CN212569821U - Non-contact fingerprint acquisition device - Google Patents
Non-contact fingerprint acquisition device Download PDFInfo
- Publication number
- CN212569821U CN212569821U CN202022198405.4U CN202022198405U CN212569821U CN 212569821 U CN212569821 U CN 212569821U CN 202022198405 U CN202022198405 U CN 202022198405U CN 212569821 U CN212569821 U CN 212569821U
- Authority
- CN
- China
- Prior art keywords
- finger
- fingerprint
- image
- acquisition device
- image acquisition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Image Input (AREA)
Abstract
The present disclosure provides a non-contact fingerprint acquisition device. In some embodiments, a contactless fingerprint acquisition device includes: a housing including a first component and a second component connected to each other in an L-shape; the image acquisition device is arranged on the shell and used for shooting a fingerprint image, and a shooting area for image acquisition is positioned on one side of the first composition part, which is close to the second composition part; and the processing device is arranged in the shell and used for identifying and processing the image shot by the image acquisition device. The implementation of the present disclosure can improve the definition of a photographed fingerprint image.
Description
Technical Field
The disclosure relates to the technical field of non-contact fingerprint acquisition, in particular to a non-contact fingerprint acquisition device.
Background
With the development of artificial intelligence, identity authentication techniques relying on biometrics have been widely used in recent years, and related application scenarios include face recognition, voiceprint recognition, and the like.
The identity authentication technology based on the human face is developed quickly, namely 1: n face comparison and 1: 1, identity verification, and a plurality of landing scenes, such as identity card authentication and verification, gate machine passing, offline payment and the like.
Although the progress of artificial intelligence has led to great progress and rapid popularization of face recognition technology in recent years, face recognition technology is widely disputed among society and masses due to the special personal privacy of faces and the age and ethnic bias problems involved behind the faces, and face recognition technology is not accurate enough when the library capacity is large.
Fingerprints, as a biometric identification technology occupying over 50% of market share in biometric identification, have wide and deep applications in the fields of criminal investigation, entry and exit, personal consumer electronics, security, financial banking and the like. The traditional optical or capacitive touch fingerprint acquisition equipment has the problems of low acquisition quality, small acquisition area, sensitivity to skin dryness and humidity, low acquisition consistency and the like besides the sanitation risk brought by touch acquisition.
Various public health events occur, so that the contact type fingerprint acquisition equipment has small health risks; face recognition cannot be performed with the mask. And the individual user has a lot of inconvenience when carrying out the biological characteristic identification.
Present fingerprint collection equipment, for example the punched-card machine, the fingerprint of binaryzation is gathered to the mode that adopts contact fingerprint collection, and the speed of fingerprint is gathered to this kind of mode is slow, and the definition of fingerprint image receives dynamics when pressing the fingerprint easily to gather remaining fingerprint on the back equipment and influence subsequent collection and reveal fingerprint information easily. In recent years, in order to solve the inconvenience caused by the contact type fingerprint collection mode, a non-contact type fingerprint collection mode is proposed, and when a fingerprint is adopted by the non-contact type fingerprint collection mode, the collected fingerprint image needs to be ensured to have enough definition.
SUMMERY OF THE UTILITY MODEL
The present disclosure has been made to solve the above problems, and provides a non-contact fingerprint acquisition device.
The present disclosure adopts the following technical solutions.
In some embodiments, the present disclosure provides a contactless fingerprint acquisition device comprising:
a housing including a first component and a second component connected to each other in an L-shape;
the image acquisition device is arranged on the shell and used for shooting a fingerprint image, and a shooting area for image acquisition is positioned on one side of the first composition part, which is close to the second composition part;
and the processing device is arranged in the shell and used for identifying and processing the image shot by the image acquisition device.
The embodiment of the disclosure provides a non-contact fingerprint acquisition device, wherein a shell comprises a first component and a second component which are mutually connected into an L shape; the shooting area of image acquisition is located one side that first component part is close to the second component part, and first component part and second component part shelter from the veiling glare that has the interference to fingerprint collection jointly to be favorable to improving the definition of the fingerprint image of shooting.
Drawings
FIG. 1 is an exemplary system architecture diagram in which the present disclosure may be applied;
FIG. 2 is a flow chart of a method of non-contact fingerprint identification according to one embodiment of the present disclosure;
FIG. 3 is a schematic view of a finger recognition box of one embodiment of the present disclosure;
FIG. 4 is a flow chart of a method of non-contact fingerprint identification according to another embodiment of the present disclosure;
FIG. 5 is a flow chart of a fingerprint acquisition method of one embodiment of the present disclosure;
FIG. 6 is a flow chart of the normalization step of one embodiment of the present disclosure;
FIG. 7 is a flowchart of a fingerprint comparison method according to an embodiment of the present disclosure;
FIG. 8 is a flowchart of a fingerprint area detection method of one embodiment of the present disclosure;
FIG. 9 is a flow chart of a fingerprint identification method of one embodiment of the present disclosure;
FIG. 10 is a schematic illustration of a contour line of an embodiment of the present disclosure;
FIG. 11 is a schematic diagram of a non-contact fingerprint identification device according to an embodiment of the present disclosure;
FIG. 12 is a schematic diagram of a fingerprint matching apparatus according to an embodiment of the present disclosure;
fig. 13 is a schematic structural diagram of a fingerprint area detection device according to an embodiment of the present disclosure;
FIG. 14 is a schematic diagram of a fingerprint recognition device according to an embodiment of the present disclosure;
FIG. 15 is a schematic view of a non-contact fingerprint acquisition device according to one embodiment of the present disclosure;
FIG. 16 is a schematic view of another non-contact fingerprint acquisition device according to one embodiment of the present disclosure;
FIG. 17 is a schematic view of another non-contact fingerprint acquisition device according to one embodiment of the present disclosure;
FIG. 18 is a schematic view of a non-contact fingerprint acquisition device having an optical path adjustment device according to one embodiment of the present disclosure;
fig. 19 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.
The labels in the figure are: 1.a housing; 11. a first component; 12. a second component; 13. a third component; 2. an image acquisition device; 3.a processing device; 4. an illumination device; 41. an illumination section; 5. an optical path adjusting device; 6. a structured light projection device; 7. time of Flight (Time of Flight) devices.
The above and other features, advantages and aspects of various embodiments of the present disclosure will become more apparent by referring to the following detailed description when taken in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numbers refer to the same or similar elements. It should be understood that the drawings are schematic and that elements and elements are not necessarily drawn to scale.
Detailed Description
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs; the terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure; the terms "including" and "having," and any variations thereof, in the description and claims of this disclosure and the description of the above figures are intended to cover non-exclusive inclusions. The terms "first," "second," and the like in the description and claims of the present disclosure or in the above-described drawings are used for distinguishing between different objects and not for describing a particular order.
Reference herein to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the disclosure. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it is to be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the disclosure are for illustration purposes only and are not intended to limit the scope of the disclosure.
It should be understood that various steps recited in method embodiments of the present disclosure may be performed in parallel and/or in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the present disclosure is not limited in this respect.
