CN108416342A - A kind of fingerprint identification method of combination minutiae point and filament structure - Google Patents
A kind of fingerprint identification method of combination minutiae point and filament structure Download PDFInfo
- Publication number
- CN108416342A CN108416342A CN201810524261.1A CN201810524261A CN108416342A CN 108416342 A CN108416342 A CN 108416342A CN 201810524261 A CN201810524261 A CN 201810524261A CN 108416342 A CN108416342 A CN 108416342A
- Authority
- CN
- China
- Prior art keywords
- fingerprint
- minutiae
- image
- point
- filament structure
- 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
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
- G06V40/1353—Extracting features related to minutiae or pores
-
- 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
- G06V10/28—Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
-
- 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
- G06V10/30—Noise filtering
-
- 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
- G06V40/1371—Matching features related to minutiae or pores
Abstract
The invention discloses a kind of fingerprint identification methods of combination minutiae point and filament structure.The present invention includes the following steps：Step 1, all finger print images for registering users of acquisition, and pretreatment and feature extraction are carried out to fingerprint image respectively, minutiae point and filament structure feature are obtained, and be saved in database respectively, establishes finger print database；Step 2, the finger print image for acquiring user to be identified, and pretreatment and feature extraction are carried out to fingerprint image, obtain the minutiae feature and filament structure feature of the finger print image of user to be known；Step 3, the finger print image for treating knowledge user match.The present invention makes full use of the directional field information around minutiae point so that minutiae point has stronger distinguishing ability, reduces false detail point to matched influence；And minutiae feature can quickly extract the matching reference minutiae of two fingerprints, reduce the time of fast fingerprint matching.
Description
Technical field
The invention belongs to living things feature recognition and field of information security technology, more particularly to a kind of combination minutiae point and filament
The fingerprint identification method of structure.
Background technology
Fingerprint identification technology is a kind of biometrics identification technology, and because its is safe, stability is high, and generality is strong, simultaneous
Capacitive is strong and the convenient research hotspot as foreign scholar of collecting device.Finger print identification technology includes mainly acquisition finger
Fingerprint image, image preprocessing extract feature and match cognization.Currently, for high quality fingerprint image, fingerprint recognition system has
The performance of almost Perfect, high-precision, high efficiency, performance are stablized.In actual acquisition fingerprint image, due to collecting device, ring
The influence in border can not obtain high quality fingerprint image completely, and there are image incompleteness, the fingerprints of texture fracture for partial fingerprints therein
Ropy problem.For low quality fingerprint image there are a large amount of interference noises, effective information is small, the standard for causing single features to identify
True rate is relatively low.Therefore, it how to be directed to low quality fingerprint, small area fingerprint distorts fingerprint, the characteristics of latent fingerprint, improves fingerprint
The accuracy rate and efficiency of identification become the difficult point studied at present.
The match party of matching process most basic, most widely used in Automated Fingerprint Identification System based on fingerprint minutiae at present
Formula, i.e. Point Pattern Matching.But minutiae feature pattern comes with some shortcomings：First, minutiae point is unevenly distributed, in having for fingerprint
A little regions minutiae feature is seldom, especially when the region occupies the main region of fingerprint image, is known with minutiae feature
Other fingerprint can frequently result in the matching of mistake；Second is that low quality region is easily extracted the minutiae point of more falseness, and it is existing
Fingerprint pretreatment algorithm, when the region false detail point is far more than correct minutiae point be difficult it is accurate distinguish or retain compared with
More false detail points or a small number of correct minutiae points are all removed, to reduce fingerprint recognition accuracy；Third, seriously turning round
In bent fingerprint, the position and direction of minutiae point can all have greatly changed, although some matching algorithms adapt to fingerprint shape
The problem of change, but false acceptance rate can be made to increase while reducing false rejection rate.
In conclusion since minutiae feature comes with some shortcomings so that in feelings such as low quality, small area, serious distortions
Discrimination is extremely difficult to satisfied effect under condition.In fingerprint identification process, filament structure is also fingerprint important feature, from filament
For in structure, minutiae point represents the Characteristics of Mutation of filament structure, i.e. interruption or bifurcated, this feature occur suddenly for filament
It is significant, convenient for distinguishing different fingerprints, but less stable again.On the one hand in the region that no filament changes just without details
Point, on the other hand since the process of acquisition image causes the region having different from actual fingerprint and produces the change of mistake in turn
Extract the node of mistake；What the basic configuration of fingerprint image, lines trend etc. were characterized in being determined by filament structure, to more
Reliability.But if utilizing filament structure matching merely, since filament structure datum mark pair is difficult to determine again, filament structure is influenced
Alignment, can increase the time of fingerprint matching.
Invention content
The purpose of the present invention is in view of the deficiencies of the prior art, for low quality fingerprint, small area fingerprint, distortion fingerprint with
And latent fingerprint, single details Point Pattern Matching are difficult to meet the requirement of Automated Fingerprint Identification System, and it is thin to provide a kind of combination
The fingerprint identification method of node and filament structure.
The technical solution adopted by the present invention to solve the technical problems includes the following steps：
Step 1, all finger print images for registering users of acquisition, and pretreatment and feature are carried out to fingerprint image respectively
Extraction, obtains minutiae point and filament structure feature, and be saved in database respectively, establishes finger print database；
Step 2, the finger print image for acquiring user to be identified, and pretreatment and feature extraction are carried out to fingerprint image,
Obtain the minutiae feature and filament structure feature of the finger print image of user to be known；
Step 3, the finger print image for treating knowledge user match：
Step 1 is implemented as follows：
The finger print image of 1-1. acquisition registration users obtains the original of fingerprint image to fingerprint image gray proces
Gray level image；
1-2. obtains the field of direction of original-gray image by the orientation field computation method based on gradient, and to original gradation
Image is split, and obtains fingerprint ridge area image；
1-3. carries out image enhancement to fingerprint ridge area image；
Enhanced fingerprint ridge area image is changed into binary image by 1-4.；
1-5. filters out noise in binary image, then carries out Skeleton processing, then to Skeleton treated image into
Row deburring is handled, and obtains fingerprint filament structural images；
1-6. finds the minutiae point of fingerprint filament structural images by Ridge following method, obtains fingerprint filament structural images
Minutiae point, and fingerprint filament structural images will be obtained and the corresponding minutiae point of fingerprint filament structural images is stored in as a pair
In database；
1-7. repeats step 1-1~1-6, and all registration users acquisition is finished, fingerprint database is established.
