CN112651304B - Revocable palm print template generation method, device, equipment and storage medium based on feature fusion - Google Patents
Revocable palm print template generation method, device, equipment and storage medium based on feature fusion Download PDFInfo
- Publication number
- CN112651304B CN112651304B CN202011443892.4A CN202011443892A CN112651304B CN 112651304 B CN112651304 B CN 112651304B CN 202011443892 A CN202011443892 A CN 202011443892A CN 112651304 B CN112651304 B CN 112651304B
- Authority
- CN
- China
- Prior art keywords
- palm print
- fusion
- feature
- point
- features
- 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.)
- Active
Links
Images
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
- G06V40/1347—Preprocessing; Feature extraction
- G06V40/1359—Extracting features related to ridge properties; Determining the fingerprint type, e.g. whorl or loop
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/75—Organisation of the matching processes, e.g. simultaneous or sequential comparisons of image or video features; Coarse-fine approaches, e.g. multi-scale approaches; using context analysis; Selection of dictionaries
- G06V10/751—Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
-
- 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
- G06V40/1365—Matching; Classification
- G06V40/1376—Matching features related to ridge properties or fingerprint texture
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Human Computer Interaction (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a method, a device, equipment and a storage medium for generating a revocable palm print template based on feature fusion, wherein the method comprises the following steps: acquiring an original palm print image corresponding to original palm print information; intercepting a palm print ROI image from the original palm print image; extracting the palm print direction characteristics in the palm print ROI image; extracting palm print point features in the palm print ROI image; performing feature fusion on the palm print direction features and the palm print point features to obtain fusion features; and carrying out irreversible transformation on the fusion characteristics to obtain a revocable palm print template. The method obtains the fusion characteristic by fusing the palm print direction characteristic and the palm print point characteristic, obtains the revocable palm print template by carrying out irreversible transformation on the fusion characteristic, effectively protects the safety of the original palm print information, and further improves the authentication performance of the palm print.
Description
Technical Field
The invention belongs to the technical field of palm print recognition, and particularly relates to a revocable palm print template generation method, a revocable palm print template generation device, revocable palm print template generation equipment and a storage medium based on feature fusion.
Background
The biometric identification technology is closely combined with high-tech means such as optics, acoustics, biosensors and the principle of biometrics through a computer, and personal identity is identified by utilizing the inherent physiological characteristics (such as fingerprints, human faces, irises and the like) and behavior characteristics (such as handwriting, voice, gait and the like) of a human body.
The palm print is a novel non-contact biological feature recognition technology, which is a technology for acquiring the hand palm texture by using modes such as infrared irradiation and the like, extracting features by using a special algorithm and recognizing the features as user identification marks. Compared with other biological characteristics used for identification, the palm print has the advantage of being irreplaceable. Currently, the identity authentication process of biometric systems is mainly based on feature extraction and matching techniques, which all involve the calculation and authentication of biometric information of a user.
The biometric features contain a large amount of user information, and permanent information leakage can be caused after the user information is lost, so how to ensure that the safety and privacy of the user biometric information are realized on the basis of effectively identifying the user identity becomes a problem to be solved urgently in the biometric feature identification technology.
Disclosure of Invention
In order to solve the above problems in the prior art, the present invention provides a revocable palmprint template generating method, apparatus, device and storage medium based on feature fusion. The technical problem to be solved by the invention is realized by the following technical scheme:
a revocable palm print template generation method based on feature fusion comprises the following steps:
acquiring an original palm print image corresponding to original palm print information;
intercepting a palm print ROI image from the original palm print image;
extracting the palm print direction characteristics in the palm print ROI image;
extracting the characteristics of the palm print points in the palm print ROI image;
performing feature fusion on the palm print direction features and the palm print point features to obtain fusion features;
and carrying out irreversible transformation on the fusion characteristics to obtain a revocable palm print template.
In one embodiment of the present invention, extracting the palm print direction feature in the palm print ROI image comprises: and extracting the palm print direction characteristics in the palm print ROI image by using an adjacent direction fusion method.
In one embodiment of the present invention, extracting the palm print direction feature in the palm print ROI image by using a neighboring direction fusion method includes:
filtering the palm print ROI image by using MFRAT templates with different main direction angles to obtain a minimum response value, a secondary minimum response value, a minimum response direction index value and a secondary minimum response direction index value;
fusing different main direction angles of the MFRAT template by using the adjacent direction fusion method to obtain a plurality of fusion angles in adjacent directions;
and obtaining the palm print direction characteristic by utilizing the minimum response value, the secondary small response value, the minimum response direction index value, the secondary small response direction index value and the fusion angle of the plurality of adjacent directions based on an adjacent direction fusion method.
