CN107729863A - Human body refers to vein identification method - Google Patents
Human body refers to vein identification method Download PDFInfo
- Publication number
- CN107729863A CN107729863A CN201711042722.3A CN201711042722A CN107729863A CN 107729863 A CN107729863 A CN 107729863A CN 201711042722 A CN201711042722 A CN 201711042722A CN 107729863 A CN107729863 A CN 107729863A
- Authority
- CN
- China
- Prior art keywords
- value
- target image
- point set
- gray
- minimum
- 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 title claims abstract description 34
- 210000003462 vein Anatomy 0.000 title claims abstract description 25
- 230000000694 effects Effects 0.000 claims abstract description 44
- 238000010586 diagram Methods 0.000 claims abstract description 12
- 238000012217 deletion Methods 0.000 claims abstract description 4
- 230000037430 deletion Effects 0.000 claims abstract description 4
- 238000003064 k means clustering Methods 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 3
- 210000004204 blood vessel Anatomy 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- 210000001367 artery Anatomy 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
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
- G06V40/12—Fingerprints or palmprints
- G06V40/13—Sensors therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
-
- 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
-
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/14—Vascular patterns
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a kind of human body to refer to vein identification method, and it includes target image being contracted to setting ratio, and all pixels in target image are clustered for the first time by its gray value；The cluster set corresponding to center of maximum value and minimum central value that deletion clusters for the first time, and the gray value in remaining cluster set is clustered again respectively；The class that central value is maximum in clustering again is deleted, and remaining class in each classification is merged into gray scale collection, according to maximum gradation value and minimum gradation value in each classification, the gray value concentrated to gray scale is updated to obtain feature diagram data；The pixel value more than zero in feature diagram data is extracted, and a spatial point set is formed respectively for each feature diagram data；Target image and total registration effect value of fingerprint image template are calculated using ICP algorithm；When minimum value is less than given threshold in all total registration effect values, then target image is same personnel with generating the picker corresponding to the fingerprint image template of minimum registration effect value.
Description
Technical field
The present invention relates to fingerprint identification technology, and in particular to a kind of human body refers to vein identification method.
Background technology
In fingerprint identification process, the acquisition because referring to vein image needs sensitive chip to work in high ISO (high photosensitive) mould
Under formula, the high photobehavior of image can make to be superimposed stronger shot noise on the image of acquisition, cause to assume based on smoothed image
Numerous Digital Image Processing algorithms be difficult to or can not work；Meanwhile recognition time, the degree of accuracy hard constraints under, it is also difficult
To use the intensive algorithms such as conventional noise remove, images match.
Therefore, it is a kind of for shot noise, illumination-insensitive, and the finger for being adapted to the computing platform of low-power consumption low performance is quiet
Arteries and veins recognizer is urgently developed.
The content of the invention
For above-mentioned deficiency of the prior art, the invention provides one kind identification, accurately human body refers to hand vein recognition side
Method.
In order to reach foregoing invention purpose, the technical solution adopted by the present invention is：
A kind of human body is provided and refers to vein identification method, it includes：
The fingerprint image of current picker is obtained, and intercepts fingerprint region as target image；
Target image is contracted to setting ratio, and using k-means clustering algorithms to all pixels in target image
Clustered for the first time by its gray value, and generate at least five classifications；
The cluster set corresponding to center of maximum value and minimum central value that deletion clusters for the first time, and clustered using k-means
Algorithm carries out the cluster again of at least three classes to the gray value in remaining cluster set respectively；
The class that central value is maximum in clustering again is deleted, and remaining class in each classification is merged into gray scale collection, same to markers
Remember the gray scale collection corresponding to two center of maximum values in the first cluster retained, remaining gray scale collection is merged into a gray scale collection；
According to maximum gradation value and minimum gradation value in each classification, the gray value concentrated to gray scale is updated to obtain spy
Levy diagram data；
The pixel value more than zero in feature diagram data is extracted, and a spatial point is formed respectively for each feature diagram data
Set；
According to the templatespace point set of fingerprint image template in the spatial point set of target image and database, use
ICP algorithm calculates target image and total registration effect value of fingerprint image template；
When minimum value is less than given threshold in all total registration effect values, then target image is with generating minimum registration effect
Picker corresponding to the fingerprint image template of value is same personnel.
Further, the human body refers to vein identification method and also included when minimum value is more than or equal in all total registration effect values
During given threshold, then the average value of all total registration effect values is calculated；
If minimum value is less than the setting multiple of average value in all total registration effect values, target image is matched somebody with somebody with generation minimum
Picker corresponding to the fingerprint image template of quasi- Effect value is same personnel, and otherwise current picker can not be known by fingerprint
Not.
