A method, apparatus and device for fingerprint pose estimation

Abstract

The application discloses a fingerprint posture estimation method, device and equipment, and the method comprises the following steps: acquiring a finger image of a finger shot when a fingerprint is collected; extracting a finger non-fingerprint area feature from the finger image, wherein the finger non-fingerprint area feature at least comprises one of a finger contour, a knuckle texture, a knuckle wrinkle and a nail edge; and calculating posture information of the collected fingerprint on the finger through the finger non-fingerprint area feature. The technical scheme provided by the application can improve the accuracy of fingerprint posture estimation.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of fingerprint recognition, and in particular to a fingerprint posture estimation method, device and equipment. BACKGROUND

[0002] Fingerprint image is one of the important biological information for identity recognition. Since most of the fingerprint collection devices do not strictly limit the collection posture of the fingerprint, the fingerprint image with uniform posture (angle or position on the finger, etc.) cannot be obtained. Therefore, in the fingerprint image matching algorithm, estimating the posture of the fingerprint image or the relative posture transformation relationship between images is a relatively important link.

[0003] The prior art mainly estimates the fingerprint posture based on the feature points (such as singular points, etc.) or texture feature rules of the fingerprint. However, the arch type fingerprint and the like has unstable feature points, and can only rely on manual experience to extract points such as the highest curvature point on the ridge line as a reference. The method relying on the texture feature rule often needs to uniformly pre-process the fingerprint posture and then statistically and learn the rule. In addition, when part of the fingerprint image is missing or distorted, it is more difficult to obtain stable posture information. Therefore, it is limited to obtain the posture information of the fingerprint on the finger from the fingerprint texture only, and the effect is not good, and a new fingerprint posture estimation method is needed to improve the accuracy of the fingerprint posture estimation. SUMMARY

[0004] Therefore, the embodiments of the present application provide a fingerprint posture estimation method, device and equipment, thereby improving the accuracy of the fingerprint posture estimation.

[0005] According to a first aspect, the embodiments of the present application provide a fingerprint posture estimation method, which comprises: acquiring a finger image photographed on a finger during fingerprint collection; extracting a finger non-fingerprint area feature from the finger image, the finger non-fingerprint area feature comprising at least one of a finger contour, a finger joint texture, a finger joint wrinkle and a nail edge; and calculating posture information of the collected fingerprint on the finger through the finger non-fingerprint area feature.

[0006] Optionally, the acquiring the finger image photographed on the finger during the fingerprint collection comprises: acquiring a finger front image and a finger side image photographed on the finger during the fingerprint collection.

[0007] Optionally, the acquiring the finger front image and the finger side image photographed on the finger during the fingerprint collection comprises: acquiring a first top view and a first side view of a fingerprint collection device; acquiring a second top view and a second side view when the finger is placed on the fingerprint collection device; and obtaining the finger front image by subtracting the second top view from the first top view, and obtaining the finger side image by subtracting the second side view from the first side view.

[0008] Optionally, the extracting the finger non-fingerprint area feature from the finger image comprises: extracting a pixel point set of the finger front image, and calculating a dimension direction with maximum projection variance of the pixel point set; extracting a position pixel point where a first knuckle fold line and a finger edge intersect from the finger side image, and extracting two side endpoints of a segment where the finger contacts the fingerprint collection device from the finger side image, the first knuckle fold line being a knuckle fold line between a finger tip segment and a middle segment of the finger; and taking the dimension direction, the position pixel point and the two side endpoints as the finger non-fingerprint area feature.

[0009] Optionally, the extracting the position pixel point where the first knuckle fold line and the finger edge intersect from the finger side image comprises: identifying the first knuckle fold line between the finger tip segment and the middle segment of the finger from the finger side image through a preset convolutional neural network module; obtaining an edge image of the finger side image; performing an AND operation on the edge image and the first knuckle fold line to obtain an AND result; and calculating a centroid position of the AND result to obtain the position pixel point.

[0010] Optionally, the extracting the two side endpoints of the segment where the finger contacts the fingerprint collection device from the finger side image comprises: detecting straight lines in the finger side image through a Hough straight line detection operator; selecting a target straight line located at the lowermost part of the image from the detected straight lines, and taking two side endpoints of the target straight line as the two side endpoints of the segment where the finger contacts the fingerprint collection device.

