CN109145810A - Details in fingerprint point detecting method, device, equipment, system and storage medium - Google Patents
Details in fingerprint point detecting method, device, equipment, system and storage medium Download PDFInfo
- Publication number
- CN109145810A CN109145810A CN201810941053.1A CN201810941053A CN109145810A CN 109145810 A CN109145810 A CN 109145810A CN 201810941053 A CN201810941053 A CN 201810941053A CN 109145810 A CN109145810 A CN 109145810A
- Authority
- CN
- China
- Prior art keywords
- fingerprint
- minutiae
- point
- image
- details
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/12—Fingerprints or palmprints
- G06V40/13—Sensors therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a kind of details in fingerprint point detecting methods, comprising: receives the fingerprint image collected;Fingerprint image is input to the fingerprint minutiae detection model of pre-training;Wherein, fingerprint minutiae detection model is the full convolutional neural networks obtained to sample fingerprint image based on convolutional neural networks training;It is taken the fingerprint the fingerprint minutiae feature of image by downsampling unit in fingerprint minutiae detection model;The location information that minutiae feature is rebuild by up-sampling unit in fingerprint minutiae detection model, obtains minutiae point location drawing picture.This method carries out detection positioning to fingerprint minutiae by full convolutional neural networks, and accuracy is higher, anti-interference ability and strong robustness.The invention also discloses a kind of details in fingerprint point detection device, system and a kind of readable storage medium storing program for executing, have above-mentioned beneficial effect.
Description
Technical field
The present invention relates to field of biological recognition, in particular to a kind of details in fingerprint point detecting method, device, equipment, system
And a kind of readable storage medium storing program for executing.
Background technique
In recent years, based on the biological identification technology based on fingerprint recognition using more and more extensive, including fingerprint access control, refer to
Line unlock etc., excellent fingerprint recognition and matching algorithm are particularly significant for the safety of information and certification.
Fingerprint minutiae detection is the basis of common fingerprint recognition and fingerprint matching algorithm in the market, and carries out fingerprint
One basic premise of structural matching algorithm.Currently used fingerprint minutiae detection algorithm is all based on to collected finger
Print image carries out traditional images Processing Algorithm (including image preprocessing, binaryzation, refinement, the extraction of characteristic details point) and obtains fingerprint
Minutiae point.
This algorithm process process is complicated, extracts result vulnerable to fingerprint fold, scratch, dry hand during fingerprint image acquisition
Refer to or many factors such as wet finger, image are fuzzy influence, anti-interference ability is weak under undesirable element, the details in fingerprint detected
Point feature precision is low.
Therefore, how accurate detection positioning is carried out to fingerprint minutiae, improves anti-interference ability and robustness, is ability
Field technique personnel's technical issues that need to address.
Summary of the invention
The object of the present invention is to provide a kind of details in fingerprint point detecting methods, and this method is by full convolutional neural networks to finger
Line minutiae point carries out detection positioning, and accuracy is higher, anti-interference ability and strong robustness;It is a further object of the present invention to provide
A kind of details in fingerprint point detection device, equipment, system and a kind of readable storage medium storing program for executing have above-mentioned beneficial effect.
The present invention provides a kind of details in fingerprint point detecting method, comprising:
Receive the fingerprint image collected;
The fingerprint image is input to the fingerprint minutiae detection model of pre-training;Wherein, the fingerprint minutiae inspection
Surveying model is the full convolutional neural networks obtained to sample fingerprint image based on convolutional neural networks training;
The fingerprint minutiae for extracting the fingerprint image by downsampling unit in the fingerprint minutiae detection model is special
Sign;
The location information that the minutiae feature is rebuild by up-sampling unit in the fingerprint minutiae detection model, obtains
To minutiae point location drawing picture.
Preferably, the details in fingerprint point detecting method further include:
It receives minutiae point position and extracts instruction, and the extraction of minutiae point position coordinates is carried out to the minutiae point location drawing picture.
Preferably, described to include: to the progress minutiae point position coordinates extraction of minutiae point location drawing picture
The pixel of the minutiae point location drawing picture is traversed, if the pixel value of current pixel point is greater than minimum valve
When being worth, and being greater than the pixel value of other pixels in current pixel point default neighborhood, using current pixel point as local maximum
Point;
Each local maximum point that will acquire carries out coordinate calculating, obtains the coordinate of each minutiae point.
Preferably, the details in fingerprint point detecting method further include:
Operation is filtered to the minutiae point location drawing picture, obtains filtered image;
Then the pixel of the minutiae point location drawing picture is traversed specifically: to the pixel of the filtered image
Point is traversed.
Preferably, the details in fingerprint point detecting method further include:
Gauss is carried out to the minutiae point location drawing picture according to the position coordinates of extraction and exports setting, generation is sat with minutiae point
It is designated as the Gaussian Profile circle minutiae point location drawing picture in the center of circle.
Preferably, the structure of the fingerprint minutiae detection model specifically:
It is connect after five layers of downsampling unit are serial with five layers of serial up-sampling unit.