Further, set multiple and be more than zero and less than one.
Further, according to maximum gradation value and minimum gradation value in each classification, the gray value that gray scale is concentrated is carried out
Renewal specific formula be：
New_v=(v-min_v)/max_v*255
Wherein, new_v is the gray value after renewal；V is the gray value that gray scale is concentrated；Min_v is maximum gradation value；max_
V is minimum gradation value.
Further, according to the templatespace point set of fingerprint image template in the spatial point set of target image and database
Close, target image is calculated using ICP algorithm and total registration effect value of fingerprint image template further comprises：
Spatial point set and the templatespace point set of fingerprint image template in database are matched somebody with somebody using ICP algorithm
It is accurate：
new_sample_rd0_ points=Rk*sample_rd0_points+tk
Wherein, new_sample_rd0_ points is spatial point set；RkFor spin matrix；tkFor translation vector；
According to the registration effect value generated in spatial point set registration process, target image and fingerprint image in database are calculated
As total registration effect value of template：
fitk=n0*fitk0+n1*fitk1…+nx*fitkx
Wherein, fitk0、fitk1…fitkxFor the registration effect value of each spatial point set, n0、n1…nxFor each space
The weight coefficient of point set；fitkFor the registration effect value of target image.
Further, when clustering for the first time, at least clustered for the first time three times, generate at least five classifications every time, chosen
What Clustering Effect was best at least clustering three times is once used as first cluster result, enters back into the maximum deleted and clustered for the first time afterwards
Cluster set step corresponding to central value and minimum central value.
Further, the templatespace point set acquisition methods of fingerprint image template and the space of target image in database
Point set acquisition methods are identical.
Further, using the fingerprint image and fingerprint image template for referring to hand vein recognition machine acquisition picker.
Further, it is specially that target image is contracted into original target image target image to be contracted into setting ratio
0.25 times.
Beneficial effects of the present invention are：Solve noise jamming and reduction operand by reduce target image
Problem, the effective information on original image is also effectively maintained, does not cause the decline of recognition correct rate.In generation characteristic pattern number
According in processing procedure, deleting processing and can progressively isolate more complete blood vessel by k-means clustering algorithms and gray value
Image.
Grey scale pixel value segmentation figure picture is based on using the method for k-means automatic clusters in this programme, in a variety of illumination bars
It can effectively realize that image is split under part；Directly using the central value each clustered, which cluster result can be quickly recognized
In pixel belong to useless background information, so as to extract useful blood-vessel image information.
By the pixel data of image, according to its position and gray value information on image, it is converted on three dimensions
Spatial point set (also referred to as puts cloud), is then configured using the ICP algorithm and templatespace point set of point cloud registering so that
The recognizer even in sampling when finger position and posture it is skimble-scamble in the case of, still be able to accurately be identified, have
Higher recognition accuracy and reliability.
It is other without carrying out using the registration effect value of ICP algorithm output as the measuring similarity between two point cloud datas
Dedicated calculation, reduce amount of calculation；Total registration effect value is calculated by using the mode of weighting, takes full advantage of each cut section
Domain include information number and quality height, improve final recognition accuracy.
Brief description of the drawings
Fig. 1 is the flow chart that human body refers to vein identification method one embodiment.
Embodiment
The embodiment of the present invention is described below, in order to which those skilled in the art understand this hair
It is bright, it should be apparent that the invention is not restricted to the scope of embodiment, for those skilled in the art,
As long as various change in the spirit and scope of the present invention that appended claim limits and determines, these changes are aobvious and easy
See, all are using the innovation and creation of present inventive concept in the row of protection.
With reference to figure 1, Fig. 1 shows that human body refers to the flow chart of vein identification method one embodiment；As shown in figure 1, the party
Method 100 includes step 101 to step 108.
In a step 101, the fingerprint image of current picker is obtained, and intercepts fingerprint region as target image；
Fingerprint image therein, which can use, refers to the acquisition of hand vein recognition machine.
In a step 102, target image is contracted to setting ratio, and using k-means clustering algorithms to target image
In all pixels clustered for the first time by its gray value, and generate at least five classifications.
During implementation, it is specially that target image is contracted into original mesh that target image is preferably contracted to setting ratio by this programme
0.25 times of logo image.