The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions for other terms will be given in the following description.
It should be noted that the terms "first", "second", and the like in the present disclosure are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence relationship of the functions performed by the devices, modules or units.
It is noted that references to "a" or "an" in this disclosure are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that reference to "one or more" unless the context clearly dictates otherwise.
The names of messages or information exchanged between devices in the embodiments of the present disclosure are for illustrative purposes only, and are not intended to limit the scope of the messages or information.
In order to make the technical solutions of the present disclosure better understood by those skilled in the art, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.
[ System Structure ]
First, the structure of the system of one embodiment of the present disclosure is explained. As shown in fig. 1, the system architecture 100 may include terminal devices 101, 102, 103, 104, a network 105, and a server 106. The network 105 serves as a medium for providing communication links between the terminal devices 101, 102, 103, 104 and the server 106.
In this embodiment, an electronic device (e.g., terminal device 101, 102, 103, or 104 shown in fig. 1) on which the method of the present disclosure operates may perform transmission of various information through the network 105. Network 105 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few. It is noted that the wireless connection means may include, but is not limited to, a 3G/4G/5G connection, a Wi-Fi connection, a bluetooth connection, a WiMAX connection, a Zigbee connection, a UWB connection, a local area network ("LAN"), a wide area network ("WAN"), an internet network (e.g., the internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as other now known or later developed network connection means. The network 105 may communicate using any currently known or future developed network Protocol, such as HTTP (Hyper Text Transfer Protocol), and may interconnect any form or medium of digital data communication (e.g., a communications network).
A user may use terminal devices 101, 102, 103, 104 to interact with a server 106 via a network 105 to receive or send messages or the like. Various client applications, such as a video live and play application, a web browser application, a shopping application, a search application, an instant messaging tool, a mailbox client, social platform software, etc., may be installed on the terminal device 101, 102, 103, or 104.
The terminal device 101, 102, 103, or 104 may be various electronic devices having a touch display screen and/or supporting web browsing, including, but not limited to, a smart phone, a tablet computer, an e-book reader, an MP3 (moving picture experts group compression standard audio layer 3) player, an MP4 (moving picture experts group compression standard audio layer 4) player, a head mounted display device, a notebook computer, a digital broadcast receiver, a PDA (personal digital assistant), a PMP (portable multimedia player), a vehicle terminal (e.g., a car navigation terminal), and the like, and a mobile terminal such as a digital TV, a desktop computer, and the like. In addition, the terminal device 101, 102, 103 or 104 may also be a dedicated device dedicated to performing the inventive method.
The server 106 may be a server that provides various services, such as a background server that provides support for pages displayed or data transferred on the terminal device 101, 102, 103, or 104.
It should be understood that the number of terminal devices, networks, and servers in fig. 1 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.
Here, the terminal device may implement the embodiment method of the present disclosure independently or by running applications in various operating systems, such as an android system, in cooperation with other electronic terminal devices, or may run applications in other operating systems, such as applications in an iOS system, a Windows system, a hong meng system, and the like, to implement the embodiment method of the present disclosure.
[ non-contact fingerprint recognition method ]
As shown in fig. 2, fig. 2 is a flowchart of a non-contact fingerprint identification method according to an embodiment of the disclosure, which includes the following steps.
S100, a finger identification step, namely shooting the hand of the identification object, judging whether a finger exists in the shot picture, and determining the position of the finger when the finger exists.
Specifically, the embodiment of the present disclosure may include a step of obtaining a finger recognition model; judging whether the picture comprises the finger or not according to the finger identification model; and determining the position of the distal phalanx segment of each finger when the picture comprises the finger; wherein the orientation of the finger includes an arbitrary direction parallel to the screen. Wherein the step of obtaining a finger recognition model may comprise the step of obtaining a sample set; wherein the sample set comprises data-augmented finger images acquired under different conditions including different angles, different distances, different illumination, different sharpness, and different acquisition devices; and training according to the finger image to obtain the finger recognition model.
In addition, the method can further include the steps of scanning the picture by using a plurality of identification frames with different areas, and in the scanning process of each identification frame, when the confidence level of the finger contained in the area defined by the identification frame is greater than a preset value, keeping the identification frame in the area as a mark and continuing to scan until the picture is scanned, so that each identification frame as the mark at least comprises a part of the finger; sorting the recognition frames according to the credibility to obtain a sorting result; acquiring a target identification frame from the identification frame according to the sorting result; tracking the finger within a preset range including the target recognition frame; wherein each target recognition frame comprises a minimum complete area of one finger. More specifically, the step of tracking the finger within a preset range including the target recognition box may include detecting, by the finger recognition model, the confidence level that includes the finger within the preset range; when the reliability is greater than the preset value, moving the target recognition frame to a position in the preset range, where the reliability is greater than the preset value; and adjusting the target recognition frame so that the minimum complete area of the target recognition frame containing one finger is reduced from an outer dotted frame to an inner dotted frame in fig. 3.
For example, embodiments of the present disclosure may provide a calling function based on the SDK form. The upper layer application may call this function to bring up an acquisition page. At this time, whether a finger is present or not can be captured from each frame. When the human hand is placed at a set position and is stable, the current image can be obtained, and data containing the finger image is returned through a series of models and calculations. Based on the deep learning training model, reasoning calculation can be carried out on each frame of video stream in real time to detect whether fingers and finger positions exist in the image, and for each frame of image, whether a finger frame can be identified is judged through the model and calculation when the image enters a preview picture. If no finger box is recognized for a continuous period of time, the user may be interactively prompted to find no finger in the screen; if the finger is not recognized for a longer period of time, a timeout may be made. Including but not limited to identifying fingers that distinguish between different finger positions; identifying that the used frame contains the finger head area from the end point to the first knuckle line and may not contain the rest finger area; the hand can rotate freely in the picture without influencing correct recognition; according to the difference of the fingers, the left hand and the right hand are automatically identified. The finger components supporting identification can comprise two thumbs, one five fingers, one four fingers, one finger and the like. Finger recognition may include finger detection and finger tracking, among others. For example, a preview image is obtained through a camera, finger detection model inference calculation is performed on a first frame image, and if the image is divided into a plurality of frames with combined length and width, each frame has a value: confidence level of whether the finger is included; the method comprises the steps of obtaining a plurality of finger frames through processing, sequencing all the finger frames corresponding to the figure through credibility, merging the finger frames through business logic (such as a 300x300 finger frame and a 310x310 finger frame, wherein the two finger frames contain the same finger), adjusting the finger to only correspond to one finger frame, and returning the identified finger frame. Starting from the second frame image, if the finger frame is detected in the previous frame image, finger tracking can be performed. For example, model matching of the finger is performed near the target position using the previous finger position, then accurate finger frame positioning is performed according to the detection model, whether the finger is a finger is determined by the reliability, and the recognized finger frame is returned. The finger tracking can further improve the identification accuracy and reduce the time required by single-frame image identification. In order to adapt to different acquisition devices, such as different cameras, etc., the embodiments of the present disclosure may further perform generalization processing on the model, such as acquiring data samples of different scenes, such as different angles, different distances, different illuminations, and whether the focusing is clear, acquiring data of different devices, such as different cameras, and performing data augmentation, such as translation, rotation, noise addition, smoothing processing, etc. After the model is obtained based on the data training, stable recognition can be achieved under the conditions of poor image definition (for example, incomplete out-of-focus state), equipment with different resolutions and within a certain distance (for example, 10-40cm) between the hand and the camera.