In one embodiment of the present invention, the extracting palm print point features in the palm print ROI image comprises: and extracting the palm print point features in the palm print ROI image by utilizing a SURF algorithm.
In one embodiment of the present invention, the extracting the palm print point feature in the palm print ROI image by using SURF algorithm comprises:
carrying out non-overlapping blocking on the palm print ROI image to obtain a palm print block image;
judging whether the pixel point of the palm print block image is an SURF feature point or not by utilizing an SURF algorithm, if so, recording the pixel point of the palm print block image to a SURF feature point set, and if not, not recording the pixel point of the palm print block image;
selecting a representative point from the SURF feature point set by using a minimum distance method, wherein the sum of the distances between the representative point and other SURF feature points in the SURF feature point set is minimum;
and obtaining a target coding value by using the representative point and the LBP operator, wherein the target coding value is the feature of the palm print point.
In an embodiment of the present invention, performing feature fusion on the palm print direction feature and the palm print point feature to obtain a fusion feature includes:
normalizing the palm print direction characteristic to obtain a normalized palm print direction characteristic;
normalizing the palm print point features to obtain normalized palm print point features;
and performing feature fusion on the normalized palm print direction features and the normalized palm print point features to obtain fusion features.
In an embodiment of the present invention, the irreversible transformation of the fused feature to obtain a revocable palmprint template includes: acquiring q random Gaussian projection vectors, and establishing a random Gaussian projection matrix by using the q random Gaussian projection vectors;
and obtaining an index set by using the fusion characteristics and the random Gaussian projection matrix, wherein the index set is the revocable palm print template.
A revocable palm print template generation device based on feature fusion comprises:
the original palm print processing module is used for acquiring an original palm print image corresponding to original palm print information and intercepting a palm print ROI image from the original palm print image;
the palm print feature extraction module is used for respectively extracting palm print direction features and palm print point features from the palm print ROI image;
and the palmprint characteristic fusion module is used for fusing the palmprint direction characteristic and the palmprint point characteristic to obtain a revocable palmprint template.
An electronic device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any one of claims 1 to 7 when executing the computer program.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the feature fusion based revocable palm print template generating method of any one of claims 1 to 7.
The invention has the beneficial effects that:
the invention aims at the problems and provides a method, a device, equipment and a storage medium for generating a revocable palm print template based on feature fusion.
The present invention will be described in further detail with reference to the accompanying drawings and examples.
Drawings
Fig. 1 is a flowchart of a revocable palmprint template generation method based on feature fusion according to an embodiment of the present invention;
FIG. 2 is a diagram of a process for acquiring a palm print ROI image according to an embodiment of the present invention;
FIG. 3 is a grid diagram of different MFRAT templates respectively filtered with ROI according to the embodiment of the present invention;
FIG. 4 is a schematic diagram of MFRAT template neighboring principal direction angle fusion provided by an embodiment of the present invention;
FIG. 5 is a flowchart of a method for generating a revocable palm print template by fusing a palm print direction feature and a palm print point feature according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a revocable palm print template generating apparatus based on feature fusion according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of an electronic device based on a feature fusion revocable palmprint template according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to specific examples, but the embodiments of the present invention are not limited thereto.
Example one
Referring to fig. 1 and fig. 2, fig. 1 is a flowchart of a revocable palm print template generation method based on feature fusion according to an embodiment of the present invention, and fig. 2 is a process diagram of acquiring a palm print ROI image according to an embodiment of the present invention. A revocable palm print template generation method based on feature fusion comprises the following steps:
Specifically, the method for acquiring the original palm print image comprises the following steps: the palm is put into an acquisition window of an acquisition instrument, palm print information is acquired through the acquisition instrument, the acquired palm print information is transmitted to a control computer of the acquisition instrument, an image of 384mm multiplied by 284mm is acquired, and the type of the acquisition instrument can be MS500, for example. The palm print collection can be referred to a PolyU palm print database established at hong Kong university of science.
And 2, intercepting a palm print ROI image from the original palm print image.
Further, the palm print ROI image is an image of the interphalangeal valley point.