The present invention discloses a kind of details in fingerprint point detection device, comprising:
Image receiving unit, for receiving the fingerprint image collected;
Image input units, for the fingerprint image to be input to the fingerprint minutiae detection model of pre-training;Wherein,
The fingerprint minutiae detection model is the full convolutional Neural net obtained to sample fingerprint image based on convolutional neural networks training
Network;
Minutiae detection unit, for extracting the fingerprint by downsampling unit in the fingerprint minutiae detection model
The fingerprint minutiae feature of image;
Location information reconstruction unit is described thin for being rebuild by up-sampling unit in the fingerprint minutiae detection model
The location information of node diagnostic obtains minutiae point location drawing picture.
The present invention discloses a kind of fingerprint minutiae detection device, comprising:
Memory, for storing computer program;
Processor, the step of details in fingerprint point detecting method is realized when for executing the computer program.
The present invention discloses a kind of fingerprint minutiae detection system, comprising:
Fingerprint collecting equipment is sent to fingerprint minutiae inspection for acquiring fingerprint image, and by collected fingerprint image
Measurement equipment;
The fingerprint minutiae detection device, for receiving the fingerprint image collected;The fingerprint image is inputted
To the fingerprint minutiae detection model of pre-training;Wherein, the fingerprint minutiae detection model is to be based on to sample fingerprint image
The full convolutional neural networks that convolutional neural networks training obtains;It is mentioned by downsampling unit in the fingerprint minutiae detection model
Take the fingerprint minutiae feature of the fingerprint image;By described in up-sampling unit reconstruction in the fingerprint minutiae detection model
The location information of minutiae feature obtains minutiae point location drawing picture.
The present invention discloses a kind of readable storage medium storing program for executing, and program is stored on the readable storage medium storing program for executing, and described program is located
The step of reason device realizes the details in fingerprint point detecting method when executing.
In order to solve the above technical problems, the present invention provides a kind of details in fingerprint point detecting method, the fingerprint image that will acquire
It is input to the fingerprint minutiae detection model of pre-training, fingerprint minutiae detection model is that convolution mind is based on to sample fingerprint image
The full convolutional neural networks obtained through network training, downsampling unit is by down-sampling to fingerprint image in full convolutional neural networks
Take the fingerprint minutiae feature, and up-sampling unit carries out feature locations reconstruction by up-sampling according to the minutiae feature of extraction can
To obtain the location drawing picture of accurate fingerprint minutiae.Detection positioning is carried out to fingerprint minutiae by full convolutional neural networks,
Compared to common detection methods, there is better anti-interference ability and robustness, especially for finger common in fingerprint detection
Line fold, scratch, wet finger have a more stable detection performance compared with traditional detection method when image is fuzzy, and can be with
Realize accurate detection positioning.
The invention also discloses a kind of details in fingerprint point detection device, equipment, system and a kind of readable storage medium storing program for executing, have
Above-mentioned beneficial effect, details are not described herein.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the flow chart of details in fingerprint point detecting method provided in an embodiment of the present invention;
Fig. 2 is downsampling unit structural schematic diagram provided in an embodiment of the present invention;
Fig. 3 is up-sampling unit structural schematic diagram provided in an embodiment of the present invention;
Fig. 4 is a kind of downsampling unit provided in an embodiment of the present invention, the channel for connecting combining unit and up-sampling unit
Number schematic diagram;
Fig. 5 is to search local maximum point schematic diagram in a kind of 3x3 neighborhood provided in an embodiment of the present invention;
Fig. 6 is a kind of maximum value detection mark schematic diagram provided in an embodiment of the present invention;
Fig. 7 is fingerprint minutiae structure of the detecting device block diagram provided in an embodiment of the present invention;
Fig. 8 is the structural block diagram of fingerprint minutiae detection device provided in an embodiment of the present invention;
Fig. 9 is the structural schematic diagram of fingerprint minutiae detection device provided in an embodiment of the present invention;
Figure 10 is the structural block diagram of fingerprint minutiae detection system provided in an embodiment of the present invention.
Specific embodiment
Core of the invention is to provide a kind of details in fingerprint point detecting method, and this method is by full convolutional neural networks to finger
Line minutiae point carries out detection positioning, and accuracy is higher, anti-interference ability and strong robustness;Another core of the invention is to provide
A kind of details in fingerprint point detection device, system and a kind of readable storage medium storing program for executing have above-mentioned beneficial effect.
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
Referring to FIG. 1, Fig. 1 is the flow chart of details in fingerprint point detecting method provided in an embodiment of the present invention;This method can
To include:
Step s110, the fingerprint image collected is received.
Fingerprint image can be collected by fingerprint collecting equipment, and fingerprint collecting process can refer to existing fingerprint collecting side
Method, details are not described herein.
Step s120, fingerprint image is input to the fingerprint minutiae detection model of pre-training.