[0011] Optionally, the calculating the posture information of the collected fingerprint on the finger through the finger non-fingerprint area feature comprises: calculating a first straight line distance between an endpoint of the two side endpoints close to the first knuckle fold line and the position pixel point; calculating a second straight line distance between the two side endpoints; determining an interval range where the fingerprint is located on the finger by taking the position pixel point as a starting point of the interval, and taking a sum of the first straight line distance and the second straight line distance as a length of the interval; determining a fingerprint offset angle based on the dimension direction; and taking the interval range and the fingerprint offset angle as the posture information.

[0012] According to a second aspect, an embodiment of the present application provides a fingerprint posture estimation device, the device comprising: an image collection module configured to acquire a finger image of a finger shot when collecting a fingerprint; a non-fingerprint area feature extraction module configured to extract a finger non-fingerprint area feature from the finger image, the finger non-fingerprint area feature comprising at least one of a finger contour, a knuckle texture, a knuckle fold line and a nail edge; and a posture estimation module configured to calculate posture information of a collected fingerprint on a finger through the finger non-fingerprint area feature.

[0013] According to a third aspect, an embodiment of the present application provides a fingerprint posture estimation device, comprising: a fingerprint collection device, a finger camera device and a controller; the fingerprint collection device is configured to collect a fingerprint, the finger camera device is configured to take a picture of a finger when the fingerprint collection device collects the fingerprint; the controller comprises a memory and a processor, the memory and the processor are in communication connection with each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the method of the first aspect or any one of the optional embodiments of the first aspect.

[0014] According to a fourth aspect, an embodiment of the present application provides a computer readable storage medium, which stores computer instructions, and the computer instructions are used to make the computer execute the method of the first aspect or any one of the optional embodiments of the first aspect.

[0015] The technical scheme provided in the present application has the following advantages:

[0016] The technical scheme provided in the present application acquires a finger image taken by the finger camera device when collecting the fingerprint; extracts a finger non-fingerprint area feature from the finger image, the finger non-fingerprint area feature at least includes one of a finger contour, a texture on a back side of a finger joint, a finger joint wrinkle line and a nail edge; and finally calculates posture information of the collected fingerprint on the finger through the finger non-fingerprint area feature. Compared with the prior art which only acquires the posture information of the fingerprint on the finger from the fingerprint texture, the embodiment of the present application starts from the features of the finger non-fingerprint area (such as the finger contour, the texture of the finger joint or the nail edge), and the features of the multiple finger non-fingerprint areas help to locate the posture of the fingerprint, so that the present application proposes another method for estimating the posture of the fingerprint based on the features of the finger non-fingerprint area, and improves the accuracy of the posture estimation of the fingerprint. BRIEF DESCRIPTION OF DRAWINGS

[0017] The features and advantages of the present application will be more clearly understood through the following detailed description taken in conjunction with the accompanying drawings, which are shown by way of illustration and not limitation, and in which:

[0018] Figure 1 A step schematic diagram of a fingerprint posture estimation method in an embodiment of the present application is shown;

[0019] Figure 2 An effect schematic diagram of a first top view collected by the finger camera device in an embodiment of the present application is shown;

[0020] Figure 3 An effect schematic diagram of a second top view collected by the finger camera device in an embodiment of the present application is shown;

[0021] Figure 4An effect diagram of a finger front image in an embodiment of the present application is shown.

[0022] Figure 5 An effect diagram of a dimension direction calculated based on a finger front image in an embodiment of the present application is shown.

[0023] Figure 6 An identification result diagram of a finger side image in an embodiment of the present application is shown.

[0024] Figure 7 A structure diagram of a fingerprint posture estimation device in an embodiment of the present application is shown.

[0025] Figure 8 A structure diagram of a fingerprint posture estimation device in an embodiment of the present application is shown.

[0026] Figure 9 A structure diagram of a controller in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0027] In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0028] Please refer to Figure 1 In an embodiment, a fingerprint posture estimation method specifically includes the following steps.

[0029] Step S101: Obtain a finger image photographed to a finger during fingerprint collection.

[0030] Specifically, when a user places a finger on a fingerprint collection device for fingerprint collection, a finger image of the user is collected at the same time. The device for collecting the finger image of the user includes but is not limited to a camera, a mobile phone camera, a monitoring camera, and other devices with a photographing function.

[0031] Step S102: Extract a finger non-fingerprint area feature from the finger image, the finger non-fingerprint area feature at least including one of a finger contour, a knuckle texture, a knuckle wrinkle, and a nail edge.

