Fingerprint recognition method and electronic device for performing method
The fingerprint recognition method addresses the challenge of incomplete ridge detection by tracing and connecting ridges using a ridge estimation algorithm, enhancing accuracy and reliability in various conditions.
Patent Information
- Application Number
- PCT/KR2026/096292
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-09
- Filing Date
- 2026-04-24
- Publication Date
- 2026-11-12
AI Technical Summary
Existing fingerprint recognition technologies struggle to accurately identify fingerprint ridges when the user's finger is dry or has injuries, leading to incomplete or false detection due to missing or broken ridges.
A fingerprint recognition method using a ridge estimation algorithm that traces ridge paths from starting points, predicts and adjusts next points based on motion state parameters, and connects missing ridges to generate precise fingerprint information.
Enables robust fingerprint recognition regardless of finger condition by restoring missing or broken ridges, improving accuracy and reliability.
Smart Images

Figure KR2026096292_12112026_PF_FP_ABST
Abstract
Description
Fingerprint recognition method and electronic device performing the method
[0001] The present disclosure relates to a method for generating fingerprint information from a fingerprint image and performing fingerprint recognition, and an electronic device for performing fingerprint recognition.
[0002] Fingerprint recognition is a biometric technology that performs identity authentication by scanning and analyzing a user's fingerprint on an electronic device. Used in various devices such as smartphones, tablets, laptops, and access control systems, fingerprint recognition is a core function that provides security, convenience, and a personalized experience. When fingerprint recognition is used, there are cases where fingerprint ridges are missing due to reasons such as dryness or injuries to the user's finger, which can result in fingerprint information being missed or falsely detected. To address this, various algorithms are being applied to enhance fingerprint visibility; however, it is necessary to provide robust fingerprint recognition results regardless of the condition of the finger.
[0003] According to one aspect of the present disclosure, a fingerprint recognition method performed by an electronic device may be provided. The method may include the steps of acquiring a fingerprint image, estimating the ridges of a fingerprint represented by the fingerprint image by tracing ridge paths from one or more starting points within the fingerprint image through a ridge estimation algorithm, and acquiring fingerprint information for fingerprint recognition based on the estimated ridges. For each of the ridges of the fingerprint, the ridge estimation algorithm may include the steps of: predicting the next point of the ridge path based on a current point and a motion state parameter; adjusting the predicted next point and updating the motion state parameter based on the predicted next point and pixel values of the fingerprint image; and identifying a single ridge by tracing the ridge path through iteration of prediction and adjustment for subsequent points of the ridge path.
[0004] According to one aspect of the present disclosure, an electronic device for performing fingerprint recognition may be provided. The electronic device may include at least one processor comprising a processing circuit, a memory for storing instructions, and a display.
[0005] By executing the above instructions by the at least one processor, the electronic device can acquire a fingerprint image, estimate the ridges of the fingerprint represented by the fingerprint image by tracing ridge paths from one or more starting points within the fingerprint image through a ridge estimation algorithm, and acquire fingerprint information for fingerprint recognition based on the estimated ridges. For each of the ridges of the fingerprint, the ridge estimation algorithm can identify a single ridge by tracing the ridge path through the repetition of prediction and adjustment for subsequent points of the ridge path: predicting the next point of the ridge path based on a current point and a motion state parameter, adjusting the predicted next point and updating the motion state parameter based on the predicted next point and pixel values of the fingerprint image.
[0006] According to one aspect of the present disclosure, a computer-readable recording medium may be provided having a program recorded thereon for executing any one of the methods described above and below, wherein an electronic device generates fingerprint information from a fingerprint image and performs fingerprint recognition.
[0007] FIG. 1 is a drawing for exemplarily illustrating a fingerprint recognition operation performed by an electronic device according to one embodiment of the present disclosure.
[0008] Figure 2 is a diagram illustrating various examples of fingerprint images obtained according to the condition of the finger.
[0009] FIG. 3 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present disclosure.
[0010] Figure 4 is a diagram illustrating possible fingerprint recognition errors that may occur in a general fingerprint recognition method.
[0011] FIG. 5 is a flowchart illustrating the operation of a fingerprint recognition model in one embodiment of the present disclosure.
[0012] FIG. 6 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure estimating the ridge thickness of a fingerprint.
[0013] FIG. 7 is a drawing for explaining a fingerprint recognition model according to one embodiment of the present disclosure.
[0014] FIG. 8 is a diagram illustrating a motion state parameter used by an electronic device for fingerprint recognition according to one embodiment of the present disclosure.
[0015] FIG. 9 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure tracking a ridge path.
[0016] FIG. 10a is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure for predicting points of a ridge path.
[0017] FIG. 10b is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to check and adjust predicted points of a ridge path.
[0018] FIG. 11 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure for predicting points of a ridge path.
[0019] FIG. 12 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure determining the end point of a ridge path.
[0020] FIG. 13 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure determining the end point of a ridge path.
[0021] FIG. 14 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure determining starting points for ridge path tracking.
[0022] FIG. 15 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure optimizing the connection between ridges.
[0023] FIG. 16 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure optimizing the connection between ridges.
[0024] FIG. 17 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to estimate a fingerprint scratch.
[0025] FIG. 18 is a diagram illustrating the operation of an electronic device performing fingerprint recognition according to one embodiment of the present disclosure.
[0026] FIG. 19 is a diagram illustrating the operation of an electronic device performing fingerprint recognition according to one embodiment of the present disclosure.
[0027] The terms used in this disclosure will be briefly explained, and the disclosure will be described in detail. In this disclosure, the expression "at least one of a, b, or c" may refer to "a," "b," "c," "a and b," "a and c," "b and c," "all of a, b, and c," or variations thereof.
[0028] The terms used in this disclosure have been selected to be as widely used and general as possible, taking into account their functions within this disclosure; however, these terms may vary depending on the intent of those skilled in the art, case law, the emergence of new technologies, etc. Additionally, in specific cases, terms have been selected at the applicant's discretion, and in such cases, their meanings will be described in detail in the relevant explanatory sections. Therefore, terms used in this disclosure should be defined not merely by their names, but based on their meanings and the overall content of this disclosure.
[0029] Singular expressions may include plural expressions unless the context clearly indicates otherwise. Terms used herein, including technical or scientific terms, may have the same meaning as generally understood by those skilled in the art as described in this disclosure. Additionally, terms including ordinal numbers, such as "first" or "second," as used herein may be used to describe various components, but said components should not be limited by said terms. Such terms are used solely for the purpose of distinguishing one component from another.
[0030] In the present disclosure, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, terms such as "part" and "module" as described in the present disclosure refer to a unit that processes at least one function or operation, and may be implemented in hardware or software, or as a combination of hardware and software.
[0031] Embodiments of the present disclosure are described below with reference to the attached drawings so that those skilled in the art can easily implement them. However, the present disclosure may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, parts unrelated to the description have been omitted from the drawings to clearly explain the present disclosure. Additionally, for convenience of explanation, different reference numbers have been used throughout the entire disclosure even for identical components.
[0032] The present disclosure will be described below with reference to the attached drawings.
[0033] FIG. 1 is a drawing for exemplarily illustrating a fingerprint recognition operation performed by an electronic device according to one embodiment of the present disclosure.
[0034] In one embodiment, the electronic device may generate fingerprint information (20) based on a fingerprint image (10). In the example of FIG. 1, the fingerprint information (20) does not correspond to the fingerprint image (10) but represents fingerprint information (20) extracted from any fingerprint image drawn for example.
[0035] An electronic device may acquire fingerprint information (20) using, for example, a fingerprint recognition model that uses a ridge estimation algorithm. The fingerprint information (20) may be a minutiae containing detailed feature information of the fingerprint. The minutiae may include, for example, a bifurcation, which is a point where a single ridge splits into two, and an ending, which is a point where a ridge ends. However, the features included in the fingerprint information (20) are not limited thereto, and may further include a scar, an island, a pore, a crossing, a ridge join, etc. in the fingerprint information (20).
[0036] The fingerprint recognition model may be a Kalman filter based on a Hidden Markov Model (HMM), but is not limited thereto. The fingerprint recognition model can estimate multiple ridges included in the fingerprint by tracing ridge paths from one or more starting points within the fingerprint image. The fingerprint recognition model can estimate multiple ridges by repeating the process of estimating single ridges within the fingerprint image. In the process of estimating a single ridge, the fingerprint recognition model can predict the next point of the ridge path based on the current point and motion state parameters. The initial point at which the fingerprint recognition model begins single ridge estimation can be referred to as the starting point. Based on the predicted next point, the fingerprint recognition model identifies whether an actual observed point (pixel) exists within the fingerprint image; if the point is identified, it adjusts the position of the next point, and if the point is not identified, it skips the adjustment of the next point's position, thereby predicting subsequent points within the single ridge. A fingerprint recognition model can determine a single ridge within a fingerprint image by repeating the prediction process of subsequent points and determining the end point of the single ridge based on whether a defined condition is satisfied. As the determination of a single ridge is repeated, multiple ridges within the fingerprint image can be identified.