The size of the input fingerprint image of detection model determines by model structure type and design parameter, general every kind of net
The all with good grounds actual conditions of network model are designed as the input of fixed size, and fingerprint minutiae detection model inputs fingerprint image can root
Appropriate cut again to network inputs, certainly, if network model is to picture size no requirement (NR) or acquisition is carried out according to network parameter
Picture size and mode input size require to match can be without cutting, the fingerprint image that directly will acquire be input to inspection
Model is surveyed, it is not limited here.
Fingerprint minutiae detection model is the full convolution mind obtained to sample fingerprint image based on convolutional neural networks training
Through network (FCN), FCN allows picture to realize that end to end is predicted using convolutional layer by study, carries out Pixel-level to image
Classification, to solve the image, semantic feature forecasting problem of semantic level.The method phase of image segmentation is carried out with CNN with tradition
Than FCN can receive the input picture of arbitrary size, and not have to that all training image and test image is required to have similarly
Size;In addition it because avoiding due to the problem of using block of pixels and bring repetition storage and calculating convolution, is carried out by FCN
Image procossing is more efficient.The present embodiment carries out the detection of fingerprint minutiae using FCN, has better anti-interference ability and Shandong
Stick.
Step s130, by downsampling unit in fingerprint minutiae detection model take the fingerprint image fingerprint minutiae it is special
Sign.
Step s140, the location information of minutiae feature is rebuild by up-sampling unit in fingerprint minutiae detection model,
Obtain minutiae point location drawing picture.
Without limitation to the specific constructive form of detection model, specifically, detection model can by several downsampling units,
It is formed with quantity up-sampling unit identical as downsampling unit.
Downsampling unit and the specific network structure of up-sampling unit are without limitation.If downsampling unit mainly may include
Dry convolutional layer and several regularization layers.Can also need to configure other layers according to data processing, for example, linear active coating and
Pond layer etc., the number of plies of each layer also without limitation, can be according to data demand data processing precision and data processing in network
Amount needs self-setting, for example can be set and connect after two convolutional layers are connect with a regularization layer with a linear active coating
It connects.Since the characteristic pattern size after down-sampling is much smaller than original input image size, in upper sampling process, it can introduce
Warp lamination rebuilds target detail, expands characteristic pattern size to input original image size.Up-sampling unit is under
Sampling unit is correspondingly arranged, and this will not be repeated here.
In addition, without limitation to the quantity of downsampling unit and up-sampling unit.For example several down-sampling lists can be set
Member connection after connect with several up-sampling units, be respectively used to more refine carry out fingerprint minutiae feature extract and
The reconstruction of minutiae feature position.The number of plies is more, and feature extraction more refines, but calculative data volume can greatly increase,
The quantity of specific downsampling unit and up-sampling unit can voluntarily be set as needed.Up-sampling unit and down-sampling list
The quantity of member is identical, and downsampling unit increases, and the receptive field of feature learning is also constantly expanding, the global position information of network
Ability to express is stronger.Network architecture can be voluntarily adjusted flexibly according to different Detection task and target.
Wherein it is preferred to connect after five layers of downsampling unit are serial with five layers of serial up-sampling unit by experimental verification
When connecing, the stronger while required computing resource of global characteristics detectability is relatively reasonable.
Other than downsampling unit and up-sampling unit, other can also be arranged helps to carry out data transmission detection model
With the unit of calculating, such as the connection combining unit etc. for carrying out data transmission.
Such as fingerprint minutiae detection model specifically can be by five downsampling units, five connection combining units and five
Up-sampling unit composition.Wherein, downsampling unit structure is as shown in Fig. 2, by one or more convolutional layers (conv3*3), optional
Batch regularization layer (BatchNormalization), linear active coating (ReLU or LeakyReLU) and pond layer
(maxpool 2*2) composition.Up-sampling unit structure is as shown in figure 3, by warp lamination (convTranspose), several convolution
Layer (conv3*3) and several linear active coatings (ReLU or LeakyReLU) composition, linear active coating connection warp lamination and volume
After lamination.The output of each downsampling unit and the output of up-sampling unit warp lamination are incorporated as down by connection combining unit
One layer of input.One downsampling unit is connect by a connection combining unit with same level up-sampling unit, behind can be with
Continue other downsampling unit of continued access next stage and so on, data are from the downsampling unit input in first layer, from last
The up-sampling unit output of layer.
Each convolutional layer port number is not fixed in model, can carry out appropriate adjustment, Fig. 4 institute according to learning objective feature complexity
The port number schematic diagram for being shown as a kind of up-sampling unit, connecting combining unit and downsampling unit, other channels settings can be joined
According to the figure, details are not described herein.
Only the fingerprint minutiae detection model in the present embodiment is introduced by taking above-mentioned network architecture as an example herein,
The detection of training and fingerprint minutiae to model is introduced based on structural schematic diagram shown in Fig. 4, other knots
The detection model of structure type can refer to the introduction of the present embodiment, and details are not described herein.
Network training process can use the preferred feature extracted in advance using the fingerprint image of acquisition as network inputs
Dot image data is as source target data, using input image data collection and destination image data collection as input to network model
It is trained, the partial statistics characteristic of fingerprint image may be learned in the convolution kernel of convolutional layer in model, can in a manner of shared weight
To reduce network parameter, the pond layer in downsampling unit is real by being maximized (or average value) to characteristic pattern regional area
Now extract the translation invariance of feature.