[0032] Specifically, the non-fingerprint area features of the user are extracted from the image of the finger of the user by image recognition, contour detection and the like, at least including one of a finger contour, a knuckle texture (especially the texture on the back of the knuckle), a knuckle fold line and a nail edge, so as to determine the placement position of the finger through the non-fingerprint area features.

[0033] Step S103: calculating the posture information of the collected fingerprint on the finger through the non-fingerprint area features of the finger.

[0034] Specifically, the embodiment of the present application can determine the size, length, position and direction and the like of the non-fingerprint area features of the finger by analyzing the non-fingerprint area features of the finger collected in the previous step, so as to calculate the placement angle and position of the finger of the user by means of the non-fingerprint area features of the finger, and then the fingerprint collected by the fingerprint collection device can be corresponded to a part of the area of the finger of the user and to a certain placement angle of the finger, so as to obtain more accurate posture information of the fingerprint. Compared with the prior art which simply obtains the posture information of the fingerprint on the finger from the fingerprint texture, the embodiment of the present application starts from the features of the non-fingerprint area of the finger, and the multiple non-fingerprint area features of the finger are helpful for positioning the posture of the fingerprint, and another method for estimating the posture of the fingerprint based on the features of the non-fingerprint area of the finger is proposed, more posture factors are considered comprehensively, and the accuracy of the estimation of the posture of the fingerprint is improved.

[0035] Specifically, in an embodiment, the above step S101 specifically includes the following steps:

[0036] Step zero: obtaining the finger front image and the finger side image of the finger shot during fingerprint collection.

[0037] Specifically, the embodiment of the present application collects the non-fingerprint area features of the finger from the front and side of the finger respectively, and the finger camera device can be multiple to collect simultaneously, for example, three finger camera devices are arranged above, left and right of the fingerprint collection device respectively in the embodiment of the present application. Through the all-around shooting of the finger, the perfect non-fingerprint area features of the finger are captured, so as to improve the accuracy of the analysis of the posture of the finger according to the more perfect non-fingerprint area features of the finger.

[0038] Specifically, in an embodiment, the above step zero specifically includes the following steps:

[0039] Step one: obtaining the first top view and the first side view of the fingerprint collection device.

[0040] Specifically, in order to further improve the accuracy of the extraction of the non-fingerprint area features of the finger, the embodiment of the present application filters the background part of the finger image to reduce the influence of redundant pixels on the feature extraction step. First, the first top view and the first side view of the fingerprint collection device are shot by the finger camera device as a background image, for example Figure 2Fig. 1 shows a first top view of a fingerprint collection device taken by a finger camera device.

[0041] Step two: obtain a second top view and a second side view of the finger when the finger is placed on the fingerprint collection device.

[0042] Specifically, when the user's finger is placed on the fingerprint collection device for fingerprint collection, the second top view and the second side view are taken again by the finger camera device at the same position, as shown in Figure 3 Fig. 2 shows a second top view taken when the finger is placed on the fingerprint collection device.

[0043] Step three: difference between the second top view and the first top view to obtain a front view image of the finger.

[0044] Step four: difference between the second side view and the first side view to obtain a side view image of the finger.

[0045] Specifically, the hand region image set is obtained by background difference method in the embodiment of the present application

[0046]

[0047] In the formula, i represents the i-th image, N represents that the first top view, the second top view, the first side view and the second side view are all taken for N times, represents the i-th front view image or side view image of the finger, T is a noise threshold, is the i-th first top view or first side view, I i is the i-th corresponding second top view or second side view.

[0048] The front view image and the side view image of the finger containing only the finger region are obtained by the background difference method (for example Figure 4 Fig. 2 shows the front view image of the finger after subtraction of the background). Then, the extracted front view image and side view image of the finger are converted into a binary image, so that the non-fingerprint region features of the finger can be accurately extracted from the image, and the accuracy of the fingerprint posture estimation is improved.

[0049] Specifically, in an embodiment, the above step S102 specifically includes the following steps:

[0050] Step five: extract a pixel point set of the front view image of the finger, and calculate a dimension direction with maximum projection variance of the pixel point set.