Step 2 is implemented as follows：
Step 2, the finger print image for acquiring user to be identified, and pretreatment and feature extraction are carried out to fingerprint image,
Obtain the minutiae feature and filament structure feature of the finger print image of user to be known；
2-1. acquires the finger print image of user to be identified, to fingerprint image gray proces, obtains the original of fingerprint image
Beginning gray level image；
2-2. obtains the field of direction of original-gray image by the orientation field computation method based on gradient, and to original gradation
Image is split, and obtains fingerprint ridge area image；
2-3. carrying out image enhancement to fingerprint ridge area image；
Enhanced fingerprint ridge area image is changed into binary image by 2-4.；
2-5. filters out noise in binary image, then carries out Skeleton processing, then to Skeleton treated image into
Row deburring is handled, and obtains fingerprint filament structural images；
2-6. finds the minutiae point of fingerprint filament structural images by Ridge following method, obtains fingerprint filament structural images
Minutiae point, obtain the minutiae feature of fingerprint image；
Step 3, the finger print image for treating knowledge user match：
3-1. describes son according to minutiae point, and using the directional field information around minutiae point, it is on schedule right to obtain best base：
3-1-1. inputs two width fingerprint image A and B, A is fingerprint image to be identified, and the minutiae point of extraction is set A=
{a1,a2,···,an,···,aN}；B is the arbitrary fingerprint image in fingerprint base, and the minutiae point of extraction is set B={ b1,
b2,···,bm,,bM}。
3-1-2. chooses any one minutiae point a in fingerprint image An, all minutiae points in fingerprint image B are traversed,
If there are minutiae point b in fingerprint image Bm, and two minutiae point an、bmIt is identical to meet type, and position translation is in (± Δ x0,±
Δy0) in range, then enter step 3-1-4；If being not present and minutiae point a after traversing all minutiae points in fingerprint image BnIt is right
The minutiae point b answeredm, then the minutiae point a in fingerprint image A is abandonedn；
3-1-3. continues, from one minutiae point of selection in fingerprint image A, to repeat step 3-1-2, until having traversed fingerprint
All minutiae points in image A.
3-1-4. details of construction points description, description attached bag include three auxiliary magnets in the field of direction around minutiae point；Meter
Calculate the relative angular difference between being put two-by-two in three auxiliary magnets：
Δθk=| θi-θj| (i, j=0,1,2,3；I ＜ j) (1)
In formula (1), k is relative angular difference reference numeral in minutiae point description, in the range of 1≤k≤6.
Kth angular deviation is denoted as G (k) between two minutiae points:
In formula (2),For minutiae point anRelative angular difference,For minutiae point bmRelative angular difference.
3-1-5. judges whether two minutiae points are preliminary datum marks, if arbitrary G (k) is more than threshold value T in formula (2)1, explanation
2 points of mismatches, return to step 3-1-2；Otherwise two minutiae points are preliminary datum mark.Preliminary benchmark point set Q is recorded, logarithm is matched
For L, position offset (the Δ x between each pair of preliminary datum marki,Δyi,Δθi)。
3-1-6. is from preliminary datum mark to finding best base on schedule in collection Q.
The preliminary datum mark of set of computations is to concentrating any pair of match point (Δ xi,Δyi,Δθi) (i=1,2, L)
With the deviation d of remaining all match points in set Q (Δ x, Δ y, Δ θ)i：
In formula (3), by diIn be less than optimal criteria point threshold value T2A numerical value form set D, choose maximum in set D
Value Cm, maximum value CmCorresponding match point be best base on schedule, i.e. aiAnd bjIt is on schedule right for best base, corresponding position deviation
Parameter is (Δ xij,Δyij,Δθij)。
Step 3-2. minutiae features match
3-2-1. concentrates the position of all the points to calibrate minutiae point, and formula is as follows：
θ '=θ+Δ θij (5)
Wherein, S0For change of scale parameter.Details in fingerprint point set A={ a to be identified1,a2,···,aNBy translation and
Rotate (Δ xij,Δyij,Δθij) it is transformed into A'={ a1', a'2,···,a'N}。
Two minutiae matchings that 3-2-2. is usually defined generally are measured by the two " distance ".First calculate a 'iAnd bj's
Positional distance d (a 'i,bj) and field of direction differential seat angle dθ(a′i,bj).If d (a 'i,bj) it is less than positional distance empirical value Td, and
dθ(a′i,bj) it is less than field of direction differential seat angle empirical value Tθ, then two minutiae matchings.
dθ(a′i,bj)=min (| θ 'i-θj|,360°-|θ′i-θj|) (7)
According to (6) (7) two formula, two matched minutiae point similarity s are calculated：
In formula k, σ are constants, and d is that two details points set distance, dθFor two Minutiae Direction field differential seat angles.
3-2-3. counts the number of whole match points, minutiae matching score is calculated as follows according to (8) formula：
Wherein, NMIt counts out for matching, N and M are that details in fingerprint to be known counts out and counts out with template details respectively.
3-3. fingerprint filament structure matchings
The filament structure point set of 3-3-1. fingerprint images A extraction isThe filament of fingerprint image B extractions
Structured setPoint setIn each put and indicate the horizontal seat of point respectively containing only location information (x, y), x, y
Mark and ordinate.
3-3-2. is according to the transformation relation of formula (4), fingerprint filament structure point set to be identifiedBy
Translation and rotation (Δ xij,Δyij,Δθij) be transformed intoAfter transformationRefer to library
LineCompare to obtain two fingerprint overlapping region accounting T', if T' is more than presetting filament structure overlapping region area accounting
Threshold value T3, then in overlapping region carry out filament structure feature matching, hence into step 3-3-3 judge fingerprint whether
Match；Otherwise illustrate that two fingerprint overlapping regions are too small, be not the fingerprint of same finger, be judged to mismatching.