[0037] The electronic device of the present disclosure may be implemented as various types of devices that provide a fingerprint recognition function. For example, the electronic device may include devices capable of performing biometric authentication through fingerprint recognition and executing security functions by including a fingerprint sensor, such as a smart TV including a display, a smartphone, a tablet PC, a laptop PC, and smart signage, but is not limited thereto.
[0038] Various exemplary operations of an electronic device recognizing a fingerprint will be described in more detail through the drawings and descriptions below.
[0039] Figure 2 is a diagram illustrating various examples of fingerprint images obtained according to the condition of the finger.
[0040] Fingerprint recognition is performed by detecting the patterns of ridges and valleys of a fingerprint. However, the performance of fingerprint recognition may vary depending on the condition of the finger (dryness, injury, moisture, etc.). Figure 2 shows examples of fingerprint images acquired according to various finger conditions.
[0041] A general fingerprint image (210) represents a fingerprint obtained in a general finger condition. The general fingerprint image (210) may be an image with clear distinction between ridges and valleys and low noise. The general fingerprint image (210) may be taken while maintaining appropriate humidity on the skin surface of the finger and is characterized by having uniform pixel contrast. When fingerprint recognition is performed using the general fingerprint image (210), the fingerprint pattern is clearly detected, and a fingerprint recognition result can be obtained with high recognition accuracy.
[0042] A dry fingerprint image (220) represents a fingerprint obtained when the finger skin is dry. When the finger skin is dry, the contact between the finger and the fingerprint sensor may be incomplete, which may result in a problem where the sensing signal is weak. For example, a dry fingerprint image (220) may be an image in which the fingerprint appears blurry due to low pixel contrast between the ridges and valleys. In a dry fingerprint image (220), the ridges may appear partially broken or incomplete. Accordingly, when fingerprint recognition is performed using a dry fingerprint image (220), the accuracy of the fingerprint recognition result may be reduced.
[0043] A fingerprint image (230) containing a wound represents a fingerprint acquired with temporarily or permanently damaged skin. A fingerprint image (230) containing a wound may include patterns such as fine creases or folds that occur in areas where the finger is frequently folded or pressure is applied, which are unrelated to skin damage. A fingerprint image (230) containing a wound may result in a problem where the normal ridge pattern of the fingerprint in the wounded area disappears, and the wounded area is recognized as a valley rather than a ridge, causing inconsistency with the existing fingerprint. A fingerprint image (230) containing a wound may result in lower accuracy of the fingerprint recognition result due to the ridges being disconnected or having abnormal spacing.
[0044] The fingerprint recognition method performed by the electronic device (100) of the present disclosure enables fingerprint recognition by robustly generating fingerprint information regardless of the condition of the finger (e.g., normal condition, dry condition, or condition with a wound). The electronic device (100) is characterized by being able to generate precise fingerprint information by using a ridge estimation algorithm to restore missing or broken ridges by connecting them, and by estimating a wound and utilizing the wound information.
[0045] FIG. 3 is a block diagram illustrating the configuration of an electronic device according to one embodiment of the present disclosure.
[0046] Referring to FIG. 3, an electronic device (100) according to one embodiment may include a processor (110), a memory (120), and a display (130).
[0047] The processor (110) can control the overall operations of the electronic device (100). The processor (110) can control the overall operations that cause the electronic device (100) to estimate ridges within a fingerprint image and generate fingerprint information by executing one or more instructions of a program stored in the memory (120) of the electronic device (100). For example, the processor (110) can handle various tasks including processing arithmetic and logical operations of processes executed in the electronic device (100), managing the main memory (122), and transferring data to storage (124).
[0048] The processor (110) may include a processing circuit. The processing circuit may include an operation unit that performs arithmetic and logical operations, a control unit that interprets commands and controls execution, a register that stores data, and a cache memory, but is not limited thereto.
[0049] The processor (110) may be composed of at least one of, for example, a Central Processing Unit (CPU), a Microprocessor, a Graphic Processing Unit (GPU), ASICs (Application Specific Integrated Circuits), DSPs (Digital Signal Processors), DSPDs (Digital Signal Processing Devices), PLDs (Programmable Logic Devices), FPGAs (Field Programmable Gate Arrays), an Application Processor (AP), a Neural Processing Unit (NPU), or an AI-dedicated processor designed with a hardware structure specialized for processing AI models, but is not limited thereto.
[0050] In one embodiment, there may be one or more processors (110). If there is one or more processors (110), the operations of the present disclosure may be performed by one or more processors by executing instructions and / or programs stored in memory (120) individually or collectively. If the method according to one embodiment of the present disclosure includes a plurality of operations, the plurality of operations may be performed by one processor (110) or by a plurality of processors (110).
[0051] For example, when a first operation, a second operation, and a third operation are performed in an electronic device (100) by a method according to one embodiment, the first operation, the second operation, and the third operation may all be performed by a first processor, and some of the first to third operations may be performed by a first processor for processing general-purpose tasks and the remaining operations may be performed by a second processor for processing second machine learning-related tasks. However, the embodiments of the present disclosure are not limited thereto.
[0052] One or more processors (110) according to the present disclosure may be implemented as a single-core processor or as a multi-core processor. When a method according to one embodiment of the present disclosure includes a plurality of operations, the plurality of operations may be performed by a single core or by a plurality of cores included in one or more processors (110).
[0053] Various exemplary operations of the electronic device (100) performed by the processor (110) will be described in detail with reference to the drawings below.
[0054] The memory (120) can store data processed by the electronic device (100).
[0055] The memory (120) may include a main memory (122) that stores data currently being processed in the electronic device (100). The main memory (122) may store programs and data currently being executed by the processor (110) to allow the processor (110) to access the data quickly. The main memory (122) may include volatile memory such as, for example, RAM (Random Access Memory) or SRAM (Static Random Access Memory), but is not limited thereto.
[0056] The memory (120) may include storage (124) for permanently storing large amounts of data (e.g., programs, applications, system files, media files, etc.). The storage (124) may include non-volatile memory including at least one of, for example, a hard disk drive (HDD), a solid-state drive (SSD), an optical drive (e.g., a CD), a flash drive, a ROM (Read-Only Memory), an EEPROM (Electrically Erasable Programmable Read-Only Memory), and a PROM (Programmable Read-Only Memory), but is not limited thereto.
[0057] The display (130) can output a video signal to the screen of the electronic device (100) under the control of the processor (110). For example, the display (130) can display a graphic interface that guides fingerprint input for fingerprint recognition, and can output a video signal to the screen that indicates a security function (e.g., security content) provided to the authenticated user as the fingerprint is recognized.
[0058] The display (130) may include a touch panel. The touch panel may include one or more touch sensors that detect touch input. In one embodiment, user input for interacting with the electronic device (100) through the touch panel may be obtained.
[0059] The fingerprint sensor (140) may include, but is not limited to, a capacitive fingerprint sensor, an optical fingerprint sensor, an ultrasonic fingerprint sensor, etc. The fingerprint sensor (140) may be placed at a location such as a home button, a power button, the front or back of the electronic device (100) to sense contact with a user's finger and acquire a fingerprint image. In one embodiment, the fingerprint sensor (140) may be embedded in a display (130). In this case, the fingerprint sensor (140) may sense contact with a user's finger on the screen of the display (130) and acquire a fingerprint image.
[0060] Figure 4 is a diagram illustrating possible fingerprint recognition errors that may occur in a general fingerprint recognition method.
[0061] A general fingerprint recognition method may include processes for generating an enhanced fingerprint image (420) based on a fingerprint image (410) and generating fingerprint information from the enhanced fingerprint image (420).
[0062] For example, in a general fingerprint recognition method, a fingerprint image (410) is normalized, filtering (e.g., Gabor filter) is applied in multiple different directions to emphasize the ridges of the fingerprint, and the image is binarized to generate an enhanced fingerprint image (420) that strengthens the contrast between the ridges and valleys. In a general fingerprint recognition method, the enhanced fingerprint image (420) is thinned to make the ridges thinner, and unique feature points (e.g., bifurcations, endings) in the fingerprint are analyzed to generate fingerprint information (minutiae).
[0063] However, in the case of a general fingerprint recognition method, fingerprint recognition of a fingerprint image (410) in a normal finger condition is performed smoothly, but a problem arises where fingerprint information is generated inaccurately when the finger is dry or contains a wound. Referring to FIG. 4, when a general fingerprint recognition method is used, when an enhanced fingerprint image (420) is generated based on a fingerprint image (410), an exemplary result can be seen in which the branching point where the ridges split actually appears as the endpoint.
[0064] The fingerprint recognition method performed by the electronic device (100) of the present disclosure enables fingerprint recognition by robustly generating fingerprint information regardless of the condition of the finger (e.g., normal condition, dry condition, or condition with a wound). The electronic device (100) is characterized by being able to generate precise fingerprint information by using a ridge estimation algorithm to connect and restore missing or broken ridges, and by estimating a wound and utilizing the wound information. Below, operations of the fingerprint recognition method of the present disclosure, which differ from general fingerprint recognition methods, will be described.
[0065] FIG. 5 is a flowchart illustrating the operation of a fingerprint recognition model in one embodiment of the present disclosure.