The method for acquiring the palm print ROI image at the present stage is a palm print ROI (Region Of Interest) image extraction method related in Palmcode [ D.Zhang, W.K.Kong, J.you.Online palm print identification [ J ]. IEEE Transactions on Pattern Analysis and Machine Analysis, 2003, 25 (9): 1041-1050 ], and is not limited to the method, the palm print ROI image extraction method takes the intersection point between fingers as an ROI area reference point, and the specific steps Of ROI positioning and extracting comprise:
and 2.1, preprocessing the palm print ROI image by utilizing a low-pass Gaussian filter to remove the noise influence in the palm print ROI image.
And 2.2, binarizing the palm print ROI image by a Thresholding method to obtain a corresponding palm print binary image, wherein a palm print area is white, a non-palm print area is black, and finding out the edge of the palm print binary image, namely a palm print edge line, by using an edge detection algorithm.
And 2.3, based on the palm print binary image, taking a valley connecting point between the index finger and the middle finger and a valley connecting point between the little finger and the ring finger as a reference point, wherein the reference point is also a valley point.
In particular, for a contact-acquired image of a palm print, these two reference points can be determined by two fixed supports of the acquisition device.
And 2.4, connecting the two reference points by using a straight line as an x coordinate axis, and determining a y coordinate axis to establish a rectangular coordinate system for positioning the palm print ROI image by using the midpoint of the two reference points as an origin.
Specifically, a point on a y coordinate, which is three quarters of the distance between two reference points from the origin of a coordinate system, is taken as the center of the ROI image, the distance between the two reference points is taken as the side length, and a square area with four sides parallel to the x axis and the y axis respectively is defined as the palm print ROI image, and the palm print image of the palm print ROI image is extracted. In the low-resolution palm print image recognition, palm print ROI image extraction and preprocessing are firstly carried out.
And 3, extracting the palm print direction characteristics in the palm print ROI image.
Further, extracting the palm print direction features in the palm print ROI image by using an adjacent direction fusion method, comprising the following steps of:
and 3.1, filtering the palm print ROI image by using MFRAT templates with different main direction angles to obtain a minimum response value, a secondary minimum response value, a minimum response direction index value and a secondary minimum response direction index value.
Referring to fig. 3, fig. 3 is a grid diagram for filtering with ROIs using different MFRAT templates according to an embodiment of the present invention. The main direction angle is (pi/6) × t, t =0,1, …,5, namely, the palm print ROI image is filtered by using MFRAT templates with 6 different direction angles to obtain 6 response values, and the minimum response value R is found min The second smallest response value R sec Minimum response value R min Corresponding minimum response direction index value t min And the second smallest response value R sec Corresponding sub-minor response direction index value t sec The filtering process is to convolute the MFRAT template with the palm print ROI image by using 6 angles respectively.
Specifically, when detecting the direction of a certain pixel point in the palm print ROI image, a finite grid is first established with the point as the center, and the cumulative sum of the pixel gray values is respectively obtained along different linear directions, where the linear direction in which the cumulative sum of the pixel gray values is the direction of the pixel point.
And 3.2, fusing different main direction angles of the MFRAT template by using an adjacent direction fusion method to obtain a plurality of fusion angles in adjacent directions.
Further, the number of fusion angles is less than or equal to 12.
Specifically, please refer to fig. 4, where fig. 4 is a schematic diagram illustrating fusion of adjacent principal direction angles of an MFRAT template provided by the embodiment of the present invention, t is an index value of the direction angle, and when t =0, the principal direction angle is 0 °; t =1, the main direction angle is 30 °; t =2, the main direction angle is 60 °; t =3, the main direction angle is 90 °; when t =4, the main direction angle is 120 °; at t =5, the main direction angle is 150 °. For a certain pixel point (x, y) in the image, fusing 6 different direction angles of the MFRAT template by using an adjacent direction fusion method to obtain 12 fusion directions, and then the expression of the fusion direction index value o (x, y) of the point is as follows:
wherein o (x, y) represents a fusion direction index value obtained by fusing adjacent directions, and t min Denotes the minimum response direction index value, t sec Indicating the next smallest response direction index value, R sec Represents the second smallest response value, R min Indicates the minimum response value and r indicates the set threshold value.
In addition, the fusion direction is (π/12) o (x, y).
And 3.3, based on the adjacent direction fusion method, obtaining the palm print direction characteristics by utilizing the minimum response direction index value, the second minimum response direction index value, the minimum response value, the second minimum response value and the fusion angle of a plurality of adjacent directions.