S200, a fingerprint acquisition step, wherein under the condition that the finger is detected to exist in the finger identification step, focusing and photographing are carried out on the finger, and respective fingerprint images of the finger are acquired.
Specifically, the method may include the steps of acquiring a continuous multi-frame image including the finger, detecting a position of the finger in the continuous multi-frame image, and determining whether a variation range of the position is within a preset threshold range; if yes, detecting the respective focusing states of the fingers; if the fingers are focused at the same time, shooting with a corresponding focal length to obtain an image; and if the fingers are not focused at the same time (for example, the fingers are not on the same focal plane), respectively shooting and obtaining a plurality of images corresponding to the fingers according to the respective focal lengths of the fingers.
In the preview, the position of the finger may be displayed in the frame, as shown in fig. 3. In order to ensure the quality of the acquired image, the finger needs to be stably detected from the preview screen, and the finger needs to be in a proper position. For example, if the finger is too close to the camera, the problems of fuzzy focusing, incomplete finger shooting and the like may occur, and at this time, whether the distance of the hand is too close or not may be determined according to the resolution of the camera and the ratio of the size of the finger in the figure, and when the distance is too close, the user is prompted to move the hand farther. In contrast, if the finger is too far away from the camera, problems such as focusing on the background, unclear finger lines due to resolution limitation, and the like may occur, and at this time, whether the distance of the hand is too far may be determined according to the resolution of the camera and the ratio of the size of the finger in the figure, and when the distance is too far, the user is prompted to bring the hand closer. Furthermore, the sharpness of the fingers is relatively higher when the fingers are closed. When the fingers are separated, the fingers far away are easy to shoot blurrily due to the problem of dislocation of the focusing plane, at the moment, the separation distance of the fingers can be judged according to the position of the finger frame, whether the fingers are closed or not is judged through logic operation, and when the fingers are detected not to be closed, a user can be prompted to close the fingers. When the finger is in a proper position, the current image can be acquired, for example, a camera photographing interface can be called to photograph, or one frame of image of a preview video stream can be directly adopted, and both image acquisition forms can support an active and automatic mode. Wherein, the actively acquiring may be that the operator manually clicks a button to acquire a related image; automatically acquiring whether the position change of the finger is within a certain threshold range or not according to continuous multi-frame pictures, and judging whether the hand is relatively stable or not; if the acquisition equipment is movable equipment (such as a mobile phone) and the acceleration gyroscope is arranged, the information of the gyroscope can be read, and whether the equipment is relatively stable or not is judged. When the judgment is stable, the image acquisition can be automatically triggered. In particular, when there are a plurality of fingers on the screen, the fingers may not be in the same focal plane, and thus, some fingers may be clear and some fingers may be blurred. To achieve high quality images, embodiments of the present disclosure may perform multiple focus shots. For example, according to the finger frame identified by the model, the positions of a plurality of fingers can be obtained; respectively setting focusing points according to the positions of the fingers, and triggering focusing and photographing; and a sharp image of each finger is taken out of each photographed image, respectively.
S300, an image transmission step, namely encrypting the fingerprint images of the fingers acquired in the fingerprint acquisition step and transmitting the encrypted fingerprint images to a comparison system.
Specifically, the disclosed embodiments may include the step of assigning a key to the fingerprint image; the key is used for verifying the key when receiving the access request of the fingerprint image, passing the access request when the key verification is passed, and rejecting the access request when the key verification is failed. More specifically, before transmitting the fingerprint image subjected to the encryption processing to a comparison system, the disclosed embodiment may further include a step of selecting a transmission interface; the transmission interface comprises a fingerprint image warehousing interface, a single-finger comparison interface and a multi-finger comparison interface.
The embodiment of the disclosure can encrypt the fingerprint image. For example, before calling the collection interface, a one-time key may be obtained from the comparison system background, and the key may have a certain validity period; the key may be passed in when the acquisition interface is invoked. The key can be verified during shooting, and if the verification is passed, shooting can be started. After shooting is finished, the collected data can be encrypted by the key. More specifically, a secret key provided to an integrating party can be used for generating a one-time token (OTP), and after the acquisition is completed, encrypted fingerprint information is generated by using the OTP, so that the transmission is safe and the falsification is prevented. The embodiment of the disclosure can separately store sensitive picture data and insensitive feature file data by encryption transmission and storage of full-flow data, control access of the picture data, and perform hash processing on the feature file data, so as to ensure that the features after hash can still be compared with high precision, but any original information cannot be recovered from the hashed feature file, thereby ensuring the security of the fingerprint image data.
S400, a fingerprint comparison step, namely comparing the fingerprint image with the fingerprint data in the comparison system to obtain a comparison result.
Specifically, the disclosed embodiments may include steps of fingerprint authentication comparison and fingerprint search comparison. For example, the comparison capability of the fingerprint can be provided through an open API or the like, which may include three interfaces of fingerprint image warehousing, fingerprint image one-to-one comparison and fingerprint image one-to-many comparison. Wherein, fingerprint image one-to-one comparison can be for carrying out 1: 1, comparing to complete identity verification; the one-to-many comparison of the fingerprint images can be that 1: and N is compared, whether the fingerprint exists or not is checked, and the like. In addition, the comparison system can be arranged in a local or cloud server.
Furthermore, the present disclosure may further include an image processing step of processing the fingerprint image acquired in the fingerprint acquisition step, the image processing step further including: acquiring two end points of a distal interphalangeal joint line of each finger in the fingers according to the knuckle line model, and adjusting the knuckle line to a preset state according to the two end points; acquiring a plurality of boundary points of each finger edge according to a finger contour model; smoothly connecting the two end points of each finger and the plurality of boundary points in sequence to form a contour line of each finger; and a step of extracting the contour line from each of the plurality of images to obtain the fingerprint image.
In addition, the embodiment of the present disclosure may further include a step of acquiring a distance between the two end points of the distal interphalangeal joint line of each finger and an area of the contour line of each finger, and adjusting a proportion of the fingerprint image according to the distance or the area to obtain a fingerprint image of a preset specification; or acquiring the object distance of the finger, and unifying the object distance to acquire a fingerprint image with a preset specification. Acquiring the object distance of the finger may include acquiring a preset ratio of the focal distance to the object distance, and acquiring the object distance according to the preset ratio and the focal distance; or acquiring the depth information of the image, and acquiring the object distance according to the depth information.
In addition, the embodiment of the present disclosure may further include a step of partitioning the fingerprint image; respectively acquiring the fingerprint direction and the ridge density of each area of the fingerprint image; and performing image transformation processing on each interval to enable the difference value of the density of the lines in each area to be smaller than a preset difference value.