[0051] Specifically, the front view image of the finger is selected, and after binarization processing, a pixel point set P t , of the front view image of the finger is extracted (i.e. a pixel point set containing only the finger part, x and y are pixel coordinates), and then principal component analysis method is used to solvet The dimension direction with the maximum projection variance of each pixel point coordinate The specific steps are as follows: 1. Translate all sample points in the pixel point set to the center position of the sample space, so that the samples in the current pixel point set are centralized, in other words, subtract the sample mean from each sample in the pixel point set. 2. Calculate the projection of each sample on a certain characteristic direction vector for the centralized sample. 3. Calculate the projection variance of each sample based on the projection of each sample on a certain characteristic direction vector. 4. Calculate the total projection variance based on the projection variance of each sample. 5. Use the Lagrange multiplier method to process the optimization problem of the maximum total projection variance. 6. Take the characteristic vector corresponding to the maximum eigenvalue of the optimized pixel point set covariance matrix as the optimal characteristic direction, that is, the dimension direction The calculated dimension direction is as shown in Figure 5 The angle of the collected fingerprint on the finger can be determined through the direction.

[0052] Step six: extract the position pixel point where the first knuckle fold line and the finger edge intersect from the finger side image, and extract the two side endpoints of the contact segment of the finger and the fingerprint collection device from the finger side image. The first knuckle fold line is the knuckle fold line between the first segment of the finger tip and the middle segment of the finger.

[0053] Specifically, as shown in Figure 6 The first knuckle fold line L between the first segment of the finger tip and the middle segment of the finger is identified from the finger side image by an image recognition algorithm, and then the position pixel point X3 where the first knuckle fold line L and the finger edge intersect is calculated. At the same time, the two side endpoints X1 and X2 of the contact segment of the finger and the fingerprint collection device are extracted from the finger side image by an image edge detection algorithm.

[0054] Step seven: take the dimension direction, the position pixel point and the two side endpoints as the characteristics of the non-fingerprint area of the finger.

[0055] Finally, the dimension direction the position pixel point X3, the two side endpoints X1 and X2 identified in the above steps are taken as the characteristics of the non-fingerprint area of the finger, so that the collected fingerprint is accurately estimated to be the result of the finger in which angle state, the collected fingerprint is the result of the finger under how much pressing force, and the collected fingerprint area occupies the specific range and area of the finger, thereby improving the accuracy of the fingerprint posture estimation.

[0056] Specifically, in an embodiment, the above step six specifically includes the following steps:

[0057] Step eight: identify the first knuckle fold line between the first segment of the finger tip and the middle segment of the finger from the finger side image by a pre-set convolutional neural network module.

[0058] Specifically, the embodiment of the present application batch collects finger side surface images, and manually marks the positions of the creases. A convolutional neural network module is pre-trained to detect the finger crease L, and the connected region with the highest score is selected as the final result. The crease L is a binary image, and the part with a value of 1 is the crease region, and the part with a value of 0 is the non-crease region. The first knuckle crease between the finger tip and the middle knuckle of the finger is identified from the finger side surface image through the convolutional neural network module, so as to ensure the accuracy of the knuckle crease identification.

[0059] Step nine: obtaining an edge image of the finger side surface image.

[0060] Step ten: performing an AND operation on the edge image and the first knuckle crease to obtain an AND result.

[0061] Step eleven: calculating the centroid position of the AND result to obtain a position pixel point.

[0062] Specifically, the embodiment of the present application extracts the edge image of the finger side surface image through an edge detection algorithm including but not limited to a canny edge detection, a sobel edge detection, and the like, and stores the edge image in the form of a binary image, and the part with a value of 1 is the edge region, and the part with a value of 0 is the non-edge region. Then, the position pixel point X3 of the intersection of the first knuckle crease L and the finger is calculated. Specifically, the AND operation of the crease image L and the edge image is performed to obtain an AND result A, and finally the centroid position of A is calculated, that is, X3.

[0063] Step twelve: detecting a straight line in the finger side surface image through a Hough straight line detection operator;

[0064] Step thirteen: selecting a target straight line located at the lowermost part of the image from the detected straight lines, and taking the two side end points of the target straight line as the two side end points of the contact segment of the finger and the fingerprint collection device.

[0065] Specifically, the embodiment of the present application extracts the two side end points {X1, X2} of the contact segment of the finger and the fingerprint collection device from the finger side surface image. When the finger pulp is pressed against the fingerprint collection device, considering that the curve of the finger pulp will become a straight line adhering to the fingerprint collection device due to the pressing action, the embodiment of the present application uses the Hough straight line detection operator to detect the straight line on the finger side surface image, and the other parts of the finger are basically curves, so that the number of detected straight lines is small, and the influence on the data is small. Then, the straight line located at the lowermost part of the finger side surface image is selected, so as to accurately locate the two side end points {X1, X2} of the contact segment of the finger and the fingerprint collection device. For example, the point X1 is close to the finger tip, and the point X2 is close to the crease.

