CN115439891A - Fingerprint identification method based on small fingerprint head, low computational power and low memory chip - Google Patents

Fingerprint identification method based on small fingerprint head, low computational power and low memory chip Download PDF

Info

Publication number
CN115439891A
CN115439891A CN202211016516.6A CN202211016516A CN115439891A CN 115439891 A CN115439891 A CN 115439891A CN 202211016516 A CN202211016516 A CN 202211016516A CN 115439891 A CN115439891 A CN 115439891A
Authority
CN
China
Prior art keywords
fingerprint
image
low
computational power
points
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
Application number
CN202211016516.6A
Other languages
Chinese (zh)
Inventor
朱泽
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Individual
Original Assignee
Individual
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Individual filed Critical Individual
Priority to CN202211016516.6A priority Critical patent/CN115439891A/en
Publication of CN115439891A publication Critical patent/CN115439891A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/12Fingerprints or palmprints
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/12Fingerprints or palmprints
    • G06V40/1347Preprocessing; Feature extraction
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06VIMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/12Fingerprints or palmprints
    • G06V40/1365Matching; Classification

Landscapes

  • Engineering & Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Collating Specific Patterns (AREA)

Abstract

The invention discloses a fingerprint identification method based on a small fingerprint head, low computational power and a low memory chip, which comprises the following steps: the image processing, namely dividing and normalizing the collected fingerprint image, and solving a direction field and a frequency field of the fingerprint image; extracting features, namely performing binarization and thinning processing on the fingerprint image after image processing, and filtering out unnecessary feature points; and (3) fingerprint matching, namely judging whether two points are matched or not by comprehensively calculating the positions and directions between the characteristic points and the included angles between the ridge lines and the direction field, and finally judging whether two fingerprints belong to the same finger or not by the number of points in matching. The invention not only improves the accuracy of fingerprint identification, but also reduces the computational power of fingerprint identification, and can be operated on low computational power and low storage chips.