The improved Hausdorff distances of 3-3-3. carry out images match frequently as similarity measurement, and the present invention uses field
The recognition methods for searching for filament structure distance, utilizes filament structure distanceIt is used as the similar of two fingerprint filament structures
Index is spent, specific calculating is as follows：
Filament structure point setIt arrivesDistance definition：
By filament structure point setMiddle p point is to filament structure point setDistance be sorted in ascending order, take before i distance value
Average value as filament structure point setIt arrivesDistance value：
In formula (10), 1≤i≤p, p are filament structure point setThe number at midpoint, IthIndicate ascending sort, | | | | beBetween Euclidean distance.
Similarly, filament structure point setIt arrivesDistance definition：
By filament structure point setMiddle q point is to filament structure point setDistance be sorted in ascending order, take before j distance value
Average value as filament structure point setIt arrivesDistance value：
In formula, 1≤j≤q, q are point setThe number at midpoint, IthIndicate ascending sort, | | | | be Between Europe
Family name's distance.
Filament structure point setWith filament structure point setThe matching score of straight line is as follows：
Points in fingerprint filament structure will count far more than fingerprint characteristic, if using the method pair of each point search
Fingerprint filament structure is matched, and needs to consume a large amount of operation time.In order to reduce operation time, filament is searched for using field
The recognition methods of structure distance.
By can be seen that in (10) formula, the distance of any one point-to-point collection B in filament structure point set A' needs to calculate the point
The distance between to all the points in point set B, wherein minimum point is taken；It can thus be appreciated that when this calculation consumes a large amount of
Between.If only searching for the field of corresponding position in point-to-point collection B, (on the one hand Δ x, the distance of the point in Δ y) ranges are counted
The distance value of calculation is constant, the calculating time that on the other hand can be saved.
3-4. merges decision center
Since minutiae feature and filament structure feature are different essential features, threshold calculations mode is different,
Fusion Features are carried out using fractional layer information in decision-making level, expression formula is as follows：
Wherein, α is weight, SHThe matching score being calculated for filament structure feature；
Matching result judgement is as follows：
Wherein c is the similarity threshold for the 2 width fingerprint images that experiment obtains, is considered inhomogeneity fingerprint image less than c, otherwise
It is considered similar fingerprint image.
The present invention has the beneficial effect that：
1, the directional field information around minutiae point is made full use of so that minutiae point has stronger distinguishing ability, reduces void
False minutiae point is on matched influence；And minutiae feature can quickly extract the matching reference minutiae of two fingerprints, reduce fast
The time of fingerprint matching.
2, the carried fingerprint identification method of the present invention, in the case where misclassification rate is certain, recognition performance is apparently higher than conventional point
Mode identification method.Particularly, for low quality fingerprint, small area fingerprint, the fingerprint images such as distortion fingerprint, based on dot pattern knowledge
Other algorithm performance is decreased obviously, and combination minutiae point proposed by the present invention and filament structure recognition algorithm performance decline unobvious.
3, more meet the process that human expert compares fingerprint, energy than the level of minutiae point from the level of streakline to describe fingerprint
More features are enough provided for matching, avoid the defect of above-mentioned minutiae point, preferably distinguishing some can not be correct with minutiae point
Matched fingerprint improves the accuracy of fingerprint recognition.
4, the present invention is convenient for the change and upgrading of later stage algorithm：Minutiae point and filament structure are stored in database, when after
When phase needs to change or upgrade minutiae point algorithm, the filament structure in database can be directly read, then improved details
Point matching algorithm operates filament structure, to establish new details point data.
Description of the drawings
The flow chart of the combination minutiae point of Fig. 1 present invention and the fingerprint identification method of filament structure
Fig. 2 is the flow that fingerprint database is established in the present invention；
The fingerprint original image of Fig. 3 acquisitions
Fig. 4 Fingerprint diretion figures；
Fig. 5 fingerprint segmentation images；
Fig. 6 fingerprints enhance image；
Fig. 7 fingerprint binary images；
Fig. 8 fingerprint thinning structure feature figures；
Fig. 9 fingerprint minutiae feature figures；
Figure 10 minutiae points describe subgraph.
Specific implementation mode
Specific embodiments of the present invention are described further below in conjunction with the accompanying drawings.
As Figure 1-10 shows, a kind of fingerprint identification method of combination minutiae point and filament structure, the specific implementation process is as follows：
Step 1, all finger print images for registering users of acquisition, and pretreatment and feature are carried out to fingerprint image respectively
Extraction, obtains minutiae point and filament structure feature, and be saved in database respectively, establishes finger print database；
Step 2, the finger print image for acquiring user to be identified, and pretreatment and feature extraction are carried out to fingerprint image,
Obtain the minutiae feature and filament structure feature of the finger print image of user to be known；
Step 3, the finger print image for treating knowledge user match：
Step 1 is implemented as follows：
The finger print image of 1-1. acquisition registration users obtains the original of fingerprint image to fingerprint image gray proces
Gray level image, such as Fig. 3；
1-2. obtains the field of direction of original-gray image, such as Fig. 4 by the orientation field computation method based on gradient, and to original
Beginning gray level image is split, and obtains fingerprint ridge area image；Such as Fig. 5；
1-3. carries out image enhancement, such as Fig. 6 to fingerprint ridge area image；
Enhanced fingerprint ridge area image is changed into binary image, such as Fig. 7 by 1-4.；
1-5. filters out noise in binary image, then carries out Skeleton processing, then to Skeleton treated image into
Row deburring is handled, and obtains fingerprint filament structural images, such as Fig. 8；
1-6. finds the minutiae point of fingerprint filament structural images by Ridge following method, obtains fingerprint filament structural images
Minutiae point, such as Fig. 9, and fingerprint filament structural images and the corresponding minutiae point of fingerprint filament structural images will be obtained as a pair of
It is stored in database；
1-7. repeats step 1-1~1-6, and all registration users acquisition is finished, fingerprint database is established.