[0066] Specifically, in an embodiment, the above step S103 specifically includes the following steps:

[0067] Step fourteen: calculate the first straight line distance between the end point close to the first knuckle fold line and the position pixel point of the two side end points.

[0068] Step fifteen: calculate the second straight line distance between the two side end points.

[0069] Specifically, after identifying the two side end points {X1, X2} and the position pixel point of the segment of the finger in contact with the fingerprint collection device, the first straight line distance between the end point close to the first knuckle fold line and the position pixel point is calculated, for example Figure 6 The straight line distance between X2 and X3 is the first straight line distance. And the second straight line distance between the two side end points is calculated, for example Figure 6 The straight line distance between X1 and X2 is the second straight line distance.

[0070] Step sixteen: determine the interval range of the fingerprint on the finger with the position pixel point as the interval starting point and the sum of the first straight line distance and the second straight line distance as the interval length.

[0071] Specifically, when the finger is not pressed on the fingerprint collection device, the side shape of the finger pad is a curve, and the sum of the first straight line distance and the second straight line distance is used to approximate the curve length of the side curve of the finger pad in the embodiment of the application. With the position pixel point as the interval starting point and the sum of the first straight line distance and the second straight line distance as the interval length, an interval range (X3, X3+d1+d2) can be determined, where X3 is the coordinate of the position pixel point, indicating the interval starting point, and d1 and d2 represent the first straight line distance and the second straight line distance respectively.

[0072] Step seventeen: determine the fingerprint offset angle based on the dimension direction.

[0073] Step eighteen: take the interval range and the fingerprint offset angle as the posture information.

[0074] Finally, the dimension direction is mapped to the finger top view coordinate system, the offset angle of the fingerprint on the finger is calculated, the fingerprint offset angle is taken as the direction information of the fingerprint, and the above interval range (X3, X3+d1+d2) is taken as the position information of the fingerprint on the user's finger. Through the scheme provided by the embodiment of the application, the posture information of the fingerprint can be accurately estimated, and the accuracy of the fingerprint posture estimation is improved.

[0075] Through the above steps, the technical scheme provided by the present application obtains a finger image of a finger photographed when a fingerprint is collected; extracts a non-fingerprint area feature of the finger from the finger image (the non-fingerprint area feature of the finger at least includes one of a finger contour, a knuckle wrinkle, and a nail edge); and finally calculates posture information of the collected fingerprint on the finger through the non-fingerprint area feature of the finger. Compared with the prior art which has limitations in obtaining the posture information of the fingerprint on the finger only from a fingerprint texture, the embodiment of the present application starts from the features of the non-fingerprint area of the finger (such as the finger contour, the knuckle texture, or the nail edge), and the features of the multiple non-fingerprint areas of the finger help to locate the posture of the fingerprint, so that the present application proposes another method for estimating the posture of the fingerprint based on the features of the non-fingerprint area of the finger, and improves the accuracy of the estimation of the posture of the fingerprint.

[0076] As shown in Figure 7 The present embodiment also provides a fingerprint posture estimation device, which comprises:

[0077] The image collection module 101 is configured to obtain a finger image of a finger photographed when a fingerprint is collected. For details, refer to the related description of step S101 in the above method embodiment, which will not be repeated here.

[0078] The non-fingerprint area feature extraction module 102 is configured to extract a non-fingerprint area feature of the finger from the finger image, and the non-fingerprint area feature of the finger at least includes one of a finger contour, a knuckle wrinkle, and a nail edge. For details, refer to the related description of step S102 in the above method embodiment, which will not be repeated here.

[0079] The posture estimation module 103 is configured to calculate posture information of the collected fingerprint on the finger through the non-fingerprint area feature of the finger. For details, refer to the related description of step S103 in the above method embodiment, which will not be repeated here.

[0080] The fingerprint posture estimation device provided by the embodiment of the present application is configured to execute the fingerprint posture estimation method provided by the above embodiment, and has the same implementation manner and principle. For details, refer to the related description of the above method embodiment, which will not be repeated here.