Description

Fingerprint identification method based on small fingerprint head, low computational power and low memory chip
Technical Field
The invention relates to the field of fingerprint identification, in particular to a fingerprint identification method based on a small fingerprint head, low computational power and a low memory chip.
Background
The fingerprint identification algorithm is widely applied to identity detection nowadays due to the characteristics of stability and uniqueness of fingerprints, and has clear regulation requirements on fingerprint acquisition area to ensure accuracy in serious applications, and the acquisition area of a fingerprint sensor is continuously reduced along with the increase of the demand of a small-size fingerprint identification device, so that the effective information for fingerprint matching is reduced; meanwhile, adverse conditions such as skin cracks and stains affect the fingerprint image, so that a small-area fingerprint collector puts higher requirements on a fingerprint identification algorithm.
Disclosure of Invention
The invention aims to provide a fingerprint identification method based on a small fingerprint head, low computational complexity and a low storage chip, which improves the accuracy of fingerprint identification, reduces the computational complexity of fingerprint identification and can operate on the low computational complexity and the low storage chip.
In order to solve the above technical problems, the present invention provides a fingerprint identification method based on a small fingerprint head, low computational power and low memory chip, comprising the steps of:
the image processing, namely dividing and normalizing the collected fingerprint image, and solving a direction field and a frequency field of the fingerprint image;
extracting features, namely performing binarization and thinning processing on the fingerprint image after image processing, and filtering out unnecessary feature points;
and fingerprint matching, namely judging whether two points are matched or not by comprehensively calculating the positions and the directions between the characteristic points and the included angles between the ridge lines and the direction field, and finally judging whether two fingerprints belong to the same finger or not by the number of matched points.
Further, in the image processing step, the foreground and the background of the fingerprint image are separated after the normalization processing.
Furthermore, in the image processing step, the direction field and the frequency field of the fingerprint image are calculated, and then image enhancement is carried out through a filtering algorithm.
Further, the filtering algorithm adopts a cross filtering algorithm.
Further, in the step of feature extraction, binarization processing is carried out by utilizing the distribution condition of ridge lines and valley lines around the test pixel points.
Furthermore, in the step of feature extraction, the broken lines of the fingerprint image are used for filtering out the unnecessary feature points.
Compared with the prior art, the invention at least has the following beneficial effects:
the invention optimizes and screens the feature point extraction aiming at the collected fingerprint image, greatly improves the efficiency and the accuracy of fingerprint identification, and in addition, the invention writes part of codes using functions in an arm library by using assembly language, greatly reduces the time for storage, calling and calculation under the condition of ensuring the unchanged effect, and can operate on a low-computation and low-storage chip.
Drawings
FIG. 1 is a flow chart of a fingerprint identification method based on a small fingerprint head, low computational power and a low memory chip according to an embodiment of the present invention.
Detailed Description
The fingerprint identification method based on a small fingerprint head, low computational power and low memory chip of the present invention will be described in more detail with reference to the schematic drawings, in which preferred embodiments of the present invention are shown, it being understood that a person skilled in the art may modify the invention described herein while still achieving the advantageous effects of the present invention. Accordingly, the following description should be construed as broadly as possible to those skilled in the art and not as limiting the invention.
The invention is more particularly described in the following paragraphs with reference to the accompanying drawings by way of example. Advantages and features of the present invention will become apparent from the following description and from the claims. It is to be noted that the drawings are in a very simplified form and are not to precise scale, which is merely for the purpose of facilitating and distinctly claiming the embodiments of the present invention.
As shown in fig. 1, an embodiment of the present invention provides a fingerprint identification method based on a small fingerprint head, low computational power and a low memory chip, including the following steps:
the method comprises the following steps: and image processing, namely segmenting and normalizing the acquired fingerprint image, and solving a direction field and a frequency field of the fingerprint image.
Specifically, in the image processing process, the image segmentation, the normalization processing, and the solution of the direction field and the frequency field of the fingerprint image are included, and each processing process is described in detail below.
Image segmentation, an important step of an online fingerprint identification algorithm, is the segmentation of fingerprint images. The effective area including the ridge line and the valley line is separated from the fingerprint image to be segmented by the fingerprint image segmentation, so that the standard of fingerprint identification is achieved, the calculation complexity is reduced, and the processing speed of a fingerprint identification system is improved. The goal of fingerprint image segmentation is to segment the foreground region of the fingerprint to avoid extracting features in the noise and background regions.
And normalization processing, namely performing normalization processing on the acquired fingerprint image, namely adjusting the mean value and the variance of the gray level of the fingerprint gray level image once, so that the fingerprint image acquired by any equipment can have the expected variance and mean value, thereby shielding unnecessary noise. Fingerprint normalization does not change the quality of the fingerprint, but facilitates subsequent processing of the fingerprint and ensures accelerated convergence when the program runs.
And carrying out foreground and background separation on the fingerprint image after normalization processing. The gradient of the foreground area of the fingerprint image is generally larger, the black and white lines are clear and alternate, the gradient of the background area is smaller, the obtained gradient is highly smoothed, the noise of the gradient image is removed, and then the point can be judged to be the foreground or the background by using a proper threshold value. Effective information of the fingerprint image can be screened out through foreground and background separation.
The direction field and the frequency field of the fingerprint image are obtained, the directionality and the texture of the fingerprint are strong, and the fingerprint can be regarded as a fluid model. The direction field describes the tangential direction of a pixel point in the fingerprint image and a ridge line or a valley line where a small fingerprint is located, and because the fingerprint directions of the fingerprint image in a small area are almost the same, the direction of the pixel point is generally replaced by the direction of the small block where the point is located in the calculation. And the frequency field of the fingerprint image reflects the ridge spacing of the fingerprint. Along the direction of the ridges and valleys, the intensity of the fingerprint image can be modeled as a sine-like wave. The two are combined, and the collected fingerprint image information can be accurately reflected.
And carrying out image enhancement after solving the direction field and the frequency field of the fingerprint image. The cross filtering algorithm which is a simple algorithm based on the traditional Gabor filtering algorithm is used, and the cross filtering algorithm has the characteristics that the effect similar to that of the Gabor filtering algorithm can be achieved, but the required calculated amount and the required calculated time are greatly reduced. In the invention, the cross filtering algorithm achieves the precision similar to the original precision by utilizing the segmentation calculation in a very limited space, and greatly reduces the calculation amount of fingerprint identification on the premise of ensuring the calculation precision.
Step two: and (4) feature extraction, namely performing binarization and thinning processing on the fingerprint image after image processing, and filtering out unnecessary feature points.
Specifically, in the feature extraction process, the binarization process and the thinning process are included, and each process is described in detail below.
And (3) binarization processing, namely, the gray value of a point on the image is 0 or 255, namely, the whole image presents an obvious black-white effect, so that ridges and valleys in the fingerprint image are effectively highlighted. In the image binarization process, the method does not use the conventional threshold value to carry out binarization judgment, but carries out comprehensive judgment by using the conditions of ridge lines and valley lines around the test pixel points according to the characteristics of the fingerprint image, thereby achieving the effect better than the effect of carrying out binarization through the threshold value.
For example, gray values of eight directions around the current cell are calculated, the maximum value and the minimum value in the eight directions are obtained through comparison, and the average value of the gray values in the eight directions is calculated. And adding the maximum value, the minimum value and the gray value of the quadruple current point to obtain the test index of the current point. If the test index of the current point is less than three times the average value of the directional grays, we consider that the point is most likely on the ridge line. To prevent the algorithm from erroneously determining a pure white or dirty region as a ridge, further determination is required. If the maximum value in the eight directions is larger than the average value in the eight directions, we can determine that the current point is not in a pure white area, and in addition, to prevent some gray areas from being determined as a ridge line, we use a threshold to determine whether the current point is a point on the ridge line. By combining the above three factors, we can tell whether the current point is a ridge line or a valley line or a background.
Two binarization schemes are established according to the average frequency of the fingerprint, and when the ridge line is too thick, the range of direction calculation is expanded to prevent the appearance of a hole in the ridge line.
And thinning, wherein thinning is short for a process of reducing the lines of the image from the multi-pixel width to the unit pixel width. The skeleton is used for representing the line image, so that the data volume can be effectively reduced, and the storage difficulty and the identification difficulty of the image can be reduced. The calculated amount of fingerprint identification can be reduced, and the efficiency of fingerprint identification is improved.
In the feature extraction link, after the feature extraction is finished, the targeted processing is carried out on the feature abnormal dense area. Because the phenomenon that the number of feature points in the same area is too large is difficult to occur in the fingerprint, the feature is utilized to judge the position of the broken fingerprint in the fingerprint and filter the unnecessary feature points to achieve the purpose of filtering the feature points. The fingerprint template screened by the characteristic points needs to occupy smaller space, so that the matching speed can be greatly increased, and the hardware condition required by using the chip provided by the invention is reduced.
Step three: and fingerprint matching, namely judging whether two points are matched or not by comprehensively calculating the positions and the directions between the characteristic points and the included angles between the ridge lines and the direction field, and finally judging whether two fingerprints belong to the same finger or not by the number of matched points.
Wherein, the comprehensive calculation score is a result obtained by carrying out formula calculation on the optimal parameters obtained after multiple parameter adjustment and testing.
sim=(parameter1–diff_orient)*parameter2
+(parameter3–diff_distance)*parameter4
+(parameter5–diff_phi)*parameter6
+(parameter7–diff_theta)*parameter8
The parameters 1, 2, 3, 4, 5, 6, 7 and 8 are obtained by debugging.
Wherein, parameter1, parameter3, parameter5 and parameter7 are transformation parameters of the direction field difference, the polar coordinate position distance difference, the polar coordinate position angle difference and the ridge line included angle difference of the object to be matched and the template respectively; the parameters 2, 4, 6 and 8 are weights given after the direction field difference, the polar coordinate position distance difference, the polar coordinate position angle difference and the ridge line included angle difference of the object to be matched and the template are transformed respectively.
diff _ origin, diff _ distance, diff _ phi and diff _ theta are respectively a direction field difference value, a polar coordinate position distance difference value, a polar coordinate position angle difference value and a ridge line included angle difference value which are obtained after the object to be matched and the template are transformed by rotation, translation and the like.
In consideration of the universality and diversity of the fingerprint identification use group, different scoring standards are adopted aiming at different application objects.
For the population with large age and dry skin, the broken lines on the fingerprint are more likely to appear. Aiming at fingerprint images with more broken lines on the collected fingerprint, the broken line matching is preferentially carried out in the fingerprint matching link. The broken line matching broken line point number capable of well representing the broken line characteristics is used as the main evaluation content of the first scoring standard.
For a part of groups with few fingerprint feature points, the conventional matching process cannot acquire sufficiently accurate feature point information. And (3) storing the number of the characteristic points into joint matching in the link of matching the fingerprint with each template aiming at the fingerprint image with less number of the characteristic points on the collected fingerprint. And the comprehensive result of matching the fingerprints with the templates is used as the main evaluation content of the second grading standard.
There is also a wide difference in fingerprint image characteristics beyond the two cases described above. The scoring standards are different for different feature points and frequency fields of the image for collecting the fingerprint.
In particular, in order to minimize the storage and computation requirements, the codes of functions partially using the arm library are written in assembly language, so that the storage, calling and computation time is greatly reduced under the condition of ensuring that the effect is not changed, and the time cost of operating by using the arm library is reduced by 39% (shown in a table below).
Time consuming before assembly language optimization (ms) Time consuming (ms) after assembly language optimization
Test code 1 5187.98 3156.44
Test code 2 5175.11 3157.38
Test code 3 5174.35 3156.42
Test code 4 5167.6 3156.43
Average calculation 5176.26 3156.67
Compared with the prior art, the invention at least has the following beneficial effects:
the invention optimizes and screens the feature point extraction aiming at the acquired fingerprint image, greatly improves the efficiency and the accuracy of fingerprint identification, and in addition, the invention writes part of codes using functions in an arm library by using assembly language, greatly reduces the time of storage, calling and calculation under the condition of ensuring the unchanged effect, and can run on a low-computation and low-storage chip.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (6)