The fingerprint image of acquisition, can be directly defeated as network using the source images of acquisition in training process as source images
Enter.In order to increase the quantity of sample set, also for the diversity for increasing sample type, source images can also be subjected to image enhancement
(increasing distracter) is used as network inputs afterwards, so that network model is learnt the feature more general to fingerprint image, at this to network
The acquisition methods of mode input image are without limitation.Specifically, image enhancement is carried out for example to fingerprint image at random: increased random
Noise carries out fuzzification operation to image, random generation scratch is added to fingerprint image, fold, adjusts image light and shade etc. at random.
In training process, to input picture carry out enhancing be it is random, i.e., for same input fingerprint image, can choose difference
Training batch be added disturbance item.Input is referred to by establishing full convolutional neural networks model using depth learning technology
Print image is trained study, and network is made to have identification fingerprint characteristic minutiae point ability.
To model parameter training after the completion of, can fixed model network parameter, to test fingerprint image carry out minutiae point it is pre-
It surveys, simulated target is to export artificial mark using characteristic point position coordinate as the white circular (pixel value 255) in the center of circle, other regions
The filled black area image for being 0 for pixel value can actually be exported using characteristic point coordinate as the center of circle and target image radius
Similar to the round area image minutiae point location drawing picture of white, minutiae point is marked in the image, is and input test fingerprint image phase
With the gray level image of size, the gray level image of output is minutiae point location drawing picture, under in fingerprint minutiae detection model
Sampling unit takes the fingerprint the fingerprint minutiae feature of image;Up-sampling unit rebuilds the location information of minutiae feature, thus
Realize the prediction to the minutiae point position of fingerprint image.
In addition, the minutiae point position prediction image exported in above-described embodiment be using characteristic point coordinate as the center of circle and target
The round area image minutiae point location drawing picture of the similar white of image radius, center is clear, but details dot profile is not very
Obviously, to optimize minutiae point remarkable result, minutiae point position detection accuracy is improved, it is preferable that can sit according to the position of extraction
Mark carries out Gauss to minutiae point location drawing picture and exports setting, generates using minutiae point coordinate as the Gaussian Profile circle minutiae point in the center of circle
Location drawing picture, the Gaussian Profile circle minutiae point location drawing picture sharp outline of output, conveniently checks and calculates.
Based on the above-mentioned technical proposal, details in fingerprint point detecting method provided by the present embodiment, the fingerprint image that will acquire
It is input to the fingerprint minutiae detection model of pre-training, is that convolution is based on to sample fingerprint image in fingerprint minutiae detection model
The full convolutional neural networks that neural metwork training obtains, by downsampling unit in full convolutional neural networks by down-sampling to finger
Print image takes the fingerprint minutiae feature, and up-sampling unit carries out feature locations according to the minutiae feature of extraction by up-sampling
Rebuild available minutiae point location drawing picture.Detection positioning is carried out to fingerprint minutiae by full convolutional neural networks, compared to
Common detection methods, have better anti-interference ability and robustness, especially for fingerprint fold common in fingerprint detection,
Scratch, wet finger have more stable detection performance compared with traditional detection method when image is fuzzy, and may be implemented accurate
Detection positioning.Meanwhile network architecture can be adjusted flexibly according to different Detection task and target, be had wider
Applicability.
Based on the above embodiment, test fingerprint image is inputted into fingerprint minutiae detection model, then model output image is
As fingerprint minutiae forecast image.The position to minutiae point is generally required during utilizing to minutiae point location drawing picture
Confidence breath is compared, to realize the aspect ratio pair to fingerprint minutiae, it is preferable that can receive the extraction of minutiae point position and refers to
It enables, and position coordinates extraction is carried out to minutiae point location drawing picture.
The method extracted at this to fingerprint minutiae specific location coordinate without limitation, can refer to existing position and sit
Mark extracting method.Wherein it is preferred to which carrying out the extraction of minutiae point position coordinates to the fingerprint minutiae location drawing picture of output can lead to
It crosses and the pixel value of minutiae point location drawing picture is traversed, if the pixel value of current pixel point is greater than minimum threshold values, and is greater than and works as
In preceding pixel point default neighborhood when the pixel value of other pixels, using current pixel point as local maximum point;It will acquire
Each local maximum point carries out coordinate calculating, obtains the coordinate of each minutiae point, passes through minimum pixel threshold values and minimum neighborhood two
Restrictive condition extracts multiple local maximums, guarantees the uniqueness of characteristic point in a certain range, promotes detection accuracy.