[0066] In one embodiment, the electronic device (100) can acquire fingerprint information using a fingerprint recognition model. The fingerprint recognition model can estimate the ridges of the fingerprint using a ridge estimation algorithm. The fingerprint recognition model may be a Kalman filter based on a Hidden Markov Model (HMM), but is not limited thereto.
[0067] In operation 510, the electronic device (100) can acquire a fingerprint image and estimate ridges by tracing ridge paths within the fingerprint image. The electronic device (100) can estimate multiple ridges of the fingerprint by repeating the process of identifying a single ridge included in the fingerprint. For example, the electronic device (100) can identify a first single ridge by tracing a ridge path from a first starting point and identify a second single ridge by tracing a ridge path from a second starting point. In this case, the first single ridge and the second single ridge may be any one of the multiple ridges included in the fingerprint within the fingerprint image. The electronic device (100) can identify a single ridge by repeating operations 511 through 513. The electronic device (100) can identify a single ridge by starting from a starting point, repeating the process of predicting the next point based on the current point, and determining an end point.
[0068] In operation 511, the electronic device (100) can predict the center of the next point of the ridge path. The electronic device (100) can predict the next point of the ridge path based on the current point (e.g., the starting point) and the motion state parameters corresponding to the current point. The motion state parameters may include, for example, velocity and acceleration.
[0069] In operation 512, the electronic device (100) can adjust the ridge path by adjusting the position of the predicted point. The electronic device (100) can adjust the position of the predicted point based on the pixel values of the fingerprint image. The pixel values of the fingerprint image may have larger values (e.g., 255) as they represent ridges and smaller values (e.g., 0) as they represent valleys. The electronic device (100) can adjust the position of the predicted point and update the motion state parameters. In one embodiment, if no pixel representing a ridge is identified within a predetermined range from the next predicted point, the electronic device (100) may skip adjusting the position of the predicted point.
[0070] In operation 513, the electronic device (100) can determine the end point of a single ridge path. The electronic device (100) can determine the end point of a ridge path based on defined conditions. If the electronic device (100) determines that it is not the end point of a single ridge path, it can repeat operations 511 through 513 so that prediction and adjustment for subsequent points of the ridge path can be continued.
[0071] The defined conditions may include, for example, cases where the next point of the ridge path is not identified and it is determined that the ridge path no longer exists, or cases where the predicted next point is determined to be a point of a different ridge path other than the point of the ridge path currently being tracked. When the end point of a single ridge is determined, the electronic device (100) may update the path tracking status (e.g., end path tracking). When the end point of a single ridge is determined, the electronic device (100) may track another single ridge path starting from another starting point. To track another single ridge path from another starting point, the electronic device (100) may repeat operations 511 through 513.
[0072] In operation 520, the electronic device (100) can optimize the connectivity between ridges for a plurality of ridges identified through operation 510. For example, the electronic device (100) can optimize the connectivity between ridges through Maximum Likelihood Estimation. The process of the electronic device (100) optimizing the connectivity between a plurality of ridges is further described with reference to the drawings below.
[0073] In operation 530, the electronic device (100) can identify one or more scratches on the fingerprint within the fingerprint image. The electronic device (100) can include the identified scratches in the fingerprint information, or store the fingerprint with the identified scratches removed as fingerprint information.
[0074] In operation 540, the electronic device (100) can train a fingerprint recognition model. The fingerprint recognition model predicts the next point based on motion state parameters in each single ridge estimation process. The motion state parameters used for prediction may include velocity and acceleration and are treated as hidden parameters. Based on the data obtained in the aforementioned operations for fingerprint recognition, the electronic device (100) can learn related probability parameters so that the fingerprint recognition model can infer an observation value (e.g., a pixel value of a fingerprint image) based on the hidden parameters.
[0075] FIG. 6 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure estimating the ridge thickness of a fingerprint.
[0076] In one embodiment, the electronic device (100) can estimate the average thickness of the ridges within the fingerprint. To estimate the average thickness of the ridges within the fingerprint, the electronic device (100) can generate an arbitrary line across the fingerprint image. The electronic device (100) can generate brightness data representing pixel distance versus brightness by generating a line across the fingerprint image. In this case, in the brightness data, a value of brightness (pixel value) that is relatively large may represent a ridge, and a value of brightness that is relatively small may represent a valley. The electronic device (100) can convert the brightness data into the frequency domain to generate frequency data representing frequency versus intensity.
[0077] For example, as in the first example (610), the electronic device (100) can generate a first line across a fingerprint image. The electronic device (100) can identify pixel values of the fingerprint image along the first line, generate first brightness data (612) representing pixel distance versus brightness, and generate first frequency data (614) representing frequency versus intensity through a Fourier transform.
[0078] As in the second example (620), the electronic device (100) can generate a second line across the fingerprint image. The electronic device (100) can identify pixel values of the fingerprint image along the second line, generate second brightness data (622) representing pixel distance versus brightness, and generate second frequency data (624) representing frequency versus intensity through a Fourier transform.
[0079] In a similar manner, the electronic device (100) can generate arbitrary lines crossing a fingerprint image in each of multiple directions and perform measurement of pixel values and conversion to the frequency domain to obtain frequency data in each direction. The electronic device (100) can, for example, generate 36 lines rotating at 10-degree intervals to perform measurement of pixel values and conversion to the frequency domain, but is not limited to the example described above.
[0080] In one embodiment, the electronic device (100) can obtain average frequency data (630). For example, the electronic device (100) can calculate an average based on the first frequency data (614), the second frequency data (624), ..., the Nth frequency data. The electronic device (100) can estimate the thickness of the ridges within the fingerprint image using the average frequency data (630). For example, the electronic device (100) can identify frequency values within a predetermined range of high intensity within the average frequency data. Since frequency and period are inversely related, the electronic device (100) can identify the period based on the frequency values and identify the thickness of the ridges that appear periodically within the fingerprint image in pixel units.
[0081] In one embodiment, the electronic device (100) can track a ridge path based on ridge thickness. For example, the electronic device (100) can adjust the radius of a Gaussian function (640) for ridge path tracking and adjustment to correspond to the estimated ridge thickness. The radius of the Gaussian function (640) can be adjusted by variance, and the electronic device (100) can adjust the variance of the Gaussian function based on ridge thickness.
[0082] FIG. 7 is a drawing for explaining a fingerprint recognition model according to one embodiment of the present disclosure.
[0083] In one embodiment, the electronic device (100) can acquire fingerprint information using a fingerprint recognition model (700). The fingerprint recognition model (700) may be a Kalman filter based on a Hidden Markov Model (HMM). A Kalman filter based on a Hidden Markov Model (HMM) may be designed to estimate the next state from the previous state and correct the estimated value based on observations, and may acquire fingerprint recognition information by tracking each single ridge within the fingerprint image and identifying multiple ridges.
[0084] The fingerprint recognition model (700) may include position, velocity, and acceleration as hidden parameters.
[0085] The fingerprint recognition model (700) can repeat the process of estimating and correcting points of the ridge path along the ridge path from the starting point.
[0086] In operation 710, the fingerprint recognition model (700) can predict a point of the ridge path. The fingerprint recognition model (700) can predict the next point based on the position of the current point and the velocity and acceleration included in the motion state parameters. The initial position may be a starting point, and the initial velocity and acceleration may be arbitrary values or preset values.
[0087] In operation 720, the fingerprint recognition model (700) can adjust the path based on the next point predicted from the current point. The first operation of the process included in the path adjustment is a process of determining whether path adjustment is necessary, which is a process of determining whether the predicted next point corresponds to the observed value. This is identifying whether there is a point predicted to be on the ridge path, and may be referred to as a point identification process. The fingerprint recognition model (700) repeats the process of predicting points, and since this repetition is a repetition of predicting the next point based on the current point, the “predicted next point” may simply be referred to as the “predicted point” within this disclosure. The fingerprint recognition model (700) may use pixel values within the fingerprint image to determine whether the predicted point corresponds to the observed value. Within the fingerprint image, pixel values may have larger values (e.g., 255) as they represent ridges and smaller values (e.g., 0) as they represent valleys.
[0088] The fingerprint recognition model (700) may utilize maximum likelihood estimation. The fingerprint recognition model (700) may calculate a likelihood representing the probability of observed data occurring based on pixel values, which are a set of observed data representing ridges. For example, the fingerprint recognition model (700) may calculate likelihood values within a predetermined range from a predicted point. The fingerprint recognition model (700) may find values of position, velocity, and acceleration, which are hidden parameters that maximize and / or increase the likelihood values. If the likelihood values calculated during the maximum likelihood estimation process are below a threshold value, the fingerprint recognition model (700) may skip action 725, which is an action of adjusting the position of the next predicted point, and allow action 730, which is an action of determining whether to continue ridge path tracking, to be performed. If the likelihood value is above the threshold value during the maximum likelihood estimation process, the fingerprint recognition model (700) may identify hidden parameters that have the maximum likelihood among the likelihood values above the threshold value. The fingerprint recognition model (700) can perform operation 725 after selecting hidden parameters that have the maximum likelihood.