Specifically, a minimum response direction index value, a second minimum response direction index value, a minimum response value, a second minimum response value and a plurality of adjacent direction fusion directions are used to obtain a fusion direction index value O (x, y), the fusion direction index value of each pixel point in the image is solved to construct a fusion direction index value matrix, and the matrix is converted into a column vector to obtain the direction characteristic O of the palm print.
If the minimum response direction and the second minimum response direction are adjacent and the difference between the minimum response value and the second minimum response value is less than the set threshold R, then R min Instead of the actual minimum response value, the fusion angle obtained by fusing adjacent directions should be between the minimum response direction and the next minimum response direction, and the threshold r is set to be 8. By the adjacent direction fusion method, the direction retrieval range is from [0,5 ] on the premise of not increasing MFRAT template]Expanded to [0,11]And the accuracy of the palm print direction characteristic coding is improved.
And 4, extracting the characteristics of the palm print points in the palm print ROI image.
Further, extracting the palm print point features in the palm print ROI image by utilizing the SURF algorithm, wherein the method comprises the following steps:
and 4.1, carrying out non-overlapping blocking on the palm print ROI image to obtain a palm print block image.
Specifically, non-overlapping blocking is performed on the extracted palm print ROI image, and a padding operation is performed on the edge portion, so that m palm print block images of n × n pixel sizes are obtained.
According to the current state of the art example: the size of the palm print ROI image is 128 x 128 pixels, the size of the palm print block image is 24 x 24 pixels, the edge filling operation is completed by firstly performing the filling operation, respectively expanding 8 pixels on the upper, lower, left and right sides of the palm print ROI image, expanding the size of the palm print ROI image to 144 x 144 pixels, uniformly filling the expanded partial pixel values to be 0, and finally performing the blocking, wherein m is 36, and n is 24.
And 4.2, judging whether the pixel point of the palmprint block image is an SURF feature point or not by using an SURF algorithm, if so, recording the pixel point of the palmprint block image into an SURF feature point set, and if not, not recording the pixel point of the palmprint block image.
Specifically, whether a pixel point of the palm print block image is a SURF feature point is judged by using a SURF algorithm, so that the problem that a plurality of SURF points are arranged in a part of the palm print block image and the distances between the SURF points are short can be solved.
The method specifically comprises the following steps:
and 4.2.1, establishing an integral image for the palm print ROI image.
And 4.2.2, building a size space by enlarging the size of the filter template, and convolving the integral image with the filter template enlarged each time to obtain a corresponding image of a subsequent layer.
And 4.2.3, judging whether the target pixel point is an extreme point or not by using the Hessian matrix determinant, and when the pixel point is larger than the pixels of 26 adjacent points of the upper and lower adjacent layers of the pixel point, taking the target pixel point as the extreme point.
And 4.2.4, after all the extreme points are detected, removing unstable characteristic points to obtain a SURF characteristic point set.
And 4.3, selecting a representative point from the SURF characteristic point set by using a minimum distance method, wherein the sum of the distances between the representative point and other SURF characteristic points in the SURF characteristic point set is minimum.
Specifically, each palmprint block image has one or more SURF feature points, and for the ith image block, when there are multiple SURF points, the sum of the distances between each SURF feature point and other SURF feature points is obtained, and the SURF feature point with the minimum sum of the distances is used as the representative point p of the image block i 。
Step 4.4, utilizing the representative point p i And the LBP operator obtains a target coding value, and the target coding value is the palm print point characteristic.
Specifically, using the representative point p i And the LBP operator obtains a coded value P i As the point feature of the palm print block image, the point feature of the palm print ROI image finally obtained is P = [ P = 1 ,Ρ 2 ,...,Ρ m ]。
Because the only representative point is selected for each palm print block image, the number of the representative points of the whole palm print ROI image is determined by the number of the blocks, based on the fact that the number of the representative points extracted for different palm print ROI images is the same, in addition, because the processing is carried out block by block, the extraction sequence of the representative points is fixed, the finally obtained representative points are orderly and the number is fixed.
And 5, performing feature fusion on the palm print direction features and the palm print point features to obtain fusion features.
Further, referring to fig. 5, fig. 5 is a flowchart of generating a revocable palm print template by fusing a palm print direction feature and a palm print point feature according to an embodiment of the present invention, where step 5 further includes:
and 5.1, normalizing the palm print direction characteristics to obtain normalized palm print direction characteristics.
Specifically, the palm print direction feature is normalized to the range of [0,1], and the normalized palm print direction feature is obtained.