Based on the above embodiment, before the pixel value to minutiae point location drawing picture traverses, for the details for avoiding output
Adverse effect of the fuzzy region to traversing result in point location drawing picture, it is preferable that minutiae point location drawing picture can be filtered
Operation, obtains filtered image.A filter function is defined, Gauss twice is carried out using fixed size gaussian filtering core and filters
Wave, reusing onesize filtering core progress, etching operation (generates image dot image pixel intensities according to network and radius is fixed twice
Adopted filtering core size).Then the pixel value of minutiae point location drawing picture is traversed specifically: to the pixel value of filtered image
It is traversed.Firstly, using filter function to output prediction details point image carry out one to filtering operation twice (number of operations according to
Depending on effect).Then, image pixel is traversed one by one, when current pixel value is greater than minimum threshold values, then with current pixel
Other pixel values compare one by one in minimum neighborhood, and Fig. 5 show a kind of interior schematic diagram for searching local maximum point of 3x3 neighborhood,
When only current pixel value is all larger than other pixel values in neighborhood, a part for just finding current pixel as one is most
Big value point, successively all local maximum points and coordinate of target image can be obtained in traversal, are illustrated in figure 6 a kind of maximum value
Detection mark schematic diagram, it is seen then that the method provided through this embodiment carries out minutiae point position and determines that accuracy rate is higher, detection effect
Fruit is obviously improved.
Details in fingerprint point detection device provided by the invention is introduced below, referring to FIG. 7, Fig. 7 is that the present invention is real
The structural block diagram of the details in fingerprint point detection device of example offer is provided;The apparatus may include: image receiving unit 700, image are defeated
Enter unit 701, minutiae detection unit 702 and location information reconstruction unit 703.
Wherein, image receiving unit 700 is mainly used for receiving the fingerprint image collected;
Image input units 701 are mainly used for for fingerprint image being input to the fingerprint minutiae detection model of pre-training;Its
In, fingerprint minutiae detection model is the full convolutional Neural net obtained to sample fingerprint image based on convolutional neural networks training
Network;
Minutiae detection unit 702 is mainly used for the figure that takes the fingerprint by downsampling unit in fingerprint minutiae detection model
The fingerprint minutiae feature of picture;
Location information reconstruction unit 703 is mainly used for rebuilding details by up-sampling unit in fingerprint minutiae detection model
The location information of point feature obtains minutiae point location drawing picture.
Details in fingerprint point detection device provided by the invention detects fingerprint minutiae by full convolutional neural networks
Positioning, accuracy is higher, anti-interference ability and strong robustness.
Preferably, details in fingerprint point detection device provided in this embodiment can be with further include: position coordinates extraction unit, it is main
It is used to receive minutiae point position and extracts instruction, receive minutiae point position and extract instruction, and minutiae point location drawing picture is carried out thin
Node location coordinate extracts.
Preferably, above-mentioned position coordinates extraction unit can specifically include:
Pixel value judgment sub-unit is traversed for the pixel value to minutiae point location drawing picture, if current pixel point
Pixel value is greater than minimum threshold values, and when being greater than the pixel value of other pixels in current pixel point default neighborhood, by current pixel
Point is used as local maximum point;
Coordinate computation subunit, each local maximum point for will acquire carry out coordinate calculating, obtain each minutiae point
Coordinate.
Preferably, can be in position coordinates extraction unit further include: filtering subunit, for minutiae point location drawing picture into
Row filtering operation obtains filtered image;Filtering subunit is connect with pixel value judgment sub-unit, then pixel value judgement is single
Member is specifically used for: traversing to the pixel value of filtered image.
Details in fingerprint point detection device provided in this embodiment can be with further include: Gauss setting unit, for according to extraction
Position coordinates to minutiae point location drawing picture carry out Gauss export setting, generate using minutiae point coordinate as the center of circle Gaussian Profile circle
Shape minutiae point location drawing picture.
It should be noted that each unit in details in fingerprint point detection device in the specific embodiment of the invention,
The course of work please refers to the corresponding specific embodiment of Fig. 1, and details are not described herein.
Fingerprint minutiae detection device provided by the invention is introduced below, specifically to fingerprint minutiae detection device
Introduction the step of can refer to above-mentioned details in fingerprint point detecting method, Fig. 8 is fingerprint minutiae provided in an embodiment of the present invention inspection
The structural block diagram of measurement equipment;The equipment may include:
Memory 800, for storing computer program;
Processor 801, when for executing computer program the step of realization details in fingerprint point detecting method.
Referring to FIG. 9, the structural schematic diagram of fingerprint minutiae detection device provided in an embodiment of the present invention, the detection device
Bigger difference can be generated because configuration or performance are different, may include one or more processors (central
Processing units, CPU) 322 (for example, one or more processors) and memory 332, one or more
Store the storage medium 330 (such as one or more mass memory units) of application program 342 or data 344.Wherein, it deposits
Reservoir 332 and storage medium 330 can be of short duration storage or persistent storage.The program for being stored in storage medium 330 may include
One or more modules (diagram does not mark), each module may include to the series of instructions behaviour in data processing equipment
Make.Further, central processing unit 322 can be set to communicate with storage medium 330, executes and deposits on detection device 301
Series of instructions operation in storage media 330.