In addition, the embodiment of the present disclosure may further include a step of connecting the contour lines to form a convex function; and classifying the fingers into fingers of a left hand or a right hand according to the convex function and the length of each finger, and respectively corresponding the contour lines to a thumb, an index finger, a middle finger, a ring finger and a little finger.
In addition, the embodiment of the disclosure may further include performing equalization processing on the fingerprint image; performing signal filtering on the fingerprint image according to the frequency domain signal corresponding to the lines of the fingerprint; and calculating the frequency domain signal to obtain the quality score of the fingerprint image.
The images obtained from the camera can be subjected to finger recognition and image segmentation through the model, and finally the image of each finger is obtained. Specifically, one or more finger frames, i.e., positions of the fingers, may be obtained using the finger recognition model. Acquiring the positions of two end points of a first knuckle line of the finger by using a knuckle line model, which is shown as A and B in FIG. 3; and then, acquiring a plurality of end points of the edge of the finger by using the finger contour line model, and connecting to form the contour line of the finger. Obtaining a plurality of finger frames according to the finger recognition model, and generating one or more finger images; performing model reasoning on a finger knuckle line and a contour line of each finger image respectively to obtain end points on two sides of the finger knuckle line and a plurality of points forming the finger contour line; through processing, the direction of the finger is adjusted and the background of the finger is removed, for example, the direction is adjusted through end points on two sides of a knuckle line, the finger tip is ensured to be upward in the final output image by adjusting the knuckle line to be in a horizontal state, the hand can rotate during shooting, and then the direction is adjusted through the knuckle line, so that the direction of the output finger is ensured to be consistent; the finger area is obtained through the finger contour line, and the background area outside the finger is removed. The finger frames are connected to form a convex function, and whether the photographed hand is the left hand or the right hand can be calculated and judged according to the relation between the function and the fingers (such as the farthest thumb and the shortest little finger), and the thumb, the index finger, the middle finger, the ring finger and the little finger corresponding to each finger frame are obtained. If the left or right hand cannot be determined, the user may specify the left or right hand. The finger images can be scaled and output collectively as, for example, a 500dpi image. For example, the average value of the fingerprint areas and the finger pitch line lengths corresponding to different finger positions is obtained by counting a certain number of fingerprint images, and based on the empirical value and the finger area and the finger pitch line length in each finger frame, the image proportion of each finger image can be adjusted; calibrating and analyzing the plurality of types of equipment to obtain the relationship between the focal length read out by the equipment at the equipment end and the actual object distance, reading the focal length from a photographing interface of the equipment during photographing, and converting the actual object distance of the finger to adjust the proportion of the finger image; the depth information of the image can be acquired by means of double cameras, structured light, ToF and the like, so that the actual object distance from the finger to the equipment is estimated, and the proportion of the finger image is adjusted. Furthermore, the finger image is captured by a camera, because the finger is a three-dimensional object, and the captured image is two-dimensional, which may cause the density of the lines in different areas of the fingerprint to be inconsistent: the lines in the middle of the fingerprint are sparse, and the lines around the fingerprint are dense. In order to perform a better comparison, the finger image may be normalized, for example, by calculating the fingerprint directions and frequencies of different regions to obtain an approximate ridge density, and by performing image transformation processing, the ridge density of each region of the image reaches or approaches a fixed value. In addition, the depth information can help people to unfold the fingerprint edge, so that the fingerprint center area and the fingerprint edge area have the same line density, and higher precision can be realized when the fingerprint acquired by the contact type fingerprint acquisition instrument is compared in the later period. In order to ensure the accuracy of subsequent fingerprint comparison, the finger image needs to be relatively clear, so quality judgment can be added in an image processing mode, including equalization processing on the image, signal filtering on the image according to frequency domain signals corresponding to fingerprint lines, calculation on line signals in the image, judgment on the quality of the whole image and the like. After the image processing flow is completed, for example, a plurality of finger gray-scale images of 640 × 640500 dpi can be obtained.
In order to implement the technical solution of the non-contact fingerprint identification method in the embodiments of the present disclosure, an embodiment of the present disclosure provides a non-contact fingerprint identification device, which may be specifically applied to various electronic terminal devices, as shown in fig. 11, and further provides a non-contact fingerprint identification device, including an identification module 181, which may be used in a finger identification step, to shoot a hand of an identification object, determine whether a finger exists in a shot picture, and determine a position of the finger when the finger exists; an obtaining module 182, configured to perform a fingerprint obtaining step, perform focusing and photographing on the finger when the finger is detected in the finger identifying step, and obtain a fingerprint image of each finger; a transmission module 183, which is used in the image transmission step, for encrypting the fingerprint images of the fingers acquired in the fingerprint acquisition step, and transmitting the encrypted fingerprint images to the comparison system; and a comparison module 184 for comparing the fingerprint image with the fingerprint data in the comparison system to obtain a comparison result. In addition, the embodiment of the present disclosure may further include a processing module (not shown) configured to perform image processing on the fingerprint image acquired in the fingerprint acquiring step.
[ Living body identification method ]
The above has generally described the non-contact fingerprint identification method of the present invention. In addition, in the application of fingerprint identification, some people steal the fingerprint of a user and imitate the finger with rubber and the like so as to deceive fingerprint identification equipment, thereby causing potential safety hazards. Therefore, in the non-contact fingerprint recognition method of the present invention, it is preferable that the living body recognition is performed before the final fingerprint recognition is performed, and the fingerprint recognition is performed only when the recognition target is a living body. Hereinafter, the living body identification method in the non-contact fingerprint identification method of the present invention will be specifically described.
As shown in fig. 4, the method of the present disclosure includes the following steps.
(1) An acquisition step of acquiring a fingerprint image of at least a part of a fingerprint of at least one finger of an identified object by a non-contact shooting manner.
Specifically, in the prior art, when acquiring a fingerprint image of a finger, the finger is in contact with a fingerprint acquisition device (e.g., a fingerprint punched-card machine), and the fingerprint image is acquired in a contact manner, the existing contact-type fingerprint image is a binarized fingerprint image, in this embodiment, a non-contact shooting manner is adopted to acquire the fingerprint image, the finger is not in contact with a device for acquiring the fingerprint image, and the acquired fingerprint image may be a colored fingerprint image.
(2) And a living body identification step of judging whether the identified object is a living body in at least one mode to obtain at least one judgment result and obtaining an identification result according to the at least one judgment result.
Specifically, in this embodiment, at least one manner is adopted to determine whether the identified object is a living body, the identified object is an owner of the fingerprint image of at least a part of the fingerprint of the at least one finger, and one manner corresponds to one determination result. In this embodiment, the identification result is a final result of whether the identified object obtained in the living body identification step is a living body, that is, the determination result is used as a basis for determining the identification result, when there is only one determination result, the identification result and the determination result may be the same, when there are a plurality of determination results, the reliability of each determination result needs to be considered comprehensively to obtain the identification result, a weight may be set for each determination result in advance, and the identification result may be determined according to the weight result.
(3) A fingerprint recognition step of performing fingerprint recognition on the fingerprint image in a case where the recognition result shows that the object to be recognized is a living body.