[0089] In operation 725, the fingerprint recognition model (700) can adjust the path based on the predicted point. A second operation of the process included in the path adjustment is centering. If hidden parameters (position, velocity, and acceleration) that have the maximum likelihood are identified, the fingerprint recognition model (700) can adjust the position of the predicted point to the identified position and update the hidden parameters. When the hidden parameters are updated, the fingerprint recognition model (700) can perform operation 730, which is an operation to determine whether to continue ridge path tracking.
[0090] In operation 730, the fingerprint recognition model (700) can determine whether to continue path tracking. The fingerprint recognition model (700) can update the path tracking status based on defined conditions.
[0091] For example, the fingerprint recognition model (700) can check a first condition indicating whether point prediction and adjustment have been performed through actions 720 and 725.
[0092] For example, if both operation 720 and operation 725 are performed (the first condition is satisfied), the fingerprint recognition model (700) can check the second condition.
[0093] When the fingerprint recognition model (700) repeats the process of identifying a single ridge, one or more identified single ridges may already exist. Therefore, the second condition may be whether the point where prediction and adjustment are performed is a point on the current ridge path (i.e., not a point on another ridge path).
[0094] If the point where prediction and adjustment are performed is a point on the current ridge path rather than a point on another ridge path (the second condition is satisfied), the fingerprint recognition model (700) can maintain the path tracking state as “continue path tracking” to continue path tracking. When the path tracking state is “continue path tracking,” the fingerprint recognition model (700) can perform operation 710 again to predict the next point on the ridge path.
[0095] If the point where prediction and adjustment were performed is identified as a point on a different ridge path (the second condition is not satisfied), the point where prediction and adjustment were performed is a point on a different ridge path, not a point on the ridge path currently being tracked. The fingerprint recognition model (700) can update the path tracking status to “path tracking ended.” In this case, the point prior to the last point where prediction and adjustment were performed can be determined as the end point of the ridge path. The fingerprint recognition model (700) can identify a path including points from the start point to the end point of the ridge path as a single ridge.
[0096] For example, if only operation 720 is performed and operation 725 is not performed (first condition not satisfied), the fingerprint recognition model (700) can check the third condition.
[0097] In the process of the fingerprint recognition model (700) identifying a single ridge, there may be cases where some ridges are missing even though fingerprint ridges actually exist due to the degradation of the quality of the fingerprint image. Therefore, the third condition may be whether the number of times point adjustment is skipped is less than a preset number to reflect cases where point prediction is performed but point adjustment is not performed due to missing observation values. The preset number may be, for example, 4 times, but is not limited thereto.
[0098] For example, if there is a point (skipped point) where only prediction is performed and no adjustment is performed, the fingerprint recognition model (700) may maintain the path tracking state as “continue path tracking” to continue path tracking. When the path tracking state is “continue path tracking,” the fingerprint recognition model (700) may perform operation 710 again to predict subsequent points of the ridge path.
[0099] If the number of times point adjustments are skipped is less than a preset number (the third condition is satisfied), the fingerprint recognition model (700) may maintain the path tracking status as “path tracking continues” to continue path tracking. If the number of times point adjustments are skipped reaches a preset number (the third condition is not satisfied), the fingerprint recognition model (700) may update the path tracking status to “path tracking ended.” In this case, the point prior to the point where the last prediction was performed may be determined as the end point of the ridge path. The fingerprint recognition model (700) may identify a path including points from the start point to the end point of the ridge path as a single ridge.
[0100] A fingerprint recognition model (700) can identify multiple ridges included in a fingerprint by repeating the process of identifying a single ridge from different starting points. An electronic device (100) can obtain fingerprint information for fingerprint recognition based on the multiple ridges identified using the fingerprint recognition model (700). The fingerprint information may be a minusha containing detailed feature information of the fingerprint. The minusha may include, for example, a branching point where a single ridge splits into two branches and an end point where a ridge ends, but is not limited thereto, and may include a scar, a point, a hole, an intersection point, a connection point, etc.
[0101] FIG. 8 is a diagram illustrating a motion state parameter used by an electronic device for fingerprint recognition according to one embodiment of the present disclosure.
[0102] In one embodiment, the electronic device (100) may use velocity and acceleration as motion state parameters. The electronic device (100) may track the path of ridges within a fingerprint image using motion state parameters.
[0103] Although the specific characteristics of fingerprints vary from person to person, the ridges and valleys included in fingerprints are generally curved. In other words, fingerprints contain various curves, and if the curvature of a curve is abrupt, errors may occur in predicting the next point of the ridge path.
[0104] A first example (810) of ridge path estimation illustrates the result of an electronic device (100) repeatedly predicting the next point of a ridge path using only speed when tracking a ridge path. Referring to the first example (810), when the electronic device (100) tracks a ridge path while repeating the process of predicting the next point of a ridge path, a prediction error (812) may occur in which the predicted next point is predicted to be a point of a different single ridge path, rather than a point of the current single ridge path. In other words, if the amount of change in speed is not reflected, when predicting the next point, the predicted point may deviate from the current ridge path being tracked in sections of sharp curvature.
[0105] The electronic device (100) can precisely track a single ridge containing a sharp curvature by applying acceleration at every point of ridge path tracking to reflect the amount of change in velocity. A second example (820) of ridge path estimation illustrates the result of the electronic device (100) repeatedly predicting the next point of the ridge path using velocity and acceleration when tracking the ridge path. Referring to the second example (820), a result can be shown in which different ridges are distinguished and identified even if there is a sharp curvature in the ridge. The electronic device (100) can identify each single ridge without deviating from the ridge path by updating the velocity using acceleration at every point and predicting the next point using velocity.
[0106] FIG. 9 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure tracking a ridge path.
[0107] The example in Fig. 9 is an example corresponding to the case where the first and second conditions described in Fig. 7 are satisfied and a single ridge path is tracked.
[0108] In one embodiment, the electronic device (100) can identify single ridges (910, 920, 930) included in the fingerprint using a fingerprint recognition model. The electronic device (100) can track the ridge path by repeating the process of predicting and adjusting points on the ridge path.
[0109] The electronic device (100) can update the hidden parameters of the fingerprint recognition model at every point of ridge path tracking. For example, the electronic device (100) can update the acceleration. For example, the electronic device (100) can update the velocity using the acceleration. For example, the electronic device (100) can update the position using the velocity.
[0110] For example, when an electronic device (100) identifies a first single ridge (910) within a fingerprint image, the electronic device (100) can predict a next point (914) based on the position, velocity, and acceleration of a current point (912) of the first single ridge (910) path. After predicting the next point (914), the electronic device (100) can adjust the position of the next point (914) and update its position, velocity, and acceleration. The parameters updated at the next point (914) can be used to predict the next point. The electronic device (100) can identify the first single ridge (910) included in the fingerprint by repeating the process of predicting subsequent points of the ridge path and updating parameters while adjusting their positions.
[0111] In the same way, the electronic device (100) can identify a plurality of total ridges included in the fingerprint image by identifying a second single ridge (920), a third single ridge (930), ... a Nth single ridge, which are single ridges different from the first single ridge (910).
[0112] FIG. 10a is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure for predicting points of a ridge path.
[0113] In describing FIG. 10a, for convenience of explanation, only one ridge (1000) among the multiple ridges included in the fingerprint is shown. Also, FIG. 10a shows the state at an intermediate point in the process of the electronic device (100) starting from a starting point and tracing the ridge path. The operation described in FIG. 10a may correspond to the operation 710 of FIG. 7.
[0114] In one embodiment, the electronic device (100) can predict the next point (1020) of the ridge path based on the position and motion state parameters of the current point (1010). The prediction of the next point (1020) from the current point (1010) by the electronic device (100) can be calculated as follows.
[0115]
[0116] Here, x' represents the position of the predicted next point (1020), and v represents the velocity at the current point (1010). represents a parameter for adjusting the prediction interval between points. The position and motion state parameters of the current point (1010) may be updated after predicting the current point (1010) from the previous point. For example, the velocity at the current point (1010) may be updated based on the velocity and acceleration at the previous point.
[0117] In one embodiment, the electronic device (100) has parameters for adjusting the prediction interval between points. It can be adjusted. The electronic device (100) can adjust the trade-off between prediction accuracy and efficiency by adjusting parameters for adjusting the prediction interval between points. For example, If becomes smaller, the prediction interval between points shortens and the accuracy of the prediction can be increased, and If it increases, the prediction interval between points increases, and the amount of computation can be reduced.
[0118] Meanwhile, referring to FIG. 10a, the location of the predicted next point (1020) is a point at a different location than a point on the path of the ridge (1000). The electronic device (100) can adjust the predicted next point (1020) so that the predicted next point (1020) corresponds to the path of the ridge. This is further described with reference to FIG. 10b.
[0119] FIG. 10b is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to check and adjust predicted points of a ridge path.
[0120] In describing FIG. 10b, for convenience of explanation, only one ridge (1000) among the multiple ridges included in the fingerprint is shown. Also, the next point (1020) (hereinafter, predicted point (1020)) predicted in FIG. 10a and the next point (1030) (hereinafter, adjusted point (1030)) whose position has been adjusted are shown in FIG. 10b. The operation described in FIG. 10b may correspond to operations 720 and 725 of FIG. 7.