[0081] Through the cooperation of the above various components, the technical scheme provided by the application obtains a finger image of a finger photographed when a fingerprint is collected; extracts a finger non-fingerprint area feature (the finger non-fingerprint area feature at least includes one of a finger contour, a knuckle fold line and a nail edge) from the finger image; and finally calculates the posture information of the collected fingerprint on the finger through the finger non-fingerprint area feature. Compared with the prior art which simply obtains the posture information of the fingerprint on the finger from the fingerprint texture, the embodiment of the application starts from the features of the finger non-fingerprint area (such as the finger contour, the knuckle texture or the nail edge), and the features of the multiple finger non-fingerprint areas help to locate the fingerprint posture, so that the application proposes another method for estimating the fingerprint posture based on the features of the finger non-fingerprint area, and improves the accuracy of the fingerprint posture estimation.

[0082] Figure 8 A fingerprint posture estimation device is shown in an embodiment of the application, which comprises a fingerprint collection device 1, a finger camera device 2 and a controller 3; wherein the fingerprint collection device 1 is used to collect a fingerprint, the finger camera device 2 is used to photograph a finger when the fingerprint collection device 1 collects the fingerprint, and the controller 3 is used to execute the method in the method embodiment. Thus, when a user places a finger on the fingerprint collection device 1, the controller 3 controls the finger camera device 2 to collect a front image and a side image of the finger of the user, and then extracts the non-fingerprint area features of the finger of the user through image edge detection and other algorithms.

[0083] Specifically, in an embodiment, the controller 3 comprises a processor 901 and a memory 902, which can be connected through a bus or other means, Figure 9 For example, the connection through the bus is taken as an example.

[0084] The processor 901 can be a central processing unit (CPU). The processor 901 can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above.

[0085] The memory 902, as a non-transitory computer readable storage medium, can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the methods in the above method embodiments. The processor 901 performs various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory 902, that is, implements the methods in the above method embodiments.

[0086] The memory 902 can include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required by at least one function; and the data storage area can store data created by the processor 901 and the like. In addition, the memory 902 can include a high-speed random access memory, and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory 902 can optionally include a memory disposed remotely with respect to the processor 901, which can be connected to the processor 901 through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0087] One or more modules are stored in the memory 902, and when executed by the processor 901, the methods in the above method embodiments are performed.

[0088] The specific details of the above fingerprint posture estimation device can be understood by referring to the corresponding related descriptions and effects in the above method embodiments, which will not be described here.

[0089] Those skilled in the art can understand that all or part of the processes in the above method embodiments can be completed by a computer program instructing related hardware, and the implemented program can be stored in a computer readable storage medium. When the program is executed, it can include the processes of the above method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), etc. The storage medium can also include a combination of the above types of memories.

[0090] Although the embodiments of the present application are described in conjunction with the accompanying drawings, various modifications and changes can be made by those skilled in the art without departing from the spirit and scope of the present application, and such modifications and changes fall within the scope defined by the appended claims.

Claims

1. A method of fingerprint pose estimation, the method comprising: The method comprises: acquiring a finger image photographed on a finger during fingerprint acquisition; extracting a finger non-fingerprint area feature from the finger image, the finger non-fingerprint area feature comprising at least one of a finger contour, a knuckle texture, a knuckle wrinkle and a nail edge; calculating, by using the finger non-fingerprint area feature, posture information of the acquired fingerprint on the finger; the step of extracting the finger non-fingerprint area feature from the finger image comprises: extracting a pixel point set of a front image of the finger, and calculating a dimension direction of the pixel point set with a maximum projection variance; extracting, from a side image of the finger, a position pixel point where a first knuckle wrinkle and a finger edge intersect, and extracting, from the side image of the finger, two side endpoints of a contact segment of the finger and a fingerprint acquisition device, the first knuckle wrinkle being a knuckle wrinkle between a finger tip segment and a middle segment of the finger; using the dimension direction, the position pixel point and the two side endpoints as the finger non-fingerprint area feature; the step of extracting, from the side image of the finger, the position pixel point where the first knuckle wrinkle and the finger edge intersect comprises: identifying, by using a preset convolutional neural network module, the first knuckle wrinkle between the finger tip segment and the middle segment of the finger from the side image of the finger; acquiring an edge image of the side image of the finger; performing an AND operation on the edge image and the first knuckle wrinkle to obtain an AND result; calculating a centroid position of the AND result to obtain the position pixel point; the step of extracting, from the side image of the finger, the two side endpoints of the contact segment of the finger and the fingerprint acquisition device comprises: detecting, by using a Hough straight line detection operator, a straight line in the side image of the finger; selecting a target straight line located at a lowermost position of the image from the detected straight lines, and using two side endpoints of the target straight line as the two side endpoints of the contact segment of the finger and the fingerprint acquisition device; the step of calculating, by using the finger non-fingerprint area feature, the posture information of the acquired fingerprint on the finger comprises: calculating a first straight line distance between an endpoint of the two side endpoints close to the first knuckle wrinkle and the position pixel point; calculating a second straight line distance between the two side endpoints; determining an interval range of the fingerprint on the finger by using the position pixel point as an interval starting point, and using a sum of the first straight line distance and the second straight line distance as an interval length; determining a fingerprint offset angle based on the dimension direction; and using the interval range and the fingerprint offset angle as the posture information.