Detection device 301 can also include one or more power supplys 326, one or more wired or wireless nets
Network interface 350, one or more input/output interfaces 358, and/or, one or more operating systems 341, such as
Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM etc..
Step in details in fingerprint point detecting method described in above figure 1 can be by the knot of fingerprint minutiae detection device
Structure is realized.
Fingerprint minutiae detection system provided in an embodiment of the present invention is introduced below, details in fingerprint described below
Point detection system can correspond to each other reference with above-described fingerprint minutiae detection device.
Figure 10 is the structural block diagram of fingerprint minutiae detection system provided in an embodiment of the present invention;The system may include:
Fingerprint collecting equipment 900 and fingerprint minutiae detection device 901.
Wherein, fingerprint collecting equipment 900 is mainly used for acquiring fingerprint image, and collected fingerprint image is sent to finger
Line minutiae detection equipment;
Fingerprint minutiae detection device 901 is mainly used for receiving the fingerprint image collected;Fingerprint image is input to
The fingerprint minutiae detection model of pre-training;Wherein, fingerprint minutiae detection model is that convolution mind is based on to sample fingerprint image
The full convolutional neural networks obtained through network training;It is taken the fingerprint image by downsampling unit in fingerprint minutiae detection model
Fingerprint minutiae feature;The location information of minutiae feature is rebuild by up-sampling unit in fingerprint minutiae detection model,
Obtain minutiae point location drawing picture.
Readable storage medium storing program for executing provided in an embodiment of the present invention is introduced below, readable storage medium storing program for executing described below with
Above-described details in fingerprint point detecting method can correspond to each other reference.
A kind of readable storage medium storing program for executing disclosed by the invention, is stored thereon with program, realizes and refers to when program is executed by processor
The step of line minutiae detection method.
It is apparent to those skilled in the art that for convenience and simplicity of description, the device of foregoing description,
The specific work process of equipment, storage medium and unit, can refer to corresponding processes in the foregoing method embodiment, herein no longer
It repeats.
In several embodiments provided by the present invention, it should be understood that disclosed device, system, storage medium and
Method may be implemented in other ways.For example, apparatus embodiments described above are merely indicative, for example, single
Member division, only a kind of logical function partition, there may be another division manner in actual implementation, for example, multiple units or
Component can be combined or can be integrated into another system, or some features can be ignored or not executed.Another point is shown
The mutual coupling, direct-coupling or communication connection shown or discussed can be through some interfaces, between device or unit
Coupling or communication connection are connect, can be electrical property, mechanical or other forms.
Unit may or may not be physically separated as illustrated by the separation member, shown as a unit
Component may or may not be physical unit, it can and it is in one place, or may be distributed over multiple networks
On unit.It can some or all of the units may be selected to achieve the purpose of the solution of this embodiment according to the actual needs.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of software functional units.
It, can if integrated unit is realized in the form of SFU software functional unit and when sold or used as an independent product
To be stored in a mobile terminal.Based on this understanding, technical solution of the present invention is substantially in other words to the prior art
The all or part of the part to contribute or the technical solution can be embodied in the form of software products, which deposits
It stores up in one storage medium, including some instructions are used so that a mobile terminal (can be mobile phone or tablet computer
Deng) execute all or part of the steps of each embodiment method of the present invention.And storage medium above-mentioned includes: USB flash disk, moves firmly
Disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM),
The various media that can store program code such as magnetic or disk.
Each embodiment is described in a progressive manner in specification, the highlights of each of the examples are with other realities
The difference of example is applied, the same or similar parts in each embodiment may refer to each other.For device disclosed in embodiment
Speech, since it is corresponded to the methods disclosed in the examples, so being described relatively simple, related place is referring to method part illustration
?.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, it can be realized with the combination of electronic hardware, terminal or the two, in order to clearly demonstrate hardware and software
Interchangeability generally describes each exemplary composition and step according to function in the above description.These functions are studied carefully
Unexpectedly it is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technique people
Member can use different methods to achieve the described function each specific application, but this realization is it is not considered that super
The scope of the present invention out.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
Above to details in fingerprint point detecting method, device, equipment, system and readable storage medium storing program for executing provided by the present invention into
It has gone and has been discussed in detail.Used herein a specific example illustrates the principle and implementation of the invention, the above implementation
The explanation of example is merely used to help understand method and its core concept of the invention.It should be pointed out that for the general of the art
, without departing from the principle of the present invention, can be with several improvements and modifications are made to the present invention for logical technical staff, this
A little improvement and modification are also fallen within the protection scope of the claims of the present invention.
Claims (10)
1. a kind of details in fingerprint point detecting method characterized by comprising
Receive the fingerprint image collected;
The fingerprint image is input to the fingerprint minutiae detection model of pre-training;Wherein, the fingerprint minutiae detects mould
Type is the full convolutional neural networks obtained to sample fingerprint image based on convolutional neural networks training;
The fingerprint minutiae feature of the fingerprint image is extracted by downsampling unit in the fingerprint minutiae detection model;
The location information that the minutiae feature is rebuild by up-sampling unit in the fingerprint minutiae detection model obtains thin
Node location image.