1. A fingerprint identification method based on a small fingerprint head, low computational power and low memory chip is characterized by comprising the following steps:
the image processing, namely dividing and normalizing the collected fingerprint image, and solving a direction field and a frequency field of the fingerprint image;
extracting characteristics, namely performing binarization and thinning processing on the fingerprint image after image processing, and filtering out unnecessary characteristic points;
and fingerprint matching, namely judging whether two points are matched or not by comprehensively calculating the positions and the directions between the characteristic points and the included angles between the ridge lines and the direction field, and finally judging whether two fingerprints belong to the same finger or not by the number of matched points.
2. The fingerprint identification method based on small fingerprint head, low computational power and low memory chip as claimed in claim 1 wherein in the image processing step, the foreground and background separation of the fingerprint image is performed after the normalization process.
3. The fingerprint recognition method based on small fingerprint head, low computational power and low memory chip as claimed in claim 1, wherein in the image processing step, the direction field and frequency field of the fingerprint image are solved and then image enhancement is performed by filtering algorithm.
4. The method for small fingerprint head, low computational power and low memory chip based fingerprint identification according to claim 3 wherein said filtering algorithm employs a cross filter algorithm.
5. The fingerprint identification method based on small fingerprint heads, low computational power and low memory chips as claimed in claim 1, wherein in the feature extraction step, the binarization process is performed using the distribution of the ridges and valleys around the test pixel points.
6. The fingerprint identification method based on small fingerprint head, low computational power and low memory chip as claimed in claim 1, wherein in the feature extraction step, the unwanted feature points are filtered out by using the broken lines of the fingerprint image.
CN202211016516.6A 2022-08-24 2022-08-24 Fingerprint identification method based on small fingerprint head, low computational power and low memory chip Pending CN115439891A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211016516.6A CN115439891A (en) 2022-08-24 2022-08-24 Fingerprint identification method based on small fingerprint head, low computational power and low memory chip