Step 2 is implemented as follows：
Step 2, the finger print image for acquiring user to be identified, and pretreatment and feature extraction are carried out to fingerprint image,
Obtain the minutiae feature and filament structure feature of the finger print image of user to be known；
2-1. acquires the finger print image of user to be identified, to fingerprint image gray proces, obtains the original of fingerprint image
Beginning gray level image；
2-2. obtains the field of direction of original-gray image by the orientation field computation method based on gradient, and to original gradation
Image is split, and obtains fingerprint ridge area image；
2-3. carries out image enhancement to fingerprint ridge area image；
Enhanced fingerprint ridge area image is changed into binary image by 2-4.；
2-5. filters out noise in binary image, then carries out Skeleton processing, then to Skeleton treated image into
Row deburring is handled, and obtains fingerprint filament structural images；
2-6. finds the minutiae point of fingerprint filament structural images by Ridge following method, obtains fingerprint filament structural images
Minutiae point, obtain the minutiae feature of fingerprint image；
Step 3, the finger print image for treating knowledge user match：
3-1. is as shown in Figure 10, and son is described according to minutiae point, using the directional field information around minutiae point, obtains best base
It is on schedule right：
3-1-1. inputs two width fingerprint image A and B, A is fingerprint image to be identified, and the minutiae point of extraction is set A=
{a1,a2,···,an,···,aN}；B is the arbitrary fingerprint image in fingerprint base, and the minutiae point of extraction is set B={ b1,
b2,···,bm,,bM}。
3-1-2. chooses any one minutiae point a in fingerprint image An, all minutiae points in fingerprint image B are traversed,
If there are minutiae point b in fingerprint image Bm, and two minutiae point an、bmIt is identical to meet type, and position translation is in (± Δ x0,±
Δy0) in range, then enter step 3-1-4；If being not present and minutiae point a after traversing all minutiae points in fingerprint image BnIt is right
The minutiae point b answeredm, then the minutiae point a in fingerprint image A is abandonedn；
3-1-3. continues, from one minutiae point of selection in fingerprint image A, to repeat step 3-1-2, until having traversed fingerprint
All minutiae points in image A.
3-1-4. details of construction points description, description attached bag include three auxiliary magnets in the field of direction around minutiae point；Meter
Calculate the relative angular difference between being put two-by-two in three auxiliary magnets：
Δθk=| θi-θj| (i, j=0,1,2,3；I ＜ j) (1)
In formula (1), k is relative angular difference reference numeral in minutiae point description, in the range of 1≤k≤6.
Kth angular deviation is denoted as G (k) between two minutiae points:
In formula (2),For minutiae point anRelative angular difference,For minutiae point bmRelative angular difference.
3-1-5. judges whether two minutiae points are preliminary datum marks, if arbitrary G (k) is more than threshold value T in formula (2)1, explanation
2 points of mismatches, return to step 3-1-2；Otherwise two minutiae points are preliminary datum mark.Preliminary benchmark point set Q is recorded, logarithm is matched
For L, position offset (the Δ x between each pair of preliminary datum marki,Δyi,Δθi)。
3-1-6. is from preliminary datum mark to finding best base on schedule in collection Q.
The preliminary datum mark of set of computations is to concentrating any pair of match point (Δ xi,Δyi,Δθi) (i=1,2, L)
With the deviation d of remaining all match points in set Q (Δ x, Δ y, Δ θ)i：
In formula (3), by diIn be less than optimal criteria point threshold value T2A numerical value form set D, choose maximum in set D
Value Cm, maximum value CmCorresponding match point be best base on schedule, i.e. aiAnd bjIt is on schedule right for best base, corresponding position deviation
Parameter is (Δ xij,Δyij,Δθij)。
Step 3-2. minutiae features match
3-2-1. concentrates the position of all the points to calibrate minutiae point, and formula is as follows：
θ '=θ+Δ θij (5)
Wherein, S0For change of scale parameter.Details in fingerprint point set A={ a to be identified1,a2,···,aNBy translation and
Rotate (Δ xij,Δyij,Δθij) it is transformed into A'={ a '1, a'2,···,a'N}。
Two minutiae matchings that 3-2-2. is usually defined generally are measured by the two " distance ".First calculate a 'iAnd bj's
Positional distance d (a 'i,bj) and field of direction differential seat angle dθ(a′i,bj).If d (a 'i,bj) it is less than positional distance empirical value Td, and
dθ(a′i,bj) it is less than field of direction differential seat angle empirical value Tθ, then two minutiae matchings.
dθ(a′i,bj)=min (| θ 'i-θj|,360°-|θ′i-θj|) (7)
According to (6) (7) two formula, two matched minutiae point similarity s are calculated：
In formula k, σ are constants, and d is that two details points set distance, dθFor two Minutiae Direction field differential seat angles.
3-2-3. counts the number of whole match points, minutiae matching score is calculated as follows according to (8) formula：
Wherein, NMIt counts out for matching, N and M are that details in fingerprint to be known counts out and counts out with template details respectively.
3-3. fingerprint filament structure matchings
The filament structure point set of 3-3-1. fingerprint images A extraction isThe filament of fingerprint image B extractions
Structured setPoint setIn each put and indicate the horizontal seat of point respectively containing only location information (x, y), x, y
Mark and ordinate.
3-3-2. is according to the transformation relation of formula (4), fingerprint filament structure point set to be identifiedBy
Translation and rotation (Δ xij,Δyij,Δθij) be transformed intoAfter transformationRefer to library
LineCompare to obtain two fingerprint overlapping region accounting T', if T' is more than presetting filament structure overlapping region area accounting
Threshold value T3, then in overlapping region carry out filament structure feature matching, hence into step 3-3-3 judge fingerprint whether
Match；Otherwise illustrate that two fingerprint overlapping regions are too small, be not the fingerprint of same finger, be judged to mismatching.
The improved Hausdorff distances of 3-3-3. carry out images match frequently as similarity measurement, and the present invention uses field
The recognition methods for searching for filament structure distance, utilizes filament structure distanceIt is used as the similar of two fingerprint filament structures
Index is spent, specific calculating is as follows：
Filament structure point setIt arrivesDistance definition：
By filament structure point setMiddle p point is to filament structure point setDistance be sorted in ascending order, take before i distance value
Average value as filament structure point setIt arrivesDistance value：
In formula (10), 1≤i≤p, p are filament structure point setThe number at midpoint, IthIndicate ascending sort, | | | | beBetween Euclidean distance.