When being clustered for the first time, at least clustered for the first time three times, generate at least five classifications every time, chosen at least
What Clustering Effect was best in clustering three times is once used as first cluster result, enters back into step 103 afterwards.
In step 103, the cluster set corresponding to center of maximum value and minimum central value that deletion clusters for the first time, and use
K-means clustering algorithms carry out the cluster again of at least three classes to the gray value in remaining cluster set respectively；During first cluster
Corresponding central value is the average value of all gray values in cluster set.
At step 104, the class that central value is maximum in clustering again is deleted, and remaining class in each classification is merged into ash
Degree collection, while the gray scale collection in the first cluster of reservation corresponding to two center of maximum values is marked, remaining gray scale collection is merged into
One gray scale collection；Corresponding central value concentrates the average value of all gray values for gray scale when clustering again.
In step 105, according to maximum gradation value and minimum gradation value in each classification, the gray value concentrated to gray scale enters
Row renewal obtains feature diagram data；
In one embodiment of the invention, according to maximum gradation value and minimum gradation value in each classification, to gray scale collection
In the specific formula that is updated of gray value be：
New_v=(v-min_v)/max_v*255
Wherein, new_v is the gray value after renewal；V is the gray value that gray scale is concentrated；Min_v is maximum gradation value；max_
V is minimum gradation value.
In step 106, the pixel value in feature diagram data more than zero is extracted, and for each feature diagram data difference shape
Into a spatial point set；In particular, on each characteristic image, such as position i, the pixel value at j is vij, and vij>0, then
It is designated as point coordinates (i, j, v in three dimensional coordinate spaceij), space point set, which is combined into, meets that the three-dimensional coordinate of condition is empty in feature diagram data
Between middle point coordinates.
In step 107, according to the templatespace of fingerprint image template in the spatial point set of target image and database
Point set, target image and total registration effect value of fingerprint image template are calculated using ICP algorithm；
Wherein, the fingerprint image template of some different acquisition persons is stored with database, each fingerprint image template is both needed to
To obtain templatespace point set by the way of step 101 value step 106, namely in database fingerprint image template template
Spatial point set acquisition methods are identical with the spatial point set acquisition methods of target image.
In one embodiment of the invention, according to fingerprint image template in the spatial point set of target image and database
Templatespace point set, target image and total registration effect value of fingerprint image template are calculated using ICP algorithm and further wrapped
Include：
Spatial point set and the templatespace point set of fingerprint image template in database are matched somebody with somebody using ICP algorithm
It is accurate：
new_sample_rd0_ points=Rk*sample_rd0_points+tk
Wherein, new_sample_rd0_ points is spatial point set；RkFor spin matrix；tkFor translation vector；
According to the registration effect value generated in spatial point set registration process, target image and fingerprint image in database are calculated
As total registration effect value of template：
fitk=n0*fitk0+n1*fitk1…+nx*fitkx
Wherein, fitk0、fitk1…fitkxFor the registration effect value of each spatial point set, n0、n1…nxFor each space
The weight coefficient of point set；fitkFor the registration effect value of target image.
In step 108, when minimum value is less than given threshold in all total registration effect values, then target image and generation
Picker corresponding to the fingerprint image template of minimum registration effect value is same personnel.
During implementation, this programme body of choosing refers to vein identification method and also included：When minimum value in all total registration effect values
During more than or equal to given threshold, then the average value of all total registration effect values is calculated；
If minimum value is less than the multiple that sets of average value and (sets multiple more than zero and to be less than in all total registration effect values
One number), then target image is same personnel with generating the picker corresponding to the fingerprint image template of minimum registration effect value,
Otherwise current picker can not pass through fingerprint recognition.
Claims (10)
1. human body refers to vein identification method, it is characterised in that including：
The fingerprint image of current picker is obtained, and intercepts fingerprint region as target image；
The target image is contracted to setting ratio, and using k-means clustering algorithms to all pixels in target image
Clustered for the first time by its gray value, and generate at least five classifications；
The cluster set corresponding to center of maximum value and minimum central value that deletion clusters for the first time, and use k-means clustering algorithms
The cluster again of at least three classes is carried out to the gray value in remaining cluster set respectively；
The class that central value is maximum in clustering again is deleted, and remaining class in each classification is merged into gray scale collection, while marks guarantor
Gray scale collection in the first cluster stayed corresponding to two center of maximum values, remaining gray scale collection is merged into a gray scale collection；
According to maximum gradation value and minimum gradation value in each classification, the gray value concentrated to gray scale is updated to obtain characteristic pattern
Data；
The pixel value more than zero in feature diagram data is extracted, and a space point set is formed respectively for each feature diagram data
Close；
According to the templatespace point set of fingerprint image template in the spatial point set of target image and database, calculated using ICP
Method calculates target image and total registration effect value of fingerprint image template；
When minimum value is less than given threshold in all total registration effect values, then the target image is with generating minimum registration effect
Picker corresponding to the fingerprint image template of value is same personnel.