Specifically, in the present embodiment, whether or not to perform the fingerprint recognition result is determined based on the recognition result, and in the case where the fingerprint recognition result shows that the object to be recognized is a living body, the case where the finger is a fake non-living finger is excluded, and the fingerprint recognition is performed on the image at this time, thereby improving the security and reducing unnecessary calculations.
Here, the content of acquiring the fingerprint image is specifically described in the section of the non-contact fingerprint identification method, and only the difference point is taken as the center for description, and the rest of the content can be referred to the description in the section.
In some embodiments of the disclosure, the obtaining step comprises: acquiring a fingertip image of at least one finger of an identified object, wherein the living body identification step comprises the following steps: and inputting the fingertip image into the trained neural network model, and obtaining a first result representing whether the identified object is a living body according to the output of the neural network model.
Specifically, in some embodiments, when the obtaining step is performed, a fingerprint image of a non-contact fingerprint is captured, one or more fingertip positions are located through a finger tracking and detection algorithm, the fingertip positions may include a fingerprint region, the collected fingertip image is input into a trained neural network model for processing, and an output result of the neural network model is used to characterize whether the input fingertip image is a living body result, that is, a first result is obtained.
In some embodiments of the present disclosure, when the neural network model is established, a non-contact fingerprint image of a living body, and a 2D or 3D fingerprint image of a non-living body are collected in advance, and x _ i may be set to be used to characterize the ith fingerprint image, y _ i may be used to characterize whether the fingerprint image characterized by x _ i is a fingerprint image of a living body, y _ i is 1 when the fingerprint image is a fingerprint image of a living body, i is a variable for identifying an image, and y _ i is 0 when the fingerprint image is not a fingerprint image of a living body. Training is needed for the established neural network model, specifically, training a classifier F of the neural network including the convolutional layer, the pooling layer and the fully-connected layer, and optimizing F by using a stochastic gradient descent or other optimization methods, so that F (x _ i) is associated with y _ i, for example, the classifier can be trained by using the following formula: max _ { F }1/N _ \ sum _ i y _ i log F (x _ i) + (1-y _ i) (1-log F (x _ i)).
After the classifier F completes training, F (x _ i) represents the probability that x _ i is a non-contact fingerprint image of a living body, if F (x _ i) >0.5 represents the probability that x _ i is a non-contact fingerprint image of a living body, and if F (x _ i) <0.5 represents that x _ i is not a non-contact fingerprint image of a living body. In some embodiments, a threshold value of 0.5 is adopted as a default to be compared with F (x _ i) to determine whether the fingerprint image is a fingerprint image of a living body, and the false alarm rate and the false negative rate of the system can also be adjusted by adjusting the threshold value of the determination to adapt to the needs of a scene.
In some embodiments of the present disclosure, a non-live contactless fingerprint image is generated by at least one of a data augmentation and countermeasure generation network (GAN), and the neural network model is trained using the generated non-live contactless fingerprint image. Specifically, in reality, various non-contact fingerprint images of different non-living bodies are not easy to obtain in batch, so that more non-contact fingerprint images of the non-living bodies are generated by adopting a data augmentation and countermeasure generation network, overfitting of a neural network is reduced, and the accuracy of judging whether the fingerprint images are fingerprint images of the living bodies in practical application is improved.
In some embodiments of the present disclosure, the living body identifying step includes: the method comprises the following steps: obtaining a rPPG (Remote photoplethysmography) signal of at least one finger of the identified object, and obtaining a second result representing whether the identified object is a living body according to the rPPG signal; specifically, in this embodiment, the at least one judgment result includes the second result. The rPPG signal is a slight brightness change of the skin measured by reflected ambient light, the slight brightness change of the skin is caused by blood flow caused by heart beating, and for a fingerprint image of a living body, part of the ambient light penetrates through a cortex layer to reach a blood vessel and is reflected back to a camera, so that the brightness change caused by the blood flow can be monitored in the fingerprint image of the living body, and for a fingerprint image of a non-living body, the brightness change caused by the blood flow cannot be monitored. rPPG signals can be acquired by taking video of an identified subject.
In some embodiments, deriving from the rPPG signal a second result characterizing whether the identified subject is a living subject comprises: determining a change characteristic of an optical property of the identified subject from the rPPG signal, determining a second result from the change characteristic of the optical property. In this embodiment, the second result may be determined from different areas of the image of the finger of the recognized object, or may be determined from the image of the same area of the finger of the recognized object. For a fingerprint image of a living body, the blood flow conditions are different in different areas, so that the rPPG signals of different areas of a finger in one image have differences, while the rPPG signals of different areas of a non-living body prepared by rubber and the like have smaller differences, so that a second result can be obtained according to the rPPG signals of different areas, and on the other hand, the blood flow conditions of the same area of the living body at different times are different, but the states of the same area of the non-living body at different times are basically consistent, so that the second result can be determined according to the rPPG signals of the same area at different times. Optionally, the rPPG signal may be rPPG signals of a plurality of fingers, specifically, when acquiring the rPPG signal, the finger of the identified object may be tracked, at least two temporally different fingerprint images of one or more fingers are located and captured, for example, the fingerprint images may be acquired by using a video of the finger of the identified object, an RGB component mean value of at least one finger in each frame image is calculated, that is, each frame image obtains 3 feature values as continuous rPPG signal outputs, in the process of obtaining a second result according to the rPPG signal, the rPPG signal may be filtered first, then the filtered signal is converted into a frequency domain signal through fourier transform, a frequency domain signal is analyzed to obtain a feature, the analyzed feature is classified to determine whether the feature is a living feature or a non-living feature, so as to obtain the second result, the method for classifying the features obtained by analysis may be to classify the features by using a neural network model (e.g., a ConvLSTM model) or a support vector machine model. In this embodiment, it is considered that a living body has a heartbeat relative to a non-living body, and the heartbeat affects the optical properties of a fingerprint image, thereby realizing determination of whether an identified object is a living body.
In some embodiments of the disclosure, the obtaining step further comprises: images of a non-contact fingerprint of at least one finger of an identified object under different lighting conditions are acquired. The living body identifying step further includes: and obtaining a third result representing whether the identified object is a living body according to the images under different illumination conditions. Specifically, in this embodiment, the at least one judgment result further includes: and (5) a third result. When the lighting condition changes, the taken fingerprint image changes, wherein the change situation of the fingerprint image of the living body is different from that of the fingerprint image of the non-living body, so that the identified object corresponding to the fingerprint image can be judged to be the living body or the non-living body by changing the lighting condition and taking the fingerprint images under different lighting conditions.
In some embodiments of the present disclosure, the lighting conditions include at least one of brightness, light source color, and light source polarization state; images under different lighting conditions include: the images acquired under different brightness, the images acquired under different light source colors or the images acquired under different light source polarization states. Some non-living body recognized objects may be finger images on printed paper boards or finger images displayed through electronic screens, so that images under the condition that a flash lamp is turned on and the flash lamp is not turned on can be respectively shot to obtain images under different brightness, and because the paper boards and the electronic screens have the characteristic of approximate plane reflection, the living body recognition accuracy can be improved by changing the brightness. Since the living body and the non-living body have different physical properties and thus display states differ among different light sources, the accuracy of living body identification can be improved by changing the color of the light source. The light source may emit polarized light and unpolarized light, and determine whether it is a living body according to a difference in reflection of the object to be recognized under the polarized light.