[0121] In one embodiment, the electronic device (100) can adjust the predicted point (1020). The electronic device (100) can find values of position, velocity, and acceleration, which are hidden parameters that maximize the likelihood value, using maximum likelihood estimation. For example, the electronic device (100) can calculate likelihood values for a search range within a predetermined range from the predicted point. The electronic device (100) can obtain the likelihood value by multiplying a Gaussian function corresponding to the point and pixel values of a fingerprint image, which is a set of observed data representing a ridge.
[0122] For example, when calculating the likelihood value based on the location of the predicted point (1020), the likelihood value may be relatively low because there are no observed pixel values. When the electronic device (100) calculates the likelihood values for a search range within a predetermined range, and then calculates the likelihood value based on the location of the adjusted point (1030), the likelihood value may be highest because there are observed pixel values (operation 720 of FIG. 7).
[0123] In one embodiment, the electronic device (100) may determine whether to adjust the predicted point during the process of maximum likelihood estimation. Determining whether to adjust the predicted point (1020) means determining whether a ridge path exists, and if a ridge path does not exist, there is no need to adjust the predicted point (1020). In other words, determining whether to adjust the predicted point (1020) may be determined based on a threshold value for determining whether a ridge path exists.
[0124] For example, the electronic device (100) may determine that there is an actual observation point (ridge pixel) to be positioned around the predicted point (1020) if the likelihood value is greater than or equal to a threshold value. The electronic device (100) may determine that there is no actual observation point to be positioned around the predicted point (1020) if the likelihood value is less than or equal to a threshold value.
[0125] In one embodiment, the electronic device (100) can adjust a threshold value, which is a parameter for determining whether a ridge path exists. By adjusting the parameter for determining whether a ridge path exists, the electronic device (100) can adjust the trade-off between the accuracy and efficiency of ridge path tracking. For example, if the threshold value is reduced, a ridge path can be predicted even for areas with low observed pixel values (e.g., ridges with low pixel values due to shallow fingerprint contact), and if the threshold value is increased, a reliable ridge path prediction can be made by determining only areas with sufficiently high observed pixel values as ridge paths.
[0126] In one embodiment, the electronic device (100) can determine the radius of a Gaussian function for likelihood calculation. The radius of the Gaussian function can be adjusted by the variance, and the electronic device (100) can estimate the ridge thickness and adjust the variance of the Gaussian function so that the radius of the Gaussian function corresponds to the ridge thickness. In one embodiment, the electronic device (100) can adjust the radius of the Gaussian function for likelihood calculation. The electronic device (100) can adjust the radius of the Gaussian function to adjust the trade-off between the accuracy and efficiency of the likelihood calculation. For example, if the radius of the Gaussian function increases, the accuracy of the likelihood calculation may increase, and if the radius of the Gaussian function decreases, the amount of computation for the likelihood calculation may decrease.
[0127] In one embodiment, the electronic device (100) can identify a point where the likelihood value is maximized / increased using maximum likelihood estimation and determine a point (1030) adjusted to that location. The electronic device (100) can determine the adjusted point (1030) by adjusting the location of the predicted point (1020) and update the hidden parameters (operation 725 of FIG. 7). The updating of the hidden parameters by the electronic device (100) can be calculated as follows.
[0128]
[0129] Here, x' represents the position of the adjusted point (1030), and represents a parameter for adjusting the prediction interval between points, where v' represents the updated velocity and a' represents the updated acceleration.
[0130] In one embodiment, after determining an adjusted point (1030), the electronic device (100) can estimate the path of a ridge (1000) by performing a prediction for the next point of the adjusted point. The process of the electronic device (100) repeating the prediction and adjustment of the next point to estimate the path of the ridge (1000) can be performed until a defined condition for determining the end point of the ridge path is satisfied.
[0131] FIG. 11 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure for predicting points of a ridge path.
[0132] In describing FIG. 11, for convenience of explanation, only one ridge (1100) among the multiple ridges included in the fingerprint is illustrated. Additionally, states at an intermediate point in time are illustrated during the process in which the electronic device (100) starts from a starting point and tracks the ridge path. The operations described in FIG. 11 correspond to the operations described in the description of FIG. 7 and are intended to explain how to handle exception cases that occur during the ridge path tracking process. In other words, the example in FIG. 11 corresponds to the case where the first condition described in FIG. 7 is not satisfied and a single ridge path is tracked while operating according to the third condition.
[0133] In one embodiment, the quality of a fingerprint image obtained through a fingerprint sensor, etc., depends on the condition of the finger (e.g., wet condition, dry condition, etc.). For example, when fingerprint recognition is performed with a dry finger, the quality of the fingerprint image may deteriorate, such as some ridges being missing even though fingerprint ridges actually exist. In other words, there may be areas within the fingerprint image where pixel values corresponding to ridges are not obtained. The electronic device (100) can determine the areas where pixel values are missing as ridges and generate ridge paths through a ridge estimation algorithm.
[0134] For example, the electronic device (100) can predict a second point (1120), which is the next point, from a first point (1110) determined during the ridge path tracking process. When the second point (1120) is predicted, the electronic device (100) may apply maximum likelihood estimation to a search range within a predetermined range based on the predicted point to check whether the predicted point corresponds to an observed value. If the likelihood values calculated in the maximum likelihood estimation are below a threshold value, the electronic device (100) may determine that no actual observed point is identified around the predicted point. If no observed point corresponding to the predicted point is identified, the electronic device (100) may skip adjusting the location of the predicted point. In this case, the electronic device (100) may check a defined condition for determining the end point of the ridge path.
[0135] If the defined conditions for determining the end point of the ridge path are not met, the electronic device (100) can again predict the next point, the third point (1130), from the second point (1120). If the observed point corresponding to the third point (1130) is not identified, the adjustment of the position of the third point (1130) is skipped again and the fourth point (1140) can be predicted.
[0136] If the defined conditions for determining the end point of the ridge path are not met, the electronic device (100) can again predict the next point, the fifth point (1150), from the fourth point (1140). In the example of FIG. 11, as a result of applying maximum likelihood estimation based on the predicted fifth point (1150), pixel data, which is an observation representing the ridge (1100), exists so that the point can be identified. Once the point is identified, the electronic device (100) can adjust the position of the fifth point (1150) based on maximum likelihood estimation and update the hidden parameters, position, velocity, and acceleration.
[0137] The electronic device (100) may repeat the prediction and position adjustment of the next point, and if the observed value of the next point is not identified, skip the position adjustment of the next point and perform a prediction for the subsequent point. If the observed value of the point is identified in the prediction for the subsequent point, the electronic device (100) may adjust the position of the subsequent point and determine a ridge path including the skipped points. In the example of FIG. 11, as a result of estimating the ridge (1100), the ridge path may include a first point (1110), a second point (1120), a third point (1130), a fourth point (1140), and a fifth point (1150). Here, the first point (1110) and the fifth point (1150) are points whose positions have been adjusted because the points have been identified, and the second point (1120), the third point (1130), and the fourth point (1140) represent points whose positions have been skipped because the points have not been identified.
[0138] FIG. 12 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure determining the end point of a ridge path.
[0139] The example of FIG. 12 corresponds to a case where the first condition described in FIG. 7 is not satisfied and the ridge end point is determined by operating according to the third condition. In one embodiment, the electronic device (100) can determine the end point of the ridge path based on defined conditions. For example, the electronic device (100) can determine the last positioned point as the end point of the ridge path when the number of points whose position adjustments have been skipped exceeds a threshold value.
[0140] The electronic device (100) may repeat prediction and adjustment of points on a ridge path to identify a single ridge. For example, the electronic device (100) may determine a second point (1212) through a point prediction and adjustment process using a first point (1211) and a motion state vector, determine a third point (1213) through a point prediction and adjustment process using the second point (1212) and a motion state vector, and determine a fourth point (1214) through a point prediction and adjustment process using the third point and a motion state vector. In the manner described above, the electronic device (100) may perform a prediction for a subsequent point using the fourth point (1214) and a motion state, and as a result, the predicted location of the subsequent point may be obtained.
[0141] The electronic device (100) can perform point identification to determine whether the predicted point corresponds to an observed value when a point is predicted. For example, the electronic device (100) can calculate likelihood values within a predetermined range from the predicted point using maximum likelihood estimation to find values of position and motion state parameters (velocity and acceleration) that cause the location of the predicted point to correspond to the center of the pixels representing the ridge. In this case, the electronic device (100) can find a point that has the maximum likelihood while the likelihood values are greater than or equal to a threshold value. However, if the likelihood values are less than the threshold value, the electronic device (100) can determine that the point has not been identified and skip adjusting the location of the predicted point.
[0142] As a result, the electronic device (100) can determine the subsequent point of the fourth point (1214) as the first skipped point (1221). The location of the first skipped point (1221) can correspond to the location predicted based on the fourth point (1214). In the same way, as a result of the electronic device (100) performing a prediction and adjustment process for subsequent points after the first skipped point (1221), the second skipped point (1222), the third skipped point (1223), and the fourth skipped point (1224) can be identified, for which the point location is not adjusted because the point is not identified.