2. human body according to claim 1 refers to vein identification method, it is characterised in that also includes：
When minimum value is more than or equal to given threshold in all total registration effect values, then being averaged for all total registration effect values is calculated
Value；
If minimum value is less than the setting multiple of average value in all total registration effect values, the target image is matched somebody with somebody with generation minimum
Picker corresponding to the fingerprint image template of quasi- Effect value is same personnel, and otherwise current picker can not be known by fingerprint
Not.
3. human body according to claim 2 refers to vein identification method, it is characterised in that the setting multiple is more than zero and small
Yu Yi.
4. human body according to claim 1 refers to vein identification method, it is characterised in that maximum in each classification of basis
Gray value and minimum gradation value, the specific formula that the gray value concentrated to gray scale is updated are：
New_v=(v-min_v)/max_v*255
Wherein, new_v is the gray value after renewal；V is the gray value that gray scale is concentrated；Min_v is maximum gradation value；Max_v is
Minimum gradation value.
5. human body according to claim 1 refers to vein identification method, it is characterised in that the space according to target image
The templatespace point set of point set and fingerprint image template in database, target image and fingerprint image are calculated using ICP algorithm
As total registration effect value of template further comprises：
It is registering with the templatespace point set progress of fingerprint image template in database to spatial point set using ICP algorithm：
new_sample_rd0_ points=Rk*sample_rd0_points+tk
Wherein, new_sample_rd0_ points is spatial point set；RkFor spin matrix；tkFor translation vector；
According to the registration effect value generated in spatial point set registration process, target image and fingerprint image mould in database are calculated
Total registration effect value of plate：
fitk=n0*fitk0+n1*fitk1…+nx*fitkx
Wherein, fitk0、fitk1…fitkxFor the registration effect value of each spatial point set, n0、n1…nxFor each space point set
The weight coefficient of conjunction；fitkFor the registration effect value of target image.
6. human body according to claim 1 refers to vein identification method, it is characterised in that when clustering for the first time, at least carries out three
Secondary first cluster, generates at least five classifications every time, and Clustering Effect is best during selection at least clusters three times is once used as just
Secondary cluster result, enter back into delete the center of maximum value that clusters for the first time and the cluster set step corresponding to minimum central value afterwards.
7. vein identification method is referred to according to any described human bodies of claim 1-6, it is characterised in that fingerprint in the database
The templatespace point set acquisition methods of image template are identical with the spatial point set acquisition methods of target image.
8. vein identification method is referred to according to any described human bodies of claim 1-6, it is characterised in that use and refer to hand vein recognition machine
Obtain the fingerprint image and fingerprint image template of picker.
9. vein identification method is referred to according to any described human bodies of claim 1-6, it is characterised in that the target image contracts
As low as setting ratio is specially 0.25 times that the target image is contracted to original target image.