Specifically, in this embodiment, the brightness when the fingerprint image is captured may be changed, the color of the light projected onto the finger of the identified object may be changed, and the polarization state of the light projected onto the identified object may be changed, it should be noted that the captured fingerprint images at different brightness, the captured fingerprint images at different light source colors, and the fingerprint images at different light source polarization states may be respectively obtained, whether the identified object is a living body may be respectively determined based on the three different illumination conditions, three sub-determination results may be obtained, and the third result may be obtained according to the three sub-determination results, so that the sub-determination results under various illumination conditions may be integrated to improve the accuracy of the determination.
In some embodiments of the present disclosure, the reflectivity of the images obtained under different lighting conditions is calculated, and a third result characterizing whether the identified object is a living body is obtained according to the reflectivity. Specifically, the human skin has a specific optical characteristic that absorbs light of a specific wavelength range and generates light of a specific wavelength range, and light reflected by non-human skin and light reflected by human skin are not exactly the same, and thus, it is possible to determine whether or not an object is a living body by comparing the reflectance. Because the reflectivity of part of materials and human bodies in certain wavelength ranges may partially coincide but not coincide in all wavelength ranges, optionally, images shot without light source colors can be respectively obtained, the reflectivity of the identified object in different light source colors is calculated, so as to determine whether the identified object is a living body according to the reflectivity in different light source colors, for example, the images in different light source colors can be video images under illumination with continuously-changed wavelength in a section of wavelength range, so as to obtain the continuous change of the reflectivity along with the wavelength, and at the moment, a curve of the reflectivity along with the change of the wavelength of the illumination is obtained.
In some embodiments of the present disclosure, the images obtained under different lighting conditions are input into a deep neural network, and a third result representing whether the identified object is a living body is obtained through the deep neural network. Specifically, for the obtained images under different lighting conditions, a neural network model can be used for identification, wherein, taking the lighting condition as brightness as an example, images of a live finger taken under the conditions of turning off a flash and turning on the flash and images of a non-live finger taken under the conditions of turning off the flash and turning on the flash are acquired in advance, wherein the image taken under the condition of turning off the flash is X _ i ^1, the image taken under the condition of turning on the flash is X _ i ^2, i is a variable for distinguishing different subjects, a photograph of one subject can be recorded as (X _ i ^1, X _ i ^2), and (X _ i ^1, X _ i ^2) has a corresponding y _ i, when the subject is a live, y _ i ^1, otherwise y _ i is 0, the images are acquired in advance with (X _ i ^1, x _ i ^2) and the corresponding y _ i train the deep neural network F comprising the convolutional layer, the pooling layer and the full-link layer, so that F (X _ i ^1 ', X _ i ^ 2') to y _ i), thus an image shot by the identified object under the conditions of turning off a flash lamp and turning on the flash lamp is input into the deep neural network F, the deep neural network F outputs y _ i corresponding to the identified object, the probability that y _ i corresponding to the identified object is a living body can be set, the identified object is determined to be a living body when y _ i is larger than 0.5, otherwise, the identified object is a non-living body.
In some optional embodiments, before inputting the fingerprint images obtained under different lighting conditions into the deep neural network, the method further includes: fourier transformation is carried out on the images obtained under different illumination conditions to obtain frequency domain signals, and the frequency domain signals are also input into the deep neural network, so that the accuracy of the model living body identification is improved. The inventor of the present disclosure finds that frequency domain signals of images of a living body and a non-living body taken under different lighting conditions have very high discrimination, and the discrimination between the living body and the non-living body can be significantly improved by inputting the frequency domain signals into a deep neural network.
In some embodiments of the present disclosure, fingerprint images obtained under different lighting conditions are input into the support vector machine model, and a third result representing whether the identified object is a living body is obtained through the support vector machine model. In some optional embodiments, the images obtained under different lighting conditions are input into a deep neural network, comprising: the brightness of each area of the image obtained under different lighting conditions is analyzed, the brightness change value of each area under different lighting conditions is calculated, and the brightness change value is input into a deep neural network or a support vector machine model, so that the accuracy of living body identification can be obviously improved.
In some embodiments of the disclosure, the obtaining step comprises: acquiring an infrared image of the identified object, the living body identifying step further comprising: and obtaining a fourth result representing whether the identified object is a living body or not according to the infrared image. In this embodiment, the at least one determination result further includes: and a fourth result. When the recognized object is photographed, visible light is used for lighting when a visible light picture is taken, an infrared filter of a camera is turned off, an infrared fill-in lamp can be configured and the infrared filter of the camera is turned off when an infrared image is taken, so that the infrared image of the recognized object is taken, or an additional infrared camera can be used for taking the infrared image, the reflectivity of the skin on the surface of a finger of a human body to infrared light is low, the infrared light can partially penetrate through the skin of the finger, which is an important difference between a living body and a non-living body, and the differences are reflected on the taken infrared image, so that whether the recognized object is the living body can be determined according to the infrared image of the recognized object.
In some embodiments of the present disclosure, deriving a fourth result characterizing whether the identified object is a living body from the infrared image includes: and acquiring the optical property of the identified object under the infrared light according to the infrared image, and obtaining the fourth result according to the optical property under the infrared light. Specifically, the optical properties of the living body and the non-living body under the infrared light are different, which is particularly shown in that the skin on the surface of the human finger has low reflectivity to the infrared light, and the infrared light can penetrate through part of the skin of the finger, so that the reflectivity or the transmissivity of the identified object to the infrared light can be determined according to the shot infrared image, and whether the identified object is the living object or not is determined according to the reflectivity or the transmissivity, thereby obtaining a fourth result.
In some embodiments of the present disclosure, deriving a fourth result characterizing whether the identified object is a living body from the infrared image includes: and extracting vein features from the infrared image, and comparing the vein features with the vein features of the living body obtained in advance to obtain the fourth result. Specifically, since infrared light can penetrate the skin on the surface of a finger of a human body, if the identified object is a living body, a finger vein under the skin of the human body can be captured in the infrared image, a vein feature of the finger vein is extracted from the infrared image, the vein feature is compared with a vein feature of the human body obtained in advance, if the coincidence degree of the two is higher than a preset value, the identified object can be determined as the living body, and if the coincidence degree of the two is lower than the preset value, the identified object is determined as a non-living body, so that a fourth result is obtained. Specifically, the infrared image can be identified by using a deep neural network. In some embodiments, a target object corresponding to the image can be determined according to a fingerprint image of a finger of the identified object, a vein feature of the target object is stored in advance as the pre-obtained living vein feature, and at the same time, joint judgment is performed according to the fingerprint and vein feature of the identified object, so that the judgment accuracy is improved.