[0143] Meanwhile, points that are not identified may be due to the low quality of the fingerprint image, where pixel values are not measured in the fingerprint image even though ridges actually exist, or where the ridges actually end. In the former case, if point identification and position adjustment are performed again at a subsequent point, it is determined that the ridge was missing from the fingerprint image, and ridge path tracing continues (see FIG. 11). In the latter case, if the number of points where position adjustment was skipped exceeds a threshold value, it is determined as the end point of the ridge. For example, as shown in FIG. 12, when the threshold value is 3, the number of points where position adjustment was skipped exceeds 3, so the electronic device (100) can determine the fourth point (1214), which is the last point where position adjustment was performed, as the end point of the ridge path. In FIG. 12, the threshold value for skipped points to determine the end point of the ridge path is described as 3, which should be understood as an exemplary value.
[0144] FIG. 13 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure determining the end point of a ridge path.
[0145] The example of FIG. 13 corresponds to a case where the first condition described in FIG. 7 is not satisfied and the operation is performed according to the third condition, and the ridge end point is determined according to the second condition. In one embodiment, the electronic device (100) can determine the end point of the ridge path based on defined conditions.
[0146] The electronic device (100) can repeat prediction and adjustment for points of a ridge path to identify a single ridge. If the next point is not identified during the process of repeating prediction and adjustment for points of a ridge path, the electronic device (100) can skip adjusting the position of the next point and perform prediction and adjustment again for the subsequent point of the skipped point.
[0147] For example, in the process of the electronic device (100) repeating point prediction and adjustment for the first single ridge (1310), the first skipped point (1311), the second skipped point (1312), and the third skipped point (1313) may be identified. In this case, when the electronic device (100) performs a prediction for the subsequent point of the third skipped point (1313), the result of the point identification may be a point on another single ridge, the second single ridge (1320), rather than the first single ridge (1310). In this case, the electronic device (100) may determine the position-adjusted point (1314) prior to the first skipped point (1311) as the end point of the first single ridge (1310).
[0148] For example, the electronic device (100) may determine the last position-adjusted point as the end point of the ridge path if the angle between the ridge path including the connecting line of the position-adjusted skipped point and the ridge path of another single ridge exceeds a threshold value. For example, if the angle between the path of the first single ridge (1310) including the connecting line of the first skipped point (1311), the second skipped point (1312), and the third skipped point (1313) and the path of the second single ridge (1320) exceeds a threshold value, the electronic device (100) may determine the last position-adjusted point (1314) as the end point of the first single ridge (1310). The electronic device (100) may determine the end point of the single ridge using a motion state vector. Since the motion state vector of the first single ridge (1310) is different in direction from the motion state vector of the second single ridge (1320), the electronic device (100) can determine the last positioned point (1314) as the end point of the first single ridge (1310) when the angle of direction between the two vectors exceeds a threshold value.
[0149] FIG. 14 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure determining starting points for ridge path tracking.
[0150] In one embodiment, the electronic device (100) may determine one or more starting points within a fingerprint image to apply a ridge estimation algorithm. The electronic device (100) may identify a single ridge included within the fingerprint image by repeating point prediction and adjustment from each starting point using the ridge estimation algorithm. The electronic device (100) may identify multiple ridges within the fingerprint image by repeating the process of identifying each single ridge using the ridge estimation algorithm.
[0151] The electronic device (100) can acquire a fingerprint image. The electronic device (100) can perform peak detection on the edges of the fingerprint image. For example, the electronic device (100) can scan and store pixel values for each of the top, bottom, left, and right edges of the fingerprint image and detect peaks that indicate areas with large pixel intensity changes at each edge. The electronic device (100) can identify the coordinates within the fingerprint image of the detected peaks for each edge and determine the identified coordinates as starting points for ridge path tracing.
[0152] The electronic device (100) can track ridge paths starting from starting points. For example, the electronic device (100) can identify single ridges extending from each starting point by performing ridge path tracking from initial starting points in a direction perpendicular to each edge. Initial motion state parameters (velocity, acceleration) for ridge path tracking can be defined values or arbitrary values. As the electronic device (100) tracks the ridge path, the motion state parameters at each point can be updated by a fingerprint recognition model using a ridge estimation algorithm.
[0153] In one embodiment, the electronic device (100) may determine one or more starting points by selecting arbitrary points within the fingerprint image instead of determining starting points at the edges of the fingerprint image. For example, the electronic device (100) may determine any N starting points within the fingerprint image. In this case, even if the initial starting point does not correspond to a pixel of a ridge, the point may be adjusted to correspond to a pixel of a ridge through the point prediction and adjustment process of the ridge estimation algorithm. If the starting point is determined at an arbitrary location other than the edges of the fingerprint image, the electronic device (100) may identify multiple ridges by determining single ridges by tracing the ridge path in both directions from the starting point.
[0154] FIG. 15 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure optimizing the connection between ridges.
[0155] In one embodiment, the electronic device (100) can optimize the connection between ridges by applying maximum likelihood estimation after repeating the identification of single ridges through a ridge estimation algorithm. The electronic device (100) can estimate the most likely direction of line extension from the estimated single ridges using maximum likelihood estimation.
[0156] For example, as a result of the electronic device (100) performing identification of single ridges, the first ridge B (1515) of the first ridge A (1510) may exist as an identification result. The electronic device (100) may generate a probabilistic model of the broken portion of the first ridge A (1510) based on the ridge path, curvature, direction, and end point of the first ridge A (1510). The electronic device (100) may estimate the direction in which the broken portion of the first ridge A (1510) will be connected, and may find the most likely direction through maximum likelihood estimation by modeling the direction data as a Gaussian distribution. In this case, the first ridge B (1515) is used as an observation value, and the likelihood that the first ridge A (1510) will be connected to the first ridge B (1515) is evaluated. However, the foregoing is merely an example, and maximum likelihood estimation may be used to find a ridge that is likely to be connected to the first ridge B (1515) using the first ridge A (1510) as the observed value, and maximum likelihood estimation may be applied to each single ridge.
[0157] The electronic device (100) can find single ridges with a high probability of connection using maximum likelihood estimation, and if single ridges with a high probability of connection are identified, it can perform an optimization task to connect the ridges through pixel interpolation. In the example of FIG. 15, as a result of the electronic device (100) optimizing the connection between the ridges, the first single ridge A (1510) can be connected to the first single ridge B (1515), the second single ridge A (1520) can be connected to the second single ridge B (1525), and the third single ridge A (1530) can be connected to the third single ridge B (1535).
[0158] FIG. 16 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure optimizing the connection between ridges.
[0159] Referring to FIG. 16, the first example (1610) shows the result of performing only point prediction and position adjustment of the ridge path from multiple starting points without determining the end point of the ridge path and optimizing the connection between the ridges. Referring to the first example (1610), during the process of tracking each ridge in the fingerprint image, a result appears in which some ridge paths deviate, or a result appears in which a path that is not actually a ridge is determined as a ridge.
[0160] The electronic device (100) can obtain fingerprint information by precisely identifying multiple ridges through determining the end point of the ridge path and optimizing the connection between the ridges. Referring to the second example (1620), the electronic device (100) can prevent or reduce the abnormal extension of a single ridge by determining the end point of the ridge based on defined conditions. For example, when tracking a single ridge, the electronic device (100) can determine the end point of the single ridge path and terminate the ridge path tracking if the number of points for which position adjustment was skipped exceeds a threshold value (e.g., 2). For example, when tracking a single ridge, the electronic device (100) can determine the end point of the single ridge path and terminate the ridge path tracking if the angle between the ridge path containing the connection line of the points for which position adjustment was skipped and the ridge path of another single ridge exceeds a threshold value. The electronic device (100) can repeat the identification of single ridges and, after all ridges have been identified, optimize the connections between ridges so that single ridges with a high probability of connection can be additionally connected. Accordingly, in the second example (1620), a result may appear in which ridges within the fingerprint image are identified more accurately than in the first example (1610).
[0161] FIG. 17 is a diagram illustrating the operation of an electronic device according to one embodiment of the present disclosure to estimate a fingerprint scratch.
[0162] In one embodiment, the electronic device (100) can estimate a fingerprint scar (1700). The fingerprint scar (1700) may refer to a pattern that constitutes fingerprint information in a form different from the ridges within the fingerprint. For example, the scar (1700) may refer to a pattern caused by temporary or permanent skin damage. The scar (1700) may include patterns such as fine creases or folds that occur in areas where the finger is frequently folded or pressure is applied, which are unrelated to skin damage.
[0163] The electronic device (100) can identify points where position adjustment has been skipped after repeating the identification of a single ridge through a ridge estimation algorithm. For example, the electronic device (100) can identify areas where skipped points are densely concentrated within a fingerprint image using hierarchical clustering. The electronic device (100) can group skipped points within adjacent distances into clusters by measuring the distance between skipped points (e.g., Euclidean distance).
[0164] The electronic device (100) can estimate a fingerprint scar (1700) based on a cluster containing points where position adjustment has been skipped. For example, the electronic device (100) can perform linear regression on a cluster containing skipped points identified through hierarchical clustering. The electronic device (100) can obtain a regression line corresponding to the cluster to obtain a line representing the distribution of the skipped points. The electronic device (100) can determine the line identified through linear regression as the scar (1700). There may be one or more scars (1700).