The expression of the normalized palm print direction characteristic is as follows:
wherein, O' represents the normalized palm print direction characteristic, and O represents the palm print direction characteristic.
And 5.2, normalizing the palm print point features to obtain normalized palm print point features.
Specifically, the palm print point features are normalized to the range of [0,1], and normalized palm print point features are obtained.
The expression of the normalized palm print point features is:
wherein P' represents normalized palm print point features, and P represents palm print point features.
And 5.3, performing feature fusion on the normalized palm print direction features and the normalized palm print point features to obtain fusion features.
Specifically, after converting the normalized palm print direction features and the normalized palm print point features into column vectors, performing feature fusion to obtain fusion features, wherein the expression of the fusion features is as follows:
C＝[O′,Ρ′]；
where C denotes the fusion feature, O 'denotes the normalized palmprint orientation feature, and p' denotes the normalized palmprint point feature.
And 6, carrying out irreversible transformation on the fusion characteristics to obtain a revocable palm print template.
And 6.1, acquiring q random Gaussian projection vectors, and establishing a random Gaussian projection matrix by using the q random Gaussian projection vectors.
Specifically, q random Gaussian projection vectors are generated m times based on the distribution of Gaussian functions, and a random Gaussian projection matrix W is established i Random Gaussian projection matrix W i The expression of (a) is:
wherein i =1,2., m, j =1,2., q,
is a random gaussian projection vector.
And 6.2, obtaining an index set by using the fusion characteristics and the random Gaussian projection matrix, wherein the index set is a revocable palm print template.
Specifically, the palm print template protection technology generally converts a palm print feature template into a revocable palm print template through an irreversible transformation function, wherein a random gaussian projection vector is the irreversible transformation function, and multiplying a fusion feature by the random gaussian projection vector is the irreversible transformation function on the fusion feature, so as to protect the palm print template.
Fusing features with random Gaussian projection vectors in a random Gaussian projection matrix
Multiplying to obtain an index set, and recording an expression of an index value of the maximum value of the multiplication as follows:
wherein, X i Means of maximumThe index value of the large value, x is the input palm print feature vector.
Further, the index value set of the maximum value is an index set, that is, the palm print template can be cancelled, and the expression of the revocable palm print template is as follows:
X＝(X 1 ,X 2 ,...,X m )；
wherein X represents a revocable palm print template, X i ∈{1,2,...,q}，i＝1,2,...,m。
In summary, in the present embodiment, the palm print direction feature and the palm print point feature are respectively extracted from the obtained palm print ROI image, for a certain pixel point on the palm print image, the MFRAT template is used to obtain the main direction feature of the point, and then the relationship between the main direction and the adjacent direction is used to perform fusion to obtain the more accurate direction feature of the pixel point, and the palm print direction feature and the palm print point feature are fused and subjected to irreversible transformation to obtain the revocable palm print template, where the random gaussian projection vector is the irreversible transformation function, and the multiplication of the fusion feature and the random gaussian projection vector is the irreversible transformation of the fusion feature, so as to protect the palm print template.
Example two
Referring to fig. 6, fig. 6 is a schematic structural diagram of a revocable palm print template generating device based on feature fusion according to an embodiment of the present invention. The embodiment of the invention provides a revocable palm print template generating device based on feature fusion, which comprises:
the original palm print processing module is used for acquiring an original palm print image corresponding to original palm print information and intercepting a palm print ROI image from the original palm print image;
the palm print feature extraction module is used for respectively extracting palm print direction features and palm print point features from the palm print ROI image;
and the palmprint characteristic fusion module is used for fusing the palmprint direction characteristic and the palmprint point characteristic to obtain a revocable palmprint template.
In an embodiment of the present invention, the original palm print processing module specifically collects palm print information through a palm print collector, transmits the palm print information to a corresponding control computer in the form of a picture, and further performs ROI positioning and extraction by using intersection points between fingers as ROI region reference points through a palm print ROI image extraction method to obtain a palm print ROI image.
In one embodiment of the present invention, the palm print feature extraction module includes a palm print direction feature extraction section and a palm print point feature extraction section. Respectively obtaining a minimum response direction index value, a secondary small response direction index value, a minimum response value, a secondary small response value and a fusion direction index value of an adjacent main direction of the MFRAT template by extracting the palm print direction characteristics, wherein the expression of a fusion direction index value o (x, y) of the point is as follows: (ii) a
Wherein o (x, y) represents a fusion direction index value obtained by fusing adjacent directions, and t min Indicates the minimum response direction index value, t sec Indicating the next smallest response direction index value, R sec Representing the second smallest response value, R min The minimum response value is shown, r is the set threshold, and the fusion direction is (pi/12) o (x, y).