Similarly, filament structure point setIt arrivesDistance definition：
By filament structure point setMiddle q point is to filament structure point setDistance be sorted in ascending order, take before j distance value
Average value as filament structure point setIt arrivesDistance value：
In formula, 1≤j≤q, q are point setThe number at midpoint, IthIndicate ascending sort, | | | | be Between Europe
Family name's distance.
Filament structure point setWith filament structure point setThe matching score of straight line is as follows：
Points in fingerprint filament structure will count far more than fingerprint characteristic, if using the method pair of each point search
Fingerprint filament structure is matched, and needs to consume a large amount of operation time.In order to reduce operation time, filament is searched for using field
The recognition methods of structure distance.
By can be seen that in (10) formula, the distance of any one point-to-point collection B in filament structure point set A' needs to calculate the point
The distance between to all the points in point set B, wherein minimum point is taken；It can thus be appreciated that when this calculation consumes a large amount of
Between.If only searching for the field of corresponding position in point-to-point collection B, (on the one hand Δ x, the distance of the point in Δ y) ranges are counted
The distance value of calculation is constant, the calculating time that on the other hand can be saved.
3-4. merges decision center
Since minutiae feature and filament structure feature are different essential features, threshold calculations mode is different,
Fusion Features are carried out using fractional layer information in decision-making level, expression formula is as follows：
Wherein, α is weight, SHThe matching score being calculated for filament structure feature；
Matching result judgement is as follows：
Wherein c is the similarity threshold for the 2 width fingerprint images that experiment obtains, is considered inhomogeneity fingerprint image less than c, otherwise
It is considered similar fingerprint image.
Embodiment：
The fingerprint image for referring to 2000 registration users is acquired in experiment with collecting device, establishes abnormal fingerprint image number
According to the database in library and normal fingerprints image blend.Wherein abnormal fingerprint image data library occupies 20% or so, including low-quality
Fingerprint is measured, small area fingerprint distorts the images such as fingerprint and latent fingerprint.The fingerprint image of acquisition extracts after pretreatment to be referred to
Line characteristic information simultaneously preserves.
The fingerprint image of acquisition registered users describes son using minutiae point and finds out two fingers as user to be identified successively
The reference point of line carries out minutiae point and filament structure alignment, then ask respectively a little according to reference point locations deviation to fingerprint to be identified
The match condition under match condition and filament structure under pattern, finally carries out the fusion of two situations, judge two fingerprints whether
Match.Obtain 2000 recognition results, wherein have 4 it is unidentified go out for registered users, discrimination 99.84% is based on
VS2017 platforms calculate 4.65 milliseconds of a used time of matching.And the method for using database storage feature point data, have 124
It is a it is unidentified go out registered users, discrimination 95.04%, and the method that filament structural database is only stored only with data,
The Mean match used time is 10.12 milliseconds.
Embodiment the result shows that, the algorithm for recognizing fingerprint of this patent combined based on minutiae point and filament structure is when saving
Between while ensure that higher discrimination.
It is finally noted that the purpose for publicizing and implementing example is to help to further understand the present invention, but this field
Technical staff be appreciated that：Without departing from the spirit and scope of the invention and the appended claims, various to replace and repair
It is all possible for changing.Therefore, the present invention should not be limited to embodiment disclosure of that, and the scope of protection of present invention is to weigh
Subject to the range that sharp claim defines.
Claims (7)
1. a kind of fingerprint identification method of combination minutiae point and filament structure, it is characterised in that include the following steps：
Step 1, acquisition it is all registration users finger print images, and respectively to fingerprint image carry out pretreatment and feature carry
It takes, obtains minutiae point and filament structure feature, and be saved in database respectively, establish finger print database；
Step 2, the finger print image for acquiring user to be identified, and pretreatment and feature extraction are carried out to fingerprint image, it obtains
The minutiae feature and filament structure feature of the finger print image of user to be known；
Step 3, the finger print image for treating knowledge user match；
Step 1 is implemented as follows：
The finger print image of 1-1. acquisition registration users obtains the original gradation of fingerprint image to fingerprint image gray proces
Image；
1-2. obtains the field of direction of original-gray image by the orientation field computation method based on gradient, and to original-gray image
It is split, obtains fingerprint ridge area image；
1-3. carries out image enhancement to fingerprint ridge area image；
Enhanced fingerprint ridge area image is changed into binary image by 1-4.；
1-5. filters out noise in binary image, then carries out Skeleton processing, and then to Skeleton, treated that image is gone
Burr processing, obtains fingerprint filament structural images；
1-6. finds the minutiae point of fingerprint filament structural images by Ridge following method, obtains the thin of fingerprint filament structural images
Node, and fingerprint filament structural images will be obtained and the corresponding minutiae point of fingerprint filament structural images is stored in data as a pair
In library；
1-7. repeats step 1-1~1-6, and all registration users acquisition is finished, fingerprint database is established.
2. the fingerprint identification method of a kind of combination minutiae point and filament structure according to claim 1, it is characterised in that step
Rapid 2 are implemented as follows：
2-1. acquires the finger print image of user to be identified, to fingerprint image gray proces, obtains the original ash of fingerprint image
Spend image；
2-2. obtains the field of direction of original-gray image by the orientation field computation method based on gradient, and to original-gray image
It is split, obtains fingerprint ridge area image；
2-3. carries out image enhancement to fingerprint ridge area image；
Enhanced fingerprint ridge area image is changed into binary image by 2-4.；
2-5. filters out noise in binary image, then carries out Skeleton processing, and then to Skeleton, treated that image is gone
Burr processing, obtains fingerprint filament structural images；
2-6. finds the minutiae point of fingerprint filament structural images by Ridge following method, obtains the thin of fingerprint filament structural images
Node obtains the minutiae feature of fingerprint image.