[0165] In one embodiment, the electronic device (100) may include the identified scar (1700) in the fingerprint information. For example, the electronic device (100) may generate fingerprint information including branching points, endpoints, etc. based on a plurality of ridges. In this case, a line representing the scar (1700) may be included in the fingerprint information.
[0166] In one embodiment, the electronic device (100) can store the fingerprint with the identified wound (1700) removed as fingerprint information. For example, the electronic device (100) can remove the identified wound (1700) and restore the broken ridges by joining them together, and then generate fingerprint information including branching points, endpoints, etc. based on the restored ridges.
[0167] FIG. 18 is a diagram illustrating the operation of an electronic device performing fingerprint recognition according to one embodiment of the present disclosure.
[0168] In one embodiment, the electronic device (100) can perform fingerprint recognition by generating fingerprint information based on a fingerprint image. The fingerprint recognition process may include performing biometric authentication based on the acquired fingerprint image.
[0169] The electronic device (100) can identify the ridges (1810) of a fingerprint by using a ridge estimation algorithm to track ridge paths from one or more starting points within the fingerprint image. Additionally, the electronic device (100) can identify the scratches (1820) of the fingerprint based on clusters of skipped points identified during the ridge estimation algorithm process. The electronic device (100) can obtain fingerprint information (1830) for fingerprint recognition based on the ridges (1810) and scratches (1820). The fingerprint information (1830) may be a minusha containing detailed feature information of the fingerprint. The minusha may include, for example, a branching point where a single ridge splits into two branches and an end point where a ridge ends, but is not limited thereto, and may include scratches, islands, holes, intersections, connection points, etc.
[0170] In one embodiment, the electronic device (100) can perform biometric authentication by comparing fingerprint information (1830) with fingerprint authentication information (1840). The fingerprint authentication information (1840) may be information for biometric authentication that is pre-registered by a user of the electronic device (100) and stored in the electronic device (100). The electronic device (100) can determine whether the fingerprint information (1830) corresponds to the fingerprint authentication information (1840) by performing pattern matching of the fingerprint information (1830) and the fingerprint authentication information (1840). If the fingerprint information (1830) corresponds to the fingerprint authentication information (1840), the electronic device (100) can complete biometric authentication and enable functions that can be provided after authentication (e.g., unlocking, access to secure files, payment, financial services, etc.) to be performed.
[0171] FIG. 19 is a diagram illustrating the operation of an electronic device performing fingerprint recognition according to one embodiment of the present disclosure.
[0172] In one embodiment, the electronic device (100) can perform fingerprint recognition using a plurality of fingerprint recognition models. For example, the electronic device (100) can use a first fingerprint recognition model (e.g., fingerprint recognition model A (1910)) and a second fingerprint recognition model (e.g., fingerprint recognition model B (1920)).
[0173] Fingerprint recognition model A (1910) may be a model that receives a fingerprint image (1900) as input and outputs fingerprint information (1915) including a scar. Fingerprint recognition model A (1910) may be a model trained using fingerprint information (1915) including a scar as training data. Fingerprint recognition model A (1910) may estimate the ridges of the fingerprint by tracing ridge paths from one or more starting points within the fingerprint image (1900). Estimating the ridge paths of the fingerprint may involve repeating the process of estimating the ridge path of a single ridge. Estimating the ridge path of a single ridge may involve repeating the process of predicting and adjusting the next point of the ridge path. In this case, if the next point is not identified, the position adjustment may be skipped and a prediction for the subsequent point may be performed. Fingerprint recognition model A (1910) may obtain fingerprint scar information based on the skipped points identified during the ridge path estimation process.
[0174] Fingerprint recognition model B (1920) may be a model that receives a fingerprint image (1900) as input and outputs fingerprint information (1925) with restored scars. Fingerprint recognition model B (1920) may be a model trained using fingerprint information (1925) with restored scars (or fingerprint information without scars) as training data. Fingerprint recognition model B (1920) may estimate the ridges of the fingerprint by tracing ridge paths from one or more starting points within the fingerprint image (1900). Estimating the ridge paths of the fingerprint may involve repeating the process of estimating the ridge path of a single ridge. Estimating the ridge path of a single ridge may involve repeating the process of predicting and adjusting the next point of the ridge path. In this case, if the next point is not identified, the position adjustment may be skipped and a prediction for the subsequent point may be performed. Fingerprint recognition model B (1920) can obtain fingerprint scar information based on skipped points identified during the ridge path estimation process. When the fingerprint recognition model B (1920) identifies the scar information, it can obtain fingerprint information (1925) with the scar restored.
[0175] Fingerprint recognition model A (1910) can improve the fingerprint recognition rate for fingerprints containing scars by increasing the number of detailed features in the fingerprint information due to the scar information included in the fingerprint information. However, since the scar is included in the fingerprint information, there is a possibility of misrecognition of the fingerprint if a forged scar is included, such as by attaching a sticker to the finger. Therefore, in order to reduce cases of misrecognition while ensuring the recognition performance of fingerprints containing scars, the electronic device (100) can perform fingerprint recognition by combining the results of multiple fingerprint recognition models. For example, when a single fingerprint image (1900) is input, the electronic device (100) can perform biometric authentication by combining the results using different fingerprint recognition models. For example, the electronic device (100) may input a fingerprint image (1900) into a fingerprint recognition model A (1910) to obtain fingerprint information (1915) including a scar, and calculate a first score indicating the degree of pattern matching by comparing the fingerprint information (1915) including the scar with fingerprint authentication information stored in the electronic device (100). The electronic device (100) may input a fingerprint image (1900) into a fingerprint recognition model B (1920) to obtain fingerprint information (1925) with the scar restored, and calculate a second score indicating the degree of pattern matching by comparing the fingerprint information (1925) with the fingerprint authentication information stored in the electronic device (100). The electronic device (100) may combine the first score and the second score using an ensemble technique. For example, the electronic device (100) may calculate a final score for fingerprint authentication using the average of the first score and the second score or the weighted average of the first score and the second score. If the final score for fingerprint authentication is greater than or equal to a threshold value, the electronic device (100) can determine that the biometric authentication was successful.
[0176] The present disclosure relates to a method for performing fingerprint recognition by robustly generating fingerprint information from a fingerprint image regardless of the condition of the finger, and an electronic device for performing the method. The technical problems to be solved by the present disclosure are not limited to those mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art to which the present invention belongs from the description in this specification.
[0177] According to one aspect of the present disclosure, a fingerprint recognition method performed by an electronic device may be provided.
[0178] The above method may include the step of acquiring a fingerprint image.
[0179] The above method may include the step of estimating the ridges of the fingerprint represented by the fingerprint image by tracing ridge paths from one or more starting points within the fingerprint image through a ridge estimation algorithm.
[0180] The above method may include the step of obtaining fingerprint information for fingerprint recognition based on the estimated ridges.
[0181] The ridge estimation algorithm may include, for each of the ridges of the fingerprint: a step of predicting the next point of a ridge path based on a current point and a motion state parameter; a step of adjusting the predicted next point and updating the motion state parameter based on the predicted next point and pixel values of the fingerprint image; and a step of identifying a single ridge by tracking the ridge path through the repetition of prediction and adjustment for subsequent points of the ridge path.
[0182] The above motion state parameters may include velocity and acceleration.
[0183] The step of adjusting the predicted next point and updating the motion state parameters may involve updating the motion state parameters through maximum likelihood estimation and adjusting the position of the predicted next point so that the predicted next point corresponds to a pixel within the ridge path.
[0184] The step of adjusting the predicted next point and updating the motion state parameters may include skipping the step of adjusting the position of the predicted next point when the likelihood values calculated in the maximum likelihood estimation process are below a threshold value.
[0185] The step of identifying the single ridge above may include performing prediction and adjustment for the subsequent point of the point where the position adjustment was skipped.
[0186] The step of identifying the single ridge may include the step of determining the end point of the ridge path based on defined conditions.
[0187] The step of determining the end point of the ridge path may be to determine the last position-adjusted point as the end point of the ridge path when the number of points for which position adjustment was skipped exceeds a threshold value.
[0188] The step of determining the end point of the ridge path may be to determine the last position-adjusted point as the end point of the ridge path when the angle between the ridge path including the connecting line of the point where the position adjustment was skipped and the ridge path of another single ridge exceeds a threshold value.
[0189] The above method may include the step of estimating the average thickness of the ridges from the fingerprint image.
[0190] The above method may include the step of determining a Gaussian distribution function used in the maximum likelihood estimation for updating the motion state parameter based on the average thickness of the ridge.
[0191] The step of estimating the ridges may include the step of optimizing the connections between the ridges by applying maximum likelihood estimation after repeating the identification of the single ridge through the ridge estimation algorithm.
[0192] The above method may include the step of estimating the scar of the fingerprint based on a cluster including points where the position adjustment was skipped through the ridge estimation algorithm.
[0193] The above method may include a step of performing fingerprint recognition based on the fingerprint information and previously stored fingerprint authentication information.
[0194] According to one aspect of the present disclosure, an electronic device for performing fingerprint recognition may be provided. The electronic device may include at least one processor comprising a processing circuit, a memory for storing instructions, and a display.