2. The method of claim 1, wherein, the step of acquiring the finger image photographed on the finger during fingerprint acquisition comprises: acquiring a front image and a side image of the finger photographed on the finger during fingerprint acquisition.

3. The method of claim 2, wherein, the step of acquiring the front image and the side image of the finger photographed on the finger during fingerprint acquisition comprises: acquiring a first top view and a first side view of a fingerprint acquisition device; acquiring a second top view and a second side view of the finger placed on the fingerprint acquisition device; subtracting the second top view from the first top view to obtain the front image of the finger; subtracting the second side view from the first side view to obtain the side image of the finger.

4. A device for estimating a pose of a fingerprint, the device comprising: the device comprises: an image acquisition module configured to acquire a finger image photographed on a finger during fingerprint acquisition; a non-fingerprint area feature extraction module configured to extract a non-fingerprint area feature of a finger from the finger image, the non-fingerprint area feature of the finger comprising at least one of a finger contour, a knuckle texture, a knuckle wrinkle, and a nail edge; a pose estimation module configured to calculate pose information of a collected fingerprint on the finger based on the non-fingerprint area feature of the finger; the non-fingerprint area feature of the finger is extracted from the finger image by: extracting a pixel point set of a front image of the finger, and calculating a dimension direction of the pixel point set with a maximum projection variance; extracting a position pixel point at which a first knuckle wrinkle of the finger intersects with an edge of the finger from a side image of the finger, and extracting two end points of a contact segment of the finger with a fingerprint collection device from the side image of the finger, the first knuckle wrinkle being a knuckle wrinkle between a first segment of a finger tip and a middle segment of the finger; the dimension direction, the position pixel point, and the two end points are taken as the non-fingerprint area feature of the finger; the position pixel point at which the first knuckle wrinkle of the finger intersects with the edge of the finger is extracted from the side image of the finger by: identifying the first knuckle wrinkle between the first segment of the finger tip and the middle segment of the finger from the side image of the finger by a preset convolutional neural network module; obtaining an edge image of the side image of the finger; performing an AND operation on the edge image and the first knuckle wrinkle to obtain an AND result; calculating a centroid position of the AND result to obtain the position pixel point; the two end points of the contact segment of the finger with the fingerprint collection device are extracted from the side image of the finger by: detecting straight lines in the side image of the finger by a Hough straight line detection operator; selecting a target straight line located at a lowermost position of the image from the detected straight lines, and taking two end points of the target straight line as the two end points of the contact segment of the finger with the fingerprint collection device; the pose information of the collected fingerprint on the finger is calculated based on the non-fingerprint area feature of the finger by: calculating a first straight line distance between an end point of the two end points close to the first knuckle wrinkle and the position pixel point; calculating a second straight line distance between the two end points; determining an interval range of the fingerprint on the finger with the position pixel point as a starting point of the interval, and a sum of the first straight line distance and the second straight line distance as a length of the interval; determining a fingerprint offset angle based on the dimension direction; and taking the interval range and the fingerprint offset angle as the pose information.

5. A device for estimating a pose of a fingerprint, the device comprising: comprise: a fingerprint collection device, a finger camera device, and a controller; the fingerprint collection device is configured to collect a fingerprint, and the finger camera device is configured to capture a finger when the fingerprint collection device collects the fingerprint; the controller comprises a memory and a processor, the memory and the processor are communicatively connected with each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the method according to any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, the computer readable storage medium stores computer instructions for causing the computer to perform the method according to any one of claims 1-3.

Images