Based on the adjacent direction fusion method, the palm print direction characteristic is obtained by utilizing the minimum response direction index value, the second minimum response direction index value, the minimum response value, the second minimum response value and the fusion angle of a plurality of adjacent directions, if the minimum response direction and the second minimum response direction are adjacent, the difference between the minimum response value and the second minimum response value is smaller than a set threshold value R, the set threshold value R is 8, and at the moment, R is min Instead of the actual minimum response value, the fusion direction resulting from the fusion of adjacent directions should be between the minimum response direction and the next minimum response direction. By the adjacent direction fusion method, the direction retrieval range is from [0,5 ] on the premise of not increasing MFRAT template]Expanded to [0,11]。
The palm print ROI image is subjected to non-overlapping blocking through a palm print point feature extraction part to obtain a palm print block image, whether pixel points of the palm print block image are SURF feature points or not is judged through a SURF algorithm, the pixel points of the palm print block image judged to be the SURF feature points are recorded into a SURF feature point set, and then representative points are selected from the SURF feature point set through a minimum distance method.
Further, there are one or more SURF feature points within each palmprint block image. For the ith image block, when a plurality of SURF points exist, the sum of the distances between each SURF characteristic point and other SURF characteristic points is obtained, and the SURF characteristic point with the minimum sum of the distances is taken as a representative point p i 。
Finally, using the representative point p i And LBP operator obtains target coding value P i The target coding value is a palm print point feature, and the point feature of the palm print ROI image finally obtained is P = [ Pp = [ P = 1 ,Ρ 2 ,...,Ρ m ]And P denotes a set of target coding values.
In an embodiment of the present invention, the palm print feature fusion module is configured to obtain a normalized palm print direction feature and a normalized palm print point feature, respectively, fuse the normalized palm print direction feature and the normalized palm print point feature to obtain a fusion feature, and obtain an index set by using the fusion feature and a random gaussian projection matrix, where the index set is a revocable palm print template.
Wherein, the expression of the normalized palm print direction characteristic is as follows:
the expression of the normalized palmprint point features is:
wherein, O 'represents the normalized palm print direction characteristic, O represents the palm print direction characteristic, P' represents the normalized palm print point characteristic, and P represents the palm print point characteristic.
The expression of the fusion features is:
C＝[O′,Ρ′]；
where C denotes a fusion feature, O 'denotes a normalized palm print direction feature, and p' denotes a normalized palm print point feature.
Random Gaussian projection matrix W i The expression of (a) is:
wherein i =1,2, a.. Multidot., m, j =1,2, a.. Multidot., q,
is a random gaussian projection vector.
Multiplying the fusion characteristics by a random Gaussian projection vector in a random Gaussian projection matrix to obtain an index set, and recording an expression of an index value of a maximum value of the multiplication as follows:
wherein, X i And x is an input palm print feature vector.
Further, the index value set of the maximum value is an index set, that is, the palm print template can be cancelled, and the expression of the revocable palm print template is as follows:
X＝(X 1 ,X 2 ,...,X m )；
wherein X represents a revocable palm print template, X i ∈{1,2,...,q}，i＝1,2,...,m。
The revocable palmprint template generation device based on feature fusion provided by the embodiment of the invention can execute the method embodiment, and the implementation principle and the technical effect are similar, so that the details are not repeated.
EXAMPLE III
Referring to fig. 7, fig. 7 is a schematic structural diagram of an electronic device capable of revoking a palm print template based on feature fusion according to an embodiment of the present invention. The electronic equipment provided by the embodiment of the invention comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete mutual communication through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any one of claims 1-7 when executing the computer program.
The revocable palmprint template generating device based on feature fusion provided by the embodiment of the invention can execute the method embodiment, and the implementation principle and the technical effect are similar, and are not described again here.
Example four
Yet another embodiment of the present invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of:
acquiring an original palm print image corresponding to original palm print information;
intercepting a palm print ROI image from the original palm print image;
extracting palm print direction features in the palm print ROI image;
extracting the characteristics of the palm print points in the palm print ROI image;
and performing feature fusion on the palm print direction features and the palm print point features to obtain a revocable palm print template.