10. vein identification method is referred to according to any described human bodies of claim 1-6, it is characterised in that institute is right during first cluster
The central value answered is the average value of all gray values in cluster set, and corresponding central value is concentrated all for gray scale when clustering again
The average value of gray value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711042722.3A CN107729863B (en) | 2017-10-30 | 2017-10-30 | Human finger vein recognition method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711042722.3A CN107729863B (en) | 2017-10-30 | 2017-10-30 | Human finger vein recognition method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107729863A true CN107729863A (en) | 2018-02-23 |
| CN107729863B CN107729863B (en) | 2020-11-17 |
Family
ID=61203378
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711042722.3A Active CN107729863B (en) | 2017-10-30 | 2017-10-30 | Human finger vein recognition method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107729863B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109165639A (en) * | 2018-10-15 | 2019-01-08 | 广州广电运通金融电子股份有限公司 | A kind of finger vein identification method, device and equipment |
| CN112102210A (en) * | 2020-11-17 | 2020-12-18 | 北京圣点云信息技术有限公司 | Vein image template updating method and device based on self-learning |
| CN112200156A (en) * | 2020-11-30 | 2021-01-08 | 四川圣点世纪科技有限公司 | Vein recognition model training method and device based on clustering assistance |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103870808A (en) * | 2014-02-27 | 2014-06-18 | 中国船舶重工集团公司第七一〇研究所 | Finger vein identification method |
| CN105975974A (en) * | 2016-05-10 | 2016-09-28 | 深圳市金脉智能识别科技有限公司 | ROI image extraction method in finger vein identification |
| CN105975951A (en) * | 2016-05-27 | 2016-09-28 | 国创科视科技股份有限公司 | Finger vein and fingerprint fusion identification method of middle part of finger |
-
2017
- 2017-10-30 CN CN201711042722.3A patent/CN107729863B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103870808A (en) * | 2014-02-27 | 2014-06-18 | 中国船舶重工集团公司第七一〇研究所 | Finger vein identification method |
| CN105975974A (en) * | 2016-05-10 | 2016-09-28 | 深圳市金脉智能识别科技有限公司 | ROI image extraction method in finger vein identification |
| CN105975951A (en) * | 2016-05-27 | 2016-09-28 | 国创科视科技股份有限公司 | Finger vein and fingerprint fusion identification method of middle part of finger |
Non-Patent Citations (2)
| Title |
|---|
| DUN TAN ETC.: ""Categorizing Finger-Vein Images Using a Hierarchal Approach"", 《CHINESE CONFERENCE ON BIOMETRIC RECOGNITION》 * |
| JAYACHANDER SURBIRYALA ETC.: ""Finger vein indexing based on binary feature"", 《2015 COLOUR AND VISUAL COMPUTING SYMPOSIUM》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109165639A (en) * | 2018-10-15 | 2019-01-08 | 广州广电运通金融电子股份有限公司 | A kind of finger vein identification method, device and equipment |
| CN109165639B (en) * | 2018-10-15 | 2021-12-10 | 广州广电运通金融电子股份有限公司 | Finger vein identification method, device and equipment |
| CN112102210A (en) * | 2020-11-17 | 2020-12-18 | 北京圣点云信息技术有限公司 | Vein image template updating method and device based on self-learning |
| CN112200156A (en) * | 2020-11-30 | 2021-01-08 | 四川圣点世纪科技有限公司 | Vein recognition model training method and device based on clustering assistance |
| CN112200156B (en) * | 2020-11-30 | 2021-04-30 | 四川圣点世纪科技有限公司 | Vein recognition model training method and device based on clustering assistance |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107729863B (en) | 2020-11-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6672371B2 (en) | Method and apparatus for learning a classifier | |
| CN110348330B (en) | Face pose virtual view generation method based on VAE-ACGAN | |
| Davison et al. | Objective micro-facial movement detection using facs-based regions and baseline evaluation | |
| US20210264144A1 (en) | Human pose analysis system and method | |
| CN110569731B (en) | Face recognition method and device and electronic equipment | |
| WO2017219391A1 (en) | Face recognition system based on three-dimensional data | |
| JP2017016593A (en) | Image processing apparatus, image processing method, and program | |
| CN112329662B (en) | Multi-view saliency estimation method based on unsupervised learning | |
| CN107729863A (en) | Human body refers to vein identification method | |
| JP2021144749A (en) | Person collation device, method, and program | |
| CN110021019A (en) | A kind of thickness distributional analysis method of the AI auxiliary hair of AGA clinical image | |
| CN110033448A (en) | A kind of male bald Hamilton classification prediction analysis method of AI auxiliary of AGA clinical image | |
| CN111598144B (en) | Training method and device for image recognition model | |
| Stephen et al. | Enhancing fingerprint image through ridge orientation with neural network approach and ternarization for effective minutiae extraction | |
| KR101779642B1 (en) | Method of comparing images of irises by intelligent selection of textured zones | |
| CN108154107B (en) | Method for determining scene category to which remote sensing image belongs | |
| Saha et al. | An approach to detect the region of interest of expressive face images | |
| CN113269080B (en) | Palm vein identification method based on multi-channel convolutional neural network | |
| Chai et al. | Towards contactless palm region extraction in complex environment | |
| Hassan et al. | Enhance iris segmentation method for person recognition based on image processing techniques | |
| Sabitha et al. | Contrast enhancement for MRI image using hybrid optimization technique | |
| CN105701472A (en) | Method and device for identifying face of dynamic target | |
| Ma et al. | A lip localization algorithm under variant light conditions | |
| CN117523435A (en) | Fake video detection method and device based on graph network time sequence consistency | |
| CN117423158A (en) | Gait recognition method based on three-dimensional point cloud data enhancement |
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 |