2. details in fingerprint point detecting method as described in claim 1, which is characterized in that further include:
It receives minutiae point position and extracts instruction, and the extraction of minutiae point position coordinates is carried out to the minutiae point location drawing picture.
3. details in fingerprint point detecting method as claimed in claim 2, which is characterized in that described to be carried out to minutiae point location drawing picture
Minutiae point position coordinates extract
The pixel of the minutiae point location drawing picture is traversed, if the pixel value of current pixel point is greater than minimum threshold values, and
Greater than in current pixel point default neighborhood when the pixel value of other pixels, using current pixel point as local maximum point;
Each local maximum point that will acquire carries out coordinate calculating, obtains the coordinate of each minutiae point.
4. details in fingerprint point detecting method as claimed in claim 3, which is characterized in that further include:
Operation is filtered to the minutiae point location drawing picture, obtains filtered image;
Then the pixel of the minutiae point location drawing picture is traversed specifically: the pixel of the filtered image is clicked through
Row traversal.
5. details in fingerprint point detecting method as claimed in claim 2, which is characterized in that further include:
Gauss is carried out to the minutiae point location drawing picture according to the position coordinates of extraction and exports setting, generation is with minutiae point coordinate
The Gaussian Profile circle minutiae point location drawing picture in the center of circle.
6. details in fingerprint point detecting method as described in claim 1, which is characterized in that the fingerprint minutiae detection model
Structure specifically:
It is connect after five layers of downsampling unit are serial with five layers of serial up-sampling unit.
7. a kind of details in fingerprint point detection device characterized by comprising
Image receiving unit, for receiving the fingerprint image collected;
Image input units, for the fingerprint image to be input to the fingerprint minutiae detection model of pre-training;Wherein, described
Fingerprint minutiae detection model is the full convolutional neural networks obtained to sample fingerprint image based on convolutional neural networks training;
Minutiae detection unit, for extracting the fingerprint image by downsampling unit in the fingerprint minutiae detection model
Fingerprint minutiae feature;
Location information reconstruction unit, for rebuilding the minutiae point by up-sampling unit in the fingerprint minutiae detection model
The location information of feature obtains minutiae point location drawing picture.
8. a kind of fingerprint minutiae detection device characterized by comprising
Memory, for storing computer program;
Processor realizes that fingerprint minutiae detects as described in any one of claim 1 to 6 when for executing the computer program
The step of method.
9. a kind of fingerprint minutiae detection system characterized by comprising
Collected fingerprint image for acquiring fingerprint image, and is sent to fingerprint minutiae detection and set by fingerprint collecting equipment
It is standby;
The fingerprint minutiae detection device, for receiving the fingerprint image collected;The fingerprint image is input to pre-
Trained fingerprint minutiae detection model;Wherein, the fingerprint minutiae detection model is to be based on convolution to sample fingerprint image
The full convolutional neural networks that neural metwork training obtains;Institute is extracted by downsampling unit in the fingerprint minutiae detection model
State the fingerprint minutiae feature of fingerprint image;The details is rebuild by up-sampling unit in the fingerprint minutiae detection model
The location information of point feature obtains minutiae point location drawing picture.
10. a kind of readable storage medium storing program for executing, which is characterized in that be stored with program on the readable storage medium storing program for executing, described program is located
It manages and is realized when device executes as described in any one of claim 1 to 6 the step of details in fingerprint point detecting method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810941053.1A CN109145810A (en) | 2018-08-17 | 2018-08-17 | Details in fingerprint point detecting method, device, equipment, system and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810941053.1A CN109145810A (en) | 2018-08-17 | 2018-08-17 | Details in fingerprint point detecting method, device, equipment, system and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109145810A true CN109145810A (en) | 2019-01-04 |
Family
ID=64790229
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810941053.1A Pending CN109145810A (en) | 2018-08-17 | 2018-08-17 | Details in fingerprint point detecting method, device, equipment, system and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109145810A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113011284A (en) * | 2021-03-01 | 2021-06-22 | 杭州景联文科技有限公司 | Signature fingerprint identification method |
CN114677714A (en) * | 2022-03-30 | 2022-06-28 | 西安交通大学 | Method for matching contact fingerprint with non-contact fingerprint and related device |
WO2022236877A1 (en) * | 2021-05-14 | 2022-11-17 | 广州广电运通金融电子股份有限公司 | Deep learning-based fingerprint texture extraction method, system, and apparatus, and storage medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101408939A (en) * | 2008-10-21 | 2009-04-15 | 成都方程式电子有限公司 | Method and system for extracting fingerprint thin node in fingerprint image |