3. the fingerprint identification method of a kind of combination minutiae point and filament structure according to claim 1 or 2, it is characterised in that
Step 3 is implemented as follows：
3-1. describes son according to minutiae point, and using the directional field information around minutiae point, it is on schedule right to obtain best base：
3-2. to best base on schedule to minutiae feature match；
3-3. to best base on schedule to fingerprint filament structure matching；
3-4. carries out Fusion Features using fractional layer information in decision-making level to minutiae feature and filament structure feature.
4. the fingerprint identification method of a kind of combination minutiae point and filament structure according to claim 3, it is characterised in that step
Rapid 3-1 is implemented as follows：
3-1-1. inputs two width fingerprint image A and B, A is fingerprint image to be identified, and the minutiae point of extraction is set A={ a1,
a2,···,an,···,aN}；B is the arbitrary fingerprint image in fingerprint base, and the minutiae point of extraction is set B={ b1,
b2,···,bm,,bM}；
3-1-2. chooses any one minutiae point a in fingerprint image An, all minutiae points in fingerprint image B are traversed, if fingerprint
There are minutiae point b in image Bm, and two minutiae point an、bmIt is identical to meet type, and position translation is in (± Δ x0,±Δy0) model
In enclosing, then 3-1-4 is entered step；If being not present and minutiae point a after traversing all minutiae points in fingerprint image BnIt is corresponding thin
Node bm, then the minutiae point a in fingerprint image A is abandonedn；
3-1-3. continues, from one minutiae point of selection in fingerprint image A, to repeat step 3-1-2, until having traversed fingerprint image A
In all minutiae points；
3-1-4. details of construction points description, description attached bag include three auxiliary magnets in the field of direction around minutiae point；Calculate three
Relative angular difference between being put two-by-two in a auxiliary magnet：
Δθk=| θi-θj| (i, j=0,1,2,3；I ＜ j) (1)
In formula (1), k is relative angular difference reference numeral in minutiae point description, in the range of 1≤k≤6；
Kth angular deviation is denoted as G (k) between two minutiae points:
In formula (2),For minutiae point anRelative angular difference,For minutiae point bmRelative angular difference；
3-1-5. judges whether two minutiae points are preliminary datum marks：If arbitrary G (k) is more than threshold value T in formula (2)1, illustrate at 2 points
It mismatches, return to step 3-1-2；Otherwise two minutiae points are preliminary datum mark；Preliminary benchmark point set Q is recorded, matching logarithm is L,
Position offset (Δ x between each pair of preliminary datum marki,Δyi,Δθi)；
3-1-6. is from preliminary datum mark to finding best base on schedule in collection Q；
The preliminary datum mark of set of computations is to concentrating any pair of match point (Δ xi,Δyi,Δθi) (i=1,2, L) and collection
Close the deviation d of remaining all match points in Q (Δ x, Δ y, Δ θ)i：
In formula (3), by diIn be less than optimal criteria point threshold value T2A numerical value form set D, maximum value C in selection set Dm,
Maximum value CmCorresponding match point be best base on schedule, i.e. aiAnd bjOn schedule right for best base, corresponding position deviation parameter is
(Δxij,Δyij,Δθij)。
5. the fingerprint identification method of a kind of combination minutiae point and filament structure according to claim 4, it is characterised in that step
Rapid 3-2 is implemented as follows：
3-2-1. concentrates the position of all the points to calibrate minutiae point, and formula is as follows：
θ '=θ+Δ θij (5)
Wherein, S0For change of scale parameter；Details in fingerprint point set A={ a to be identified1,a2,···,aNBy translation and rotation
(Δxij,Δyij,Δθij) it is transformed into A'={ a '1, a'2,···,a'N}；
3-2-2. calculating a 'iAnd bjPositional distance d (a 'i,bj) and field of direction differential seat angle dθ(a′i,bj)；If d (a 'i,bj) small
In positional distance empirical value Td, and dθ(a′i,bj) it is less than field of direction differential seat angle empirical value Tθ, then two minutiae matchings；
dθ(a′i,bj)=min (| θ 'i-θj|,360°-|θ′i-θj|) (7)
According to (6) (7) two formula, two matched minutiae point similarity s are calculated：
In formula k, σ are constants, and d is that two details points set distance, dθFor two Minutiae Direction field differential seat angles；
3-2-3. counts the number of whole match points, minutiae matching score is calculated as follows according to (8) formula：
Wherein, NMIt counts out for matching, N and M are that details in fingerprint to be known counts out and counts out with template details respectively.
