognition method and identify recognition system | |
| CN110555380A (en) | Finger vein identification method based on Center Loss function | |
| CN104809450B (en) | Wrist vena identification system based on online extreme learning machine | |
| CN106340096B (en) | A kind of identification device and method for referring to hand vein recognition intelligent lock based on fingerprint | |
| CN104036245B (en) | A kind of biological feather recognition method based on online Feature Points Matching | |
| CN108197577A (en) | The finger venous image characteristic extraction method of joint Sobel and MFRAT | |
| Khan et al. | A new method to extract dorsal hand vein pattern using quadratic inference function |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190301 |