CN104156707A (en) * | 2014-08-14 | 2014-11-19 | 深圳市汇顶科技股份有限公司 | Fingerprint identification method and fingerprint identification device |
CN106056095A (en) * | 2016-06-13 | 2016-10-26 | 北京小米移动软件有限公司 | Fingerprint processing method and device |
CN106202997A (en) * | 2016-06-29 | 2016-12-07 | 四川大学 | A kind of cell division detection method based on degree of depth study |
CN107480649A (en) * | 2017-08-24 | 2017-12-15 | 浙江工业大学 | Fingerprint sweat pore extraction method based on full convolution neural network |
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 |
-
2018
- 2018-08-17 CN CN201810941053.1A patent/CN109145810A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101408939A (en) * | 2008-10-21 | 2009-04-15 | 成都方程式电子有限公司 | Method and system for extracting fingerprint thin node in fingerprint image |
CN104156707A (en) * | 2014-08-14 | 2014-11-19 | 深圳市汇顶科技股份有限公司 | Fingerprint identification method and fingerprint identification device |
CN106056095A (en) * | 2016-06-13 | 2016-10-26 | 北京小米移动软件有限公司 | Fingerprint processing method and device |
CN106202997A (en) * | 2016-06-29 | 2016-12-07 | 四川大学 | A kind of cell division detection method based on degree of depth study |
CN107480649A (en) * | 2017-08-24 | 2017-12-15 | 浙江工业大学 | Fingerprint sweat pore extraction method based on full convolution neural network |
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 |
Non-Patent Citations (4)
Title |
---|
DINH-LUAN NGUYEN ET AL: "Robust Minutiae Extractor:", 《RESEARCHGATE》 * |
OLAF RONNEBERGER ET AL: "U-Net: Convolutional Networks for Biomedical", 《ARXIV:1505.04597V1》 * |
YAO TANG ET AL: "FingerNet: An Unified Deep Network for Fingerprint Minutiae Extraction", 《2017 IEEE INTERNATIONAL JOINT CONFERENCE ON BIOMETRICS (IJCB)》 * |
YAO TANG ET AL: "Latent fingerprint minutia extraction using fully", 《ARXIV:1609.09850V1》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113011284A (en) * | 2021-03-01 | 2021-06-22 | 杭州景联文科技有限公司 | Signature fingerprint identification method |
CN113011284B (en) * | 2021-03-01 | 2023-05-26 | 杭州景联文科技有限公司 | Signature fingerprint identification method |
WO2022236877A1 (en) * | 2021-05-14 | 2022-11-17 | 广州广电运通金融电子股份有限公司 | Deep learning-based fingerprint texture extraction method, system, and apparatus, and storage medium |
CN114677714A (en) * | 2022-03-30 | 2022-06-28 | 西安交通大学 | Method for matching contact fingerprint with non-contact fingerprint and related device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
KR102109372B1 (en) | Apparatus and method for segmenting of semantic image using fully convolutional neural network based on multi scale image and multi scale dilated convolution | |
CN106709461B (en) | Activity recognition method and device based on video | |
CN103942577B (en) | Based on the personal identification method for establishing sample database and composite character certainly in video monitoring | |
CN108960214A (en) | Fingerprint enhancement binarization method, device, equipment, system and storage medium | |
CN109740413A (en) | Pedestrian recognition methods, device, computer equipment and computer storage medium again | |
CN107506761A (en) | Brain image dividing method and system based on notable inquiry learning convolutional neural networks | |
CN105512676A (en) | Food recognition method at intelligent terminal | |
CN107563328A (en) | A kind of face identification method and system based under complex environment | |
CN108921019A (en) | A kind of gait recognition method based on GEI and TripletLoss-DenseNet | |
CN110070029A (en) | A kind of gait recognition method and device | |
CN109117897A (en) | Image processing method, device and readable storage medium storing program for executing based on convolutional neural networks | |
CN110288597A (en) | Wireless capsule endoscope saliency detection method based on attention mechanism | |
CN110097029B (en) | Identity authentication method based on high way network multi-view gait recognition | |
CN105335716A (en) | Improved UDN joint-feature extraction-based pedestrian detection method | |
CN109145810A (en) | Details in fingerprint point detecting method, device, equipment, system and storage medium | |
CN109145717A (en) | A kind of face identification method of on-line study | |
CN109446889A (en) | Object tracking method and device based on twin matching network | |
CN110503613A (en) | Based on the empty convolutional neural networks of cascade towards removing rain based on single image method | |
CN109598225A (en) | Sharp attention network, neural network and pedestrian's recognition methods again | |
CN109255296A (en) | A kind of daily Human bodys' response method based on depth convolutional neural networks | |
CN114399480A (en) | Method and device for detecting severity of vegetable leaf disease | |
CN112733581B (en) | Vehicle attribute identification method and system | |
CN106778910A (en) | Deep learning system and method based on local training | |
CN112464690A (en) | Living body identification method, living body identification device, electronic equipment and readable storage medium | |
Yang et al. | A robust iris segmentation using fully convolutional network with dilated convolutions |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: 523710, 26, 188 Industrial Road, Pingshan Town, Guangdong, Dongguan, Tangxia Applicant after: Entropy Technology Co.,Ltd. Address before: 523710, 26, 188 Industrial Road, Pingshan Town, Guangdong, Dongguan, Tangxia Applicant before: ZKTECO Co.,Ltd. |
|
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190104 |