6. the fingerprint identification method of a kind of combination minutiae point and filament structure according to claim 5, it is characterised in that step
Rapid 3-1 is implemented as follows：
3-3. fingerprint filament structure matchings
The filament structure point set of 3-3-1. fingerprint images A extraction isThe filament structure of fingerprint image B extractions
SetPoint setIn each put containing only location information (x, y), x, y indicate respectively the abscissa and
Ordinate；
3-3-2. is according to the transformation relation of formula (4), fingerprint filament structure point set to be identifiedBy translation
With rotation (Δ xij,Δyij,Δθij) be transformed intoAfter transformationWith library fingerprint
Compare to obtain two fingerprint overlapping region accounting T', if T' is more than presetting filament structure overlapping region area accounting threshold value
T3, then the matching of filament structure feature is carried out in overlapping region, judges whether fingerprint matches hence into step 3-3-3；It is no
Then illustrate that two fingerprint overlapping regions are too small, is not the fingerprint of same finger, is judged to mismatching；
3-3-3. utilizes filament structure distance using the recognition methods of field search filament structure distanceIt is used as two
The index of similarity of fingerprint filament structure, specific calculating are as follows：
Filament structure point setIt arrivesDistance definition：
By filament structure point setMiddle p point is to filament structure point setDistance be sorted in ascending order, i distance value puts down before taking
Mean value is as filament structure point setIt arrivesDistance value：
In formula (10), 1≤i≤p, p are filament structure point setThe number at midpoint, IthIndicate ascending sort, | | | | be
Between Euclidean distance；
Similarly, filament structure point setIt arrivesDistance definition：
By filament structure point setMiddle q point is to filament structure point setDistance be sorted in ascending order, j distance value puts down before taking
Mean value is as filament structure point setIt arrivesDistance value：
In formula, 1≤j≤q, q are point setThe number at midpoint, IthIndicate ascending sort, | | | | be Between Euclidean away from
From；
Filament structure point setWith filament structure point setThe matching score of straight line is as follows：
7. the fingerprint identification method of a kind of combination minutiae point and filament structure according to claim 3, it is characterised in that step
Rapid 3-4 is implemented as follows：
Fusion Features are carried out to minutiae feature and filament structure feature using fractional layer information in decision-making level, expression formula is such as
Under：
Wherein, α is weight, SHThe matching score being calculated for filament structure feature；
Matching result judgement is as follows：
Wherein c is the similarity threshold for the 2 width fingerprint images that experiment obtains, is considered inhomogeneity fingerprint image less than c, otherwise it is assumed that
It is similar fingerprint image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810524261.1A CN108416342B (en) | 2018-05-28 | 2018-05-28 | Fingerprint identification method combining thin node and thin line structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810524261.1A CN108416342B (en) | 2018-05-28 | 2018-05-28 | Fingerprint identification method combining thin node and thin line structure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108416342A true CN108416342A (en) | 2018-08-17 |
| CN108416342B CN108416342B (en) | 2022-02-18 |
Family
ID=63140647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810524261.1A Active CN108416342B (en) | 2018-05-28 | 2018-05-28 | Fingerprint identification method combining thin node and thin line structure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108416342B (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109376688A (en) * | 2018-11-20 | 2019-02-22 | 连云港市公安局 | A kind of fingerprint characteristic editor inquiry combination comparison method |
| CN109657657A (en) * | 2019-02-19 | 2019-04-19 | 武汉芯盈科技有限公司 | A kind of mobile phone small area high-precision fingerprint identification method and system based on image procossing |
| CN109657579A (en) * | 2018-12-07 | 2019-04-19 | 上海爱信诺航芯电子科技有限公司 | A kind of detection of fingerprint crackle and restorative procedure |
| CN109886212A (en) * | 2019-02-25 | 2019-06-14 | 清华大学 | From the method and apparatus of rolling fingerprint synthesis fingerprint on site |
| CN109934180A (en) * | 2019-03-18 | 2019-06-25 | Oppo广东移动通信有限公司 | Fingerprint identification method and relevant apparatus |
| CN110163123A (en) * | 2019-04-30 | 2019-08-23 | 杭州电子科技大学 | One kind referring to vein fusion identification method based on single width near-infrared finger-image fingerprint |
| WO2020047717A1 (en) * | 2018-09-03 | 2020-03-12 | 深圳市汇顶科技股份有限公司 | Multi-source fingerprint image enhancement and synthesis method and related fingerprint sensor |
| CN110956468A (en) * | 2019-11-15 | 2020-04-03 | 西安电子科技大学 | Fingerprint payment system |
| CN112784809A (en) * | 2021-02-05 | 2021-05-11 | 三星（中国）半导体有限公司 | Fingerprint identification method and fingerprint identification device |
| CN112949361A (en) * | 2019-12-11 | 2021-06-11 | 杭州萤石软件有限公司 | Fingerprint identification method and device |
| CN114863493A (en) * | 2022-07-06 | 2022-08-05 | 北京圣点云信息技术有限公司 | Detection method and detection device for low-quality fingerprint image and non-fingerprint image |
| CN115497125A (en) * | 2022-11-17 | 2022-12-20 | 上海海栎创科技股份有限公司 | Fingerprint identification method, system, computer equipment and computer readable storage medium |
| CN116258842A (en) * | 2023-05-16 | 2023-06-13 | 上海海栎创科技股份有限公司 | Fingerprint template dynamic splicing optimization system and method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101329727A (en) * | 2008-06-27 | 2008-12-24 | 哈尔滨工业大学 | Fingerprint identification method combining point with line |
| CN103714159A (en) * | 2013-12-27 | 2014-04-09 | 中国人民公安大学 | Coarse-to-fine fingerprint identification method fusing second-level and third-level features |
| US20160247012A1 (en) * | 2015-02-19 | 2016-08-25 | Persistent Systems Limited | Method and a System for Matching Fingerprint Images Obtained From Different Fingerprint Image Capturing Devices |
| CN107392847A (en) * | 2017-06-07 | 2017-11-24 | 西安电子科技大学 | A kind of fingerprint image joining method based on minutiae point and range image |
| CN107748877A (en) * | 2017-11-10 | 2018-03-02 | 杭州晟元数据安全技术股份有限公司 | A kind of Fingerprint recognition method based on minutiae point and textural characteristics |
| CN107909031A (en) * | 2017-11-15 | 2018-04-13 | 张威 | A kind of scene of a crime fingerprint ridge leaves region frequency dynamic reconstruction method |
-
2018