In some embodiments of the present disclosure, the obtaining step further comprises obtaining an infrared temperature measurement result of the identified object; the living body identifying step further includes: and obtaining a fifth result representing whether the identified object is a living body according to the infrared temperature measurement result. Specifically, in this embodiment, at least one of the determination results further includes: and a fifth result. Specifically, the human body can automatically emit infrared light to the outside, the body temperature of the human body can be detected according to the infrared light emitted by the human body, the temperature of the finger of the human body is within a narrow preset range, the finger made of rubber or the temperature of a printed paper image of the finger is different from the preset range, so that a fifth result can be obtained according to the infrared temperature measurement result, and the influence of the environment on the finger of the human body is considered.
In some embodiments of the present disclosure, before the obtaining step, the method further comprises: a prompting step, prompting the identified object to perform fixed or randomly generated finger action; the living body identifying step further includes: and determining the finger motion performed by the identified object according to the acquired image, and comparing the finger motion performed by the identified object with the finger motion performed by the prompt identified object to obtain a sixth result representing whether the identified object is a living body. Specifically, in this embodiment, the at least one judgment result further includes: and a sixth result. For a forged finger or a finger image printed by paper, the finger cannot move, so that the motion prompted by the prompting step cannot be responded, and therefore whether the identified object is a living body can be determined accordingly. The fingers can be opened, closed, rotated, straightened or slightly bent, etc.
Example 1
In order to better explain the method proposed by the present disclosure, the following description takes an example that the method in one embodiment of the present disclosure is applied to a mobile phone, and the method in the embodiment may be implemented by an application in the mobile phone. In this embodiment, a plurality of fingers of an object to be identified are photographed with fingerprint images, and are judged according to the fingerprint images in a plurality of ways, that is, a plurality of judgment results are obtained, and the plurality of judgment results are integrated to obtain an identification result. Specifically, in this embodiment, a fingerprint image of an identified object is captured, then three determination results are obtained according to a color fingerprint image of the identified object, images of the identified object under different brightness conditions, and an rPPG signal of the identified object, and the three determination results are integrated to obtain an identification result, which will be described in detail below.
The following describes obtaining a determination result from a color fingerprint image of an identified object.
And starting a mobile phone camera, positioning a fingertip area of an object to be identified by the mobile phone camera, shooting a fingerprint image of the fingertip area, inputting the shot fingerprint image into the deep neural network model, and outputting the probability that the fingerprint image is a fingerprint image of a living body by the deep neural network model.
Specifically, in the embodiment, at least a part of the fingerprint image of at least one finger of the object to be identified is a color image, and the traditional contact type fingerprint identification technology is contact type total reflection imaging, so that the acquired image is a binary image, and the additional information is less, so that the fingerprint image is difficult to be used for effective living body identification; since the non-contact fingerprint image is imaged by using an RGB camera, the contained information is greatly enriched, so that the living fingerprint detection by using the deep learning method based on the image becomes possible. Collecting non-contact fingerprint images of a living body and non-contact fingerprint images of a non-living body in batches in advance, wherein the fingerprint images correspond to a parameter y _ i for representing whether the fingerprint images are the fingerprint images of the living body or not, the y _ i of the non-contact fingerprint images of the living body is equal to 1, the y _ i of the non-contact fingerprint images of the non-living body is equal to 0, training a classifier F of a deep neural network comprising a convolutional layer, a pooling layer and a full-link layer by using the collected non-contact fingerprint images of the living body and the non-living body, optimizing the classifier F of the deep neural network by using random gradient descent so that F (x _ i) -y _ i, the input value of the classifier F is the non-contact fingerprint image, the output value of the classifier F is the y _ i of the input non-contact fingerprint image, namely the output value is the probability that the input non-contact fingerprint image is the non-contact fingerprint image of the living body, if y _ i of the input non-contact fingerprint image is greater than 0.5, the input non-contact fingerprint image is considered to be a non-contact fingerprint image of a living body, otherwise, the input non-contact fingerprint image is a non-contact fingerprint image of a non-living body, and a judgment result is obtained according to the fingerprint image of the identified object. In reality, since a large number of non-contact fingerprint images of various non-living bodies are not easy to obtain, more non-contact fingerprint images of the non-living bodies can be generated through a data augmentation and countermeasure generation network, overfitting of a deep neural network is reduced, and accuracy of judgment in practical application is improved.
The following describes obtaining the determination results from images under different luminance conditions.
The method comprises the steps that under the condition that a flash lamp is turned off, a camera of a mobile phone shoots a fingerprint image of an identified object to be identified, then the flash lamp is turned on to shoot the fingerprint image of fingers again, wherein the fingerprint images of a plurality of fingers can be shot simultaneously, the fingerprint images shot under the conditions that the flash lamp is turned on and the flash lamp is turned off are input into a deep neural network model, and a judgment result of whether the identified object corresponding to the fingerprint image is a living body is output.
Specifically, the fingerprint image of the non-living body may be a fingerprint image printed on a paper board or a fingerprint image displayed on an electronic screen, and due to the fact that the paper board picture and the electronic screen picture have the characteristic of plane or near-plane reflection for reflection of light rays, diffuse reflection or mirror reflection generated by the paper board picture and the electronic screen picture for the light rays is obviously different from a finger three-dimensional entity of the living body in vision. Therefore, the living body detection of the non-contact fingerprint image on the mobile phone can be more accurately carried out without using an additional lighting device. In particular, when the recognized object is a living object, images photographed when the flash is turned off and turned on may be significantly different, and a shadow of a finger, i.e., a significant change in shadow, may be displayed on an image photographed after the flash is turned on, compared to an image photographed when the flash is turned off, and when the photographed fingerprint image is a counterfeit fingerprint image printed on a paper board or an image displayed on an electronic screen, the finger displayed thereon may not show a shadow when the flash is turned on, and a reflective area may exist due to reflection of light of the paper. On the other hand, when the object to be recognized is a finger made of a material such as rubber, although the shading can be changed when the flash lamp is turned on, the optical properties such as the reflectance of the finger are different from those of a living body, and therefore, the result of determining whether the object to be recognized is a living body can be obtained from images captured under different luminance conditions.
The following describes the determination from the rPPG signal.
The mobile phone shoots a video stream of the finger of the identified object, the position of the finger is continuously tracked in the video stream, an rPPG signal of one or more fingers is calculated, and a judgment result is obtained according to the rPPG signal.
In particular, rPPG measures subtle intensity variations of the skin of an identified subject by analyzing reflected ambient light. The slight brightness change of the skin is caused by blood flow caused by heart beating, and can be further used for living body judgment. And then filtering the rPPG continuous signal, converting the Fourier transform into a frequency domain, performing frequency domain analysis to obtain characteristics, and classifying the characteristics by using a support vector machine model or a deep neural network model so as to detect the living fingerprint. Optionally, after the position of the finger in each frame is located, a time-series deep neural network (e.g., ConvLSTM) method may also be directly used to determine whether the identified object is a living body, and this method may use more raw input information, so that the result is more stable and reliable.