The computer-readable storage medium provided by the embodiment of the present invention may implement the above method embodiments, and the implementation principle and technical effect are similar, which are not described herein again.
The foregoing is a further detailed description of the invention in connection with specific preferred embodiments and it is not intended to limit the invention to the specific embodiments described. For those skilled in the art to which the invention pertains, several simple deductions or substitutions can be made without departing from the spirit of the invention, and all shall be considered as belonging to the protection scope of the invention.
Claims (6)
1. A revocable palm print template generation method based on feature fusion is characterized by comprising the following steps:
acquiring an original palm print image corresponding to original palm print information;
intercepting a palm print ROI image from the original palm print image;
extracting the palm print direction characteristics in the palm print ROI image;
extracting the characteristics of the palm print points in the palm print ROI image;
performing feature fusion on the palm print direction features and the palm print point features to obtain fusion features;
carrying out irreversible transformation on the fusion characteristics to obtain a revocable palm print template;
extracting the palm print direction features in the palm print ROI image, comprising the following steps:
extracting the palm print direction features in the palm print ROI image by using an adjacent direction fusion method;
extracting the palm print direction features in the palm print ROI image by using an adjacent direction fusion method, wherein the method comprises the following steps:
filtering the palm print ROI image by using MFRAT templates with different main direction angles to obtain a minimum response value, a secondary minimum response value, a minimum response direction index value and a secondary minimum response direction index value;
fusing different main direction angles of the MFRAT template by using the adjacent direction fusion method to obtain a plurality of fusion angles in adjacent directions;
based on an adjacent direction fusion method, obtaining the palm print direction characteristic by using the minimum response value, the secondary small response value, the minimum response direction index value, the secondary small response direction index value and the fusion angle of the plurality of adjacent directions;
extracting the palm print point characteristics in the palm print ROI image, comprising the following steps:
extracting the palm print point features in the palm print ROI image by utilizing a SURF algorithm;
extracting the palm print point features in the palm print ROI image by utilizing a SURF algorithm, comprising the following steps of:
carrying out non-overlapping blocking on the palm print ROI image to obtain a palm print block image;
judging whether the pixel point of the palm print block image is an SURF feature point or not by utilizing an SURF algorithm, if so, recording the pixel point of the palm print block image to a SURF feature point set, and if not, not recording the pixel point of the palm print block image;
selecting a representative point from the SURF feature point set by using a minimum distance method, wherein the sum of the distances between the representative point and other SURF feature points in the SURF feature point set is minimum;
and obtaining a target coding value by using the representative point and the LBP operator, wherein the target coding value is the feature of the palm print point.
2. The method for generating the revocable palm print template based on the feature fusion as claimed in claim 1, wherein the feature fusion of the palm print direction feature and the palm print point feature to obtain the fusion feature comprises:
normalizing the palm print direction characteristic to obtain a normalized palm print direction characteristic;
normalizing the palm print point features to obtain normalized palm print point features;
and performing feature fusion on the normalized palm print direction features and the normalized palm print point features to obtain fusion features.
3. The method for generating a revocable palm print template based on feature fusion according to claim 1, wherein the step of performing irreversible transformation on the fused features to obtain the revocable palm print template comprises the following steps:
acquiring q random Gaussian projection vectors, and establishing a random Gaussian projection matrix by using the q random Gaussian projection vectors;
and obtaining an index set by using the fusion characteristics and the random Gaussian projection matrix, wherein the index set is the revocable palm print template.
4. A revocable palm print template generating apparatus based on feature fusion, for implementing the revocable palm print template generating method of any one of claims 1 to 3, the revocable palm print template generating apparatus comprising:
the original palm print processing module is used for acquiring an original palm print image corresponding to original palm print information and intercepting a palm print ROI image from the original palm print image;
the palm print feature extraction module is used for respectively extracting palm print direction features and palm print point features from the palm print ROI image;
and the palm print feature fusion module is used for fusing the palm print direction feature and the palm print point feature to obtain a fusion feature, and performing irreversible transformation on the fusion feature to obtain a revocable palm print template.
5. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any one of claims 1-3 when executing the computer program.
6. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the feature fusion based revocable palmprint template generation method of any one of claims 1 to 3.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011443892.4A CN112651304B (en) | 2020-12-11 | 2020-12-11 | Revocable palm print template generation method, device, equipment and storage medium based on feature fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011443892.4A CN112651304B (en) | 2020-12-11 | 2020-12-11 | Revocable palm print template generation method, device, equipment and storage medium based on feature fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112651304A CN112651304A (en) | 2021-04-13 |
| CN112651304B true CN112651304B (en) | 2023-02-10 |
Family
ID=75350889
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011443892.4A Active CN112651304B (en) | 2020-12-11 | 2020-12-11 | Revocable palm print template generation method, device, equipment and storage medium based on feature fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112651304B (en) |
Family Cites Families (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8965066B1 (en) * | 2013-09-16 | 2015-02-24 | Eye Verify LLC | Biometric template security and key generation |
| CN103731271B (en) * | 2013-12-30 | 2017-06-30 | 北京工业大学 | A kind of online face identity authentication based on homomorphic cryptography and Chaotic Scrambling |
| CN104268533A (en) * | 2014-10-10 | 2015-01-07 | 南昌航空大学 | Non-contact revocable palm print bimodal authentication method |
| WO2016165097A1 (en) * | 2015-04-16 | 2016-10-20 | 中国科学院自动化研究所 | Image de-aliasing system |
| KR20180093589A (en) * | 2017-02-14 | 2018-08-22 | 엘아이지넥스원 주식회사 | Apparatus and method for extracting feature point to fusing image about maneuvering target |
| JP6712247B2 (en) * | 2017-06-09 | 2020-06-17 | 株式会社日立製作所 | Biometric signature system and biometric signature method |
| CN109902586A (en) * | 2019-01-29 | 2019-06-18 | 平安科技（深圳）有限公司 | Palmmprint extracting method, device and storage medium, server |
| CN109886202B (en) * | 2019-02-22 | 2022-09-16 | 济南大学 | IoM-based revocable palm print competition code feature identification method |
| CN110287847A (en) * | 2019-06-19 | 2019-09-27 | 长安大学 | Vehicle grading search method based on Alexnet-CLbpSurf multiple features fusion |
| CN110516594B (en) * | 2019-08-27 | 2022-03-18 | 安徽大学 | Protection method and protection device for finger vein feature template capable of being cancelled |
-
2020
- 2020-12-11 CN CN202011443892.4A patent/CN112651304B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN112651304A (en) | 2021-04-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mathur et al. | Methodology for partial fingerprint enrollment and authentication on mobile devices | |
| Alheeti | Biometric iris recognition based on hybrid technique | |
| Anand et al. | PoreNet: CNN-based pore descriptor for high-resolution fingerprint recognition | |
| CN103793705B (en) | Non-contact palmprint authentication method based on iteration RANSAC algorithm and local palmmprint descriptor | |
| Fischer et al. | A novel palm vein recognition approach based on enhanced local Gabor binary patterns histogram sequence | |
| Oldal et al. | Hand geometry and palmprint-based authentication using image processing | |
| Hegde et al. | Authentication using finger knuckle prints | |
| Giełczyk et al. | Lightweight verification schema for image-based palmprint biometric systems | |
| Conti et al. | An advanced technique for user identification using partial fingerprint | |
| Wang et al. | An efficient algorithm for fingerprint matching | |
| CN112651304B (en) | Revocable palm print template generation method, device, equipment and storage medium based on feature fusion | |
| Kalluri et al. | Dynamic ROI extraction algorithm for palmprints | |
| Tran et al. | Multimodal personal verification using likelihood ratio for the match score fusion | |
| Choraś | Lip-prints feature extraction and recognition | |
| Biradar | Personal identification using palmprint biometrics based on principal line approach | |
| EP3702958B1 (en) | Method for verifying the identity of a user by identifying an object within an image that has a biometric characteristic of the user and separating a portion of the image comprising the biometric characteristic from other portions of the image | |
| Singh et al. | A line feature approach to finger knuckle image recognition | |
| Kanchana et al. | Quadtree decomposition for palm print feature representation in palmprint recognition system | |
| JP2659046B2 (en) | Identity verification device | |
| JP2007179267A (en) | Pattern matching device | |
| Amirthalingam et al. | Multimodal biometric cryptosystem for face and ear recognition based on fuzzy vault | |
| Kuban et al. | A NOVEL MODIFICATION OF SURF ALGORITHM FOR FINGERPRINT MATCHING. | |
| Chopra et al. | Finger print and finger vein recognition using repeated line tracking and minutiae | |
| Pukdesree et al. | Reviewed: The face authentication processes for accessing cloud computing services using iPhone | |
| Kazhagamani et al. | A Hough transform based feature extraction algorithm for finger knuckle biometric recognition system |
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 |