- 2018-05-28 CN CN201810524261.1A patent/CN108416342B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101329727A (en) * | 2008-06-27 | 2008-12-24 | 哈尔滨工业大学 | Fingerprint identification method combining point with line |
| CN103714159A (en) * | 2013-12-27 | 2014-04-09 | 中国人民公安大学 | Coarse-to-fine fingerprint identification method fusing second-level and third-level features |
| US20160247012A1 (en) * | 2015-02-19 | 2016-08-25 | Persistent Systems Limited | Method and a System for Matching Fingerprint Images Obtained From Different Fingerprint Image Capturing Devices |
| CN107392847A (en) * | 2017-06-07 | 2017-11-24 | 西安电子科技大学 | A kind of fingerprint image joining method based on minutiae point and range image |
| CN107748877A (en) * | 2017-11-10 | 2018-03-02 | 杭州晟元数据安全技术股份有限公司 | A kind of Fingerprint recognition method based on minutiae point and textural characteristics |
| CN107909031A (en) * | 2017-11-15 | 2018-04-13 | 张威 | A kind of scene of a crime fingerprint ridge leaves region frequency dynamic reconstruction method |
Non-Patent Citations (3)
| Title |
|---|
| GHADDAB M H ET AL.: "Fast and accurate fingerprint matching using expanded delaunay triangulation", 《IEEE/ACS 14TH INTERNATIONAL CONFERENCE ON COMPUTER SYSTEMS AND APPLICATIONS》 * |
| 田俊青: "自动指纹识别算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 田捷 等: "指纹识别技术的新进展", 《自然科学进展》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020047717A1 (en) * | 2018-09-03 | 2020-03-12 | 深圳市汇顶科技股份有限公司 | Multi-source fingerprint image enhancement and synthesis method and related fingerprint sensor |
| CN109376688A (en) * | 2018-11-20 | 2019-02-22 | 连云港市公安局 | A kind of fingerprint characteristic editor inquiry combination comparison method |
| CN109376688B (en) * | 2018-11-20 | 2021-10-22 | 连云港市公安局 | Fingerprint feature editing, inquiring, combining and comparing method |
| CN109657579A (en) * | 2018-12-07 | 2019-04-19 | 上海爱信诺航芯电子科技有限公司 | A kind of detection of fingerprint crackle and restorative procedure |
| CN109657579B (en) * | 2018-12-07 | 2023-06-09 | 上海航芯电子科技股份有限公司 | Fingerprint crack detection and repair method |
| CN109657657A (en) * | 2019-02-19 | 2019-04-19 | 武汉芯盈科技有限公司 | A kind of mobile phone small area high-precision fingerprint identification method and system based on image procossing |
| CN109886212A (en) * | 2019-02-25 | 2019-06-14 | 清华大学 | From the method and apparatus of rolling fingerprint synthesis fingerprint on site |
| CN109934180B (en) * | 2019-03-18 | 2021-06-01 | Oppo广东移动通信有限公司 | Fingerprint identification method and related device |
| CN109934180A (en) * | 2019-03-18 | 2019-06-25 | Oppo广东移动通信有限公司 | Fingerprint identification method and relevant apparatus |
| CN110163123A (en) * | 2019-04-30 | 2019-08-23 | 杭州电子科技大学 | One kind referring to vein fusion identification method based on single width near-infrared finger-image fingerprint |
| CN110163123B (en) * | 2019-04-30 | 2021-02-26 | 杭州电子科技大学 | Fingerprint finger vein fusion identification method based on single near-infrared finger image |
| CN110956468B (en) * | 2019-11-15 | 2023-05-23 | 西安电子科技大学 | Fingerprint payment system |
| CN110956468A (en) * | 2019-11-15 | 2020-04-03 | 西安电子科技大学 | Fingerprint payment system |
| CN112949361A (en) * | 2019-12-11 | 2021-06-11 | 杭州萤石软件有限公司 | Fingerprint identification method and device |
| CN112949361B (en) * | 2019-12-11 | 2023-07-25 | 杭州萤石软件有限公司 | Fingerprint identification method and device |
| CN112784809A (en) * | 2021-02-05 | 2021-05-11 | 三星（中国）半导体有限公司 | Fingerprint identification method and fingerprint identification device |
| CN114863493A (en) * | 2022-07-06 | 2022-08-05 | 北京圣点云信息技术有限公司 | Detection method and detection device for low-quality fingerprint image and non-fingerprint image |
| CN114863493B (en) * | 2022-07-06 | 2022-09-13 | 北京圣点云信息技术有限公司 | Detection method and detection device for low-quality fingerprint image and non-fingerprint image |
| CN115497125A (en) * | 2022-11-17 | 2022-12-20 | 上海海栎创科技股份有限公司 | Fingerprint identification method, system, computer equipment and computer readable storage medium |
| CN115497125B (en) * | 2022-11-17 | 2023-03-10 | 上海海栎创科技股份有限公司 | Fingerprint identification method, system, computer equipment and computer readable storage medium |
| CN116258842A (en) * | 2023-05-16 | 2023-06-13 | 上海海栎创科技股份有限公司 | Fingerprint template dynamic splicing optimization system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108416342B (en) | 2022-02-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108416342A (en) | A kind of fingerprint identification method of combination minutiae point and filament structure | |
| CN100414558C (en) | Automatic fingerprint distinguishing system and method based on template learning | |
| CN107748877B (en) | Fingerprint image identification method based on minutiae and textural features | |
| US5799098A (en) | Fingerprint identification system | |
| CN105389593B (en) | Image object recognition methods based on SURF feature | |
| Li et al. | Texture-based palmprint retrieval using a layered search scheme for personal identification | |
| CN101777128B (en) | Fingerprint minutiae matching method syncretized to global information and system thereof | |
| Chikkerur et al. | Impact of singular point detection on fingerprint matching performance | |
| CN104123537A (en) | Rapid authentication method based on handshape and palmprint recognition | |
| Cappelli et al. | A fingerprint retrieval system based on level-1 and level-2 features | |
| Feng et al. | A performance evaluation of fingerprint minutia descriptors | |
| CN114648445B (en) | Multi-view high-resolution point cloud splicing method based on feature point extraction and fine registration optimization | |
| CN114120378A (en) | Three-level classification fingerprint identification method | |
| Cao et al. | Minutia handedness: A novel global feature for minutiae-based fingerprint matching | |
| CN116704557A (en) | Low-quality fingerprint matching method based on texture information | |
| CN101833763A (en) | Method for detecting reflection image on water surface | |
| WO2016011640A1 (en) | Identification method based on handprint imaging | |
| He et al. | Modeling and analysis of local comprehensive minutia relation for fingerprint matching | |
| Jeyanthi et al. | Neural network based automatic fingerprint recognition system for overlapped latent images | |
| CN104951760B (en) | A kind of fingerprint classification method based on field of direction minutiae point description | |
| Fei et al. | Fingerprint singular points extraction based on the properties of orientation model | |
| CN115797987A (en) | Finger vein identification method based on joint loss and convolutional neural network | |
| Deshpande et al. | End-to-end automated latent fingerprint identification with improved DCNN-FFT enhancement | |
| Liu et al. | Fingerprint classification by ridgeline and singular point analysis | |
| JP2001076142A (en) | Method and device for collating fingerprint image and recording medium with this method recorded thereon |
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 |