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211016516.6A CN115439891A (en) 2022-08-24 2022-08-24 Fingerprint identification method based on small fingerprint head, low computational power and low memory chip

Publications (1)

Publication Number Publication Date
CN115439891A true CN115439891A (en) 2022-12-06

Family

ID=84244172

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211016516.6A Pending CN115439891A (en) 2022-08-24 2022-08-24 Fingerprint identification method based on small fingerprint head, low computational power and low memory chip

Country Status (1)

Country Link
CN (1) CN115439891A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117218691A (en) * 2023-11-09 2023-12-12 四川酷比通信设备有限公司 Unlocking method based on fingerprint identification

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117218691A (en) * 2023-11-09 2023-12-12 四川酷比通信设备有限公司 Unlocking method based on fingerprint identification
CN117218691B (en) * 2023-11-09 2024-01-26 四川酷比通信设备有限公司 Unlocking method based on fingerprint identification

Similar Documents

Publication Publication Date Title
CN107862282B (en) Finger vein identification and security authentication method, terminal and system
CN107748877B (en) Fingerprint image identification method based on minutiae and textural features
WO2021227331A1 (en) Finger vein machine learning recognition method and apparatus based on terrain concave and convex chjaracteristic
CN109919960B (en) Image continuous edge detection method based on multi-scale Gabor filter
CN108256456A (en) A kind of finger vein identification method based on multiple features Threshold Fusion
CN101114335A (en) Full angle rapid fingerprint identification method
CN111914755A (en) Eight-direction gradient-solving fingerprint identification model
CN111639560B (en) Finger vein feature extraction method and device based on dynamic fusion of vein skeleton line and topographic concave-convex characteristics
CN110222661B (en) Feature extraction method for moving target identification and tracking
CN107679494B (en) Fingerprint image matching method based on selective extension
CN111507206A (en) Finger vein identification method based on multi-scale local feature fusion
CN116704557A (en) Low-quality fingerprint matching method based on texture information
CN115439891A (en) Fingerprint identification method based on small fingerprint head, low computational power and low memory chip
CN112036382A (en) Finger vein feature extraction and identification method and device based on topographic point classification
Erol et al. Feature selective filtering for ridge extraction
Jiang et al. Minutiae extraction by adaptive tracing the gray level ridge of the fingerprint image
CN113723314A (en) Sugarcane stem node identification method based on YOLOv3 algorithm
Lin A novel iris recognition method based on the natural-open eyes
Fang et al. A new fingerprint thinning algorithm
KR100374695B1 (en) Automatic Fingerprint Identification System using Direct Ridge Extraction
Surmacz et al. An improved algorithm for feature extraction from a fingerprint fuzzy image
Wang et al. Fingerprint refinement model based on improved OPTA
Kuban et al. A NOVEL MODIFICATION OF SURF ALGORITHM FOR FINGERPRINT MATCHING.
CN113269020A (en) Fingertip vein image feature identification positioning method and system
George et al. A survey on prominent iris recognition systems

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