[0195] By executing the above instructions by the at least one processor, the electronic device can acquire a fingerprint image.
[0196] By executing the above instructions by the at least one processor, the electronic device can estimate the ridges of the fingerprint represented by the fingerprint image by tracing ridge paths from one or more starting points within the fingerprint image through a ridge estimation algorithm.
[0197] By executing the above instructions by the at least one processor, the electronic device can obtain fingerprint information for fingerprint recognition based on the estimated ridges.
[0198] The ridge estimation algorithm may include, for each of the ridges of the fingerprint: predicting the next point of the ridge path based on a current point and a motion state parameter, adjusting the predicted next point and updating the motion state parameter based on the predicted next point and pixel values of the fingerprint image, and identifying a single ridge by tracking the ridge path through the repetition of prediction and adjustment for subsequent points of the ridge path.
[0199] By executing the above instructions by the at least one processor, the electronic device can control the display to display a fingerprint recognition result based on the fingerprint information.
[0200] The above motion state parameters may include velocity and acceleration.
[0201] By executing the above instructions by the at least one processor, the electronic device may update the motion state parameters through maximum likelihood estimation and adjust the position of the predicted next point so that the predicted next point corresponds to a pixel within the ridge path.
[0202] By executing the above instructions by the at least one processor, the electronic device can skip adjusting the position of the predicted next point when the likelihood values calculated in the maximum likelihood estimation process are below a threshold value.
[0203] By executing the above instructions by the at least one processor, the electronic device can perform prediction and adjustment for a subsequent point of the point where the position adjustment was skipped.
[0204] By executing the above instructions by the at least one processor, the electronic device can determine the end point of the ridge path based on defined conditions.
[0205] By executing the above instructions by the at least one processor, the electronic device can determine the last position-adjusted point as the end point of the ridge path when the number of points for which the position adjustment has been skipped exceeds a threshold value.
[0206] By executing the above instructions by the at least one processor, the electronic device can determine the last position-adjusted point as the end point of the ridge path when the angle between the ridge path including the connecting line of the point where the position adjustment was skipped and the ridge path of another single ridge exceeds a threshold value.
[0207] By executing the above instructions by the at least one processor, the electronic device can determine a Gaussian distribution function used for the maximum likelihood estimation to update the motion state parameter based on the average thickness of the ridge.
[0208] By executing the above instructions by the at least one processor, the electronic device can optimize the connection between the ridges by applying maximum likelihood estimation after repeating the identification of the single ridge through the ridge estimation algorithm.
[0209] By executing the above instructions by the at least one processor, the electronic device can estimate the scratch of the fingerprint based on a cluster including points where the position adjustment was skipped through the ridge estimation algorithm.
[0210] By executing the above instructions by the at least one processor, the electronic device can perform fingerprint recognition based on the fingerprint information and previously stored fingerprint authentication information.
[0211] Meanwhile, embodiments of the present disclosure may also be implemented in the form of a recording medium containing computer-executable instructions, such as program modules executed by a computer. A computer-readable medium may be any available medium accessible by a computer and includes both volatile and non-volatile media, and both removable and non-removable media. Additionally, a computer-readable medium may include computer storage media and communication media. Computer storage media include both volatile and non-volatile, removable and non-removable media implemented by any method or technique for storing information, such as computer-readable instructions, data structures, program modules, or other data. Communication media may typically include other data of modulated data signals, such as computer-readable instructions, data structures, or program modules.
[0212] Additionally, computer-readable storage media may be provided in the form of non-transitory storage media. Here, 'non-transitory storage media' simply means that it is a tangible device and does not contain a signal (e.g., electromagnetic waves), and this term does not distinguish between cases where data is stored semi-permanently and cases where it is stored temporarily. For example, 'non-transitory storage media' may include a buffer in which data is stored temporarily.
[0213] According to one embodiment, the method according to the various embodiments disclosed herein may be provided by being included in a computer program product. The computer program product may be traded between a seller and a buyer as a product. The computer program product may be distributed in the form of a device-readable storage medium (e.g., compact disc read-only memory (CD-ROM)), or distributed online (e.g., download or upload) through an application store or directly between two user devices (e.g., smartphones). In the case of online distribution, at least a portion of the computer program product (e.g., a downloadable app) may be temporarily stored or temporarily created on a device-readable storage medium, such as the memory of a manufacturer's server, an application store's server, or a relay server.
[0214] The foregoing description of the present disclosure is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical spirit or essential features of the present disclosure. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.
[0215] The scope of the present disclosure is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof should be interpreted as being included within the scope of the present disclosure.
Claims
1. A fingerprint recognition method performed by an electronic device, Step of acquiring a fingerprint image; A step of estimating the ridges of the fingerprint in the fingerprint image by tracing ridge paths from one or more starting points in the fingerprint image through a ridge estimation algorithm (510); and The method includes the step of obtaining fingerprint information for fingerprint recognition based on the estimated ridges above, The above ridge estimation algorithm (510) is, for each of the ridges of the fingerprint: A step (511) of predicting the next point of the ridge path based on the current point and motion state parameters; Step (512) of adjusting the predicted next point and updating the motion state parameters based on the predicted next point and the pixel values of the fingerprint image; and A method comprising the step (513) of identifying a single ridge by tracking the ridge path through iterative prediction and adjustment of subsequent points of the ridge path.
2. In Paragraph 1, The above motion state parameters include velocity and acceleration, and The step of adjusting the predicted next point and updating the motion state parameters is, A method of updating the motion state parameters through maximum likelihood estimation and adjusting the position of the predicted next point so that the predicted next point corresponds to a pixel within the ridge path.
3. In Paragraph 2, The step of adjusting the predicted next point and updating the motion state parameters is, Based on the likelihood values calculated in the maximum likelihood estimation process being less than a threshold value, the method includes the step of skipping the adjustment of the predicted next point's position. The step of identifying the single ridge above is, A method in which the above position adjustment performs prediction and adjustment for a subsequent point of a skipped point.
4. In Paragraph 3, The step of identifying the single ridge above is, A method comprising the step of determining the end point of the ridge path based on defined conditions.
5. In Paragraph 4, The step of determining the end point of the above ridge path is, A method of determining the last position-adjusted point as the end point of the ridge path based on the fact that the number of skipped points in the position adjustment exceeds a threshold value.
6. In Paragraph 4, The step of determining the end point of the above ridge path is, A method of determining the last position-adjusted point as the end point of the ridge path based on the angle between the ridge path including the connecting line of the point where the position adjustment was skipped and the ridge path of another single ridge exceeding a threshold value.
7. In Paragraph 3, The above method is, A method further comprising the step of estimating the scar of the fingerprint based on a cluster including points where the position adjustment was skipped through the ridge estimation algorithm.
8. In an electronic device (100) that performs fingerprint recognition, At least one processor (110) including a processing circuit; Memory (120) for storing instructions; and Includes a display (130), By executing the above instructions by the at least one processor (110), the electronic device (100) Acquire a fingerprint image, and Estimating the ridges of the fingerprint in the fingerprint image by tracing ridge paths from one or more starting points within the fingerprint image using a ridge estimation algorithm, and Fingerprint information for fingerprint recognition is obtained based on the estimated ridges above, The above ridge estimation algorithm, for each of the ridges of the fingerprint: Based on the current point and motion state parameters, predict the next point of the ridge path, and Based on the predicted next point and the pixel values of the fingerprint image, the predicted next point is adjusted and the motion state parameters are updated, An electronic device (100) that identifies a single ridge by tracking the ridge path through the repetition of prediction and adjustment of subsequent points of the ridge path.
9. In Paragraph 8, The above motion state parameters include velocity and acceleration, and By executing the above instructions by the at least one processor, the electronic device, An electronic device that updates the motion state parameters through maximum likelihood estimation and adjusts the position of the predicted next point so that the predicted next point corresponds to a pixel within the ridge path.
10. In Paragraph 9, By executing the above instructions by the at least one processor, the electronic device, Based on the fact that the likelihood values calculated in the above maximum likelihood estimation process are below a threshold value, the adjustment of the location of the predicted next point is skipped, and An electronic device that performs prediction and adjustment for a subsequent point of the above-mentioned position adjustment that was skipped.
11. In Paragraph 10, By executing the above instructions by the at least one processor, the electronic device, An electronic device that determines the end point of the ridge path based on defined conditions.
12. In Paragraph 11, By executing the above instructions by the at least one processor, the electronic device, An electronic device that determines the last position-adjusted point as the end point of the ridge path based on the fact that the number of points skipped in the position adjustment exceeds a threshold value.
13. In Paragraph 11, By executing the above instructions by the at least one processor, the electronic device, An electronic device that determines the last position-adjusted point as the end point of the ridge path based on the angle between the ridge path including the connecting line of the point where the position adjustment was skipped and the ridge path of another single ridge exceeding a threshold value.
14. In Paragraph 8, By executing the above instructions by the at least one processor, the electronic device, An electronic device that estimates the scar of the fingerprint based on a cluster including points where the position adjustment was skipped through the ridge estimation algorithm.
15. A computer-readable recording medium having a program for executing the method of any one of paragraphs 1 through 7 on a computer.