As described above, in this embodiment, three determination results are obtained from the fingerprint image, the images under different brightness conditions, and the rPPG signal, respectively, and the three determination results are integrated to obtain the identification result, which may be used to digitally represent the determination results, where if the determination results show that the identified object is a living body, the determination result is 1, if the determination results show that the identified object is a non-living body, the determination result is 0, weights are set for the three determination results, and the sum of the weights is equal to 1, and a result value between 0 and 1 is obtained by performing weighted calculation on the three determination results, where if the result value is greater than 0.5, the identified object is a living body, and otherwise, the identified object is a non-living body. When the identification result shows that the identified object is a living body, the fingerprint image of the identified object is continuously processed, for example, the card punching operation of the identified object is completed.
Example 2
In order to better explain the method proposed by the present disclosure, the following description will be given by taking an example of the method in an embodiment of the present disclosure applied to a non-contact acquisition device, in which in the embodiment, a fingerprint image is taken of a finger of an object to be identified, and meanwhile, a plurality of ways are adopted to perform determination according to the fingerprint image, that is, a plurality of determination results are provided, and the plurality of determination results are integrated to obtain an identification result. Specifically, in this embodiment, a fingerprint image of an identified object is captured, and then five determination results are obtained according to a color fingerprint image of the identified object, images of the identified object in different light source colors and using a polarizing plate, an infrared image of the identified object, and an infrared temperature measurement result of the identified object, and the identification results are obtained by integrating the five determination results, which will be described in detail below.
The procedure of obtaining the determination result from the color fingerprint image of the identified object may be the same as embodiment 1, and the description will not be repeated here.
The following describes obtaining the determination result from the images of the identified object under different light source colors.
The method comprises the steps of using light sources with different colors to irradiate light with different colors on an identified object in sequence, shooting images of the identified object under the light with different colors, and carrying out optical feature recognition on the shot images to obtain a judgment result.
Specifically, the wavelengths of the light with different colors are different, and the objects made of different materials can selectively absorb and reflect the light with different wavelengths. The light source emits light of a particular color and, after reflection by the object, it appears in the imaging system that some wavelengths of light are lost. The imaging effect of the finger under different illumination is greatly different because the skin on the finger has different reflectivity to light with different colors, the light sources with different colors are controlled to be sequentially turned on, the images under different light source colors are collected, and after the images under different light source colors are obtained, whether the identified object is a living body can be judged through methods such as support or a deep neural network. For example, a large number of non-contact images of a living finger and a non-living finger under the irradiation of light sources of different colors can be collected, the reflectivity of the non-contact images under the irradiation of light of different colors is calculated, and whether the identified object is a living body is determined by using the difference of the reflectivity through a support vector machine.
The following describes in detail the determination results obtained from images of the identified object in different light source colors and using a polarizing plate.
The method comprises the steps of using light sources with different colors to irradiate light with different colors on an identified object in sequence, and shooting images of the identified object under the light with different colors, wherein a polaroid is arranged on a camera lens for shooting, the polaroid is arranged on the light source, and the shot images are subjected to optical feature recognition to obtain judgment results.
Specifically, different materials may exhibit different optical properties under different color light sources plus polarizers. A polaroid can be added in front of a camera lens for shooting images, another polaroid is placed in front of a light-emitting source, the polarization directions of the two polaroids form a certain angle, a plurality of light sources can be provided, different light sources correspond to the polaroids in different polarization directions, and the accuracy of living body identification is improved by obtaining more fingerprint images under different illumination conditions. Fig. 9 shows the fingerprint image of the same finger under different color light sources, with one polarizer added on the camera lens and another polarizer placed on the light source, and the polarization directions of the polarizers on the camera lens and the polarizers on the light sources are crossed. The light with different wave bands is illuminated on the finger, different illumination effects can be shown after passing through the polaroid, and whether the identified object is a living body or not is determined by analyzing the optical properties of the images of the living body finger and the non-living body finger through a support vector machine model or a deep neural network model.
The following specifically describes obtaining a determination result from an infrared image of an identified object.
Additional infrared fill light, etc. may be provided, as well as an infrared filter (IR-Cut filter) on the camera. When a visible light image of a finger is shot, visible light is used for illumination, and an infrared filter is used for filtering infrared light; when the infrared image is shot, the infrared light supplement lamp is turned on, the infrared filter is turned off to shoot the infrared image, or an additional infrared camera is adopted to shoot the infrared image.
Specifically, the human finger surface skin has a low reflectance to infrared light compared to visible light, and infrared light can penetrate even the skin of the finger and strike the finger vein under the skin. Therefore, when the recognized object is a living body, an infrared image can be captured. For the obtained infrared image of the object to be recognized, the optical property of the finger surface of the infrared image can be analyzed, and whether the object to be recognized is a living body or not is determined by using a support vector machine or a deep neural network, or the living body is detected by other methods; or, the feature of the finger vein in the infrared image may be extracted, and whether the finger vein is the same as the pre-recorded finger vein feature of the living body is determined by comparing the finger vein features, for example, the finger vein feature of the living body corresponding to the fingerprint image in the infrared image is obtained as the feature to be compared, and the finger vein feature in the infrared image is compared with the feature to be compared. The finger vein features are difficult to acquire compared with fingerprints, and the reliability of non-contact fingerprint living body identification is greatly improved by comparing the fingerprints with the finger veins. Optionally, a series of infrared photographs of a living finger and a non-living finger are collected in advance, x _ i represents a photograph, y _ i represents whether the photograph is a living photograph, if the photograph is a living photograph, y _ i is 1, otherwise y _ i is 0, and the collected photograph is used to train a deep neural network model F including a convolutional layer, a pooling layer and a fully-connected layer, so that F (x _ i) -y _ i are associated with x _ i and y _ i through the deep neural network model F, thereby obtaining a determination result by using the deep neural network F.
The following specifically describes obtaining a determination result from an infrared temperature measurement result of an identified object.
The palm area of the identified object is shot by adopting an infrared temperature measurement sensor, the temperature of the palm area is directly read, if the temperature is within a preset temperature range, the identified object is judged to be a living body, otherwise, the identified object is not the living body. When the area array temperature measurement sensor is adopted, the average temperature of the palm area can be calculated to judge whether the identified object is a living body, and an infrared temperature image can be utilized to accurately judge whether the identified object is a living body fingerprint by combining with a deep neural network.
After five judgment results of the five modes are obtained, an SVM/Boosting machine learning method can be adopted to combine the judgment results of the various methods to obtain a more accurate final result. For example, if there are m methods, the decision value represents the determination result of each method, the decision value is Z _ 1.. Z _ m, and y _ i represents whether the finger is real or not, the SVM/Boosting machine learning method enables G (Z _ 1.., Z _ m) -y _ i through the data learning function G, thereby further improving the accuracy of living body detection by using various methods.
In order to implement the technical solution of the living body identification method in the non-contact fingerprint identification method in the embodiments of the present disclosure, some embodiments of the present disclosure further provide a non-contact living body identification device, including:
an acquisition unit configured to acquire an image of at least a part of a fingerprint of at least one finger of an identified object by a non-contact photographing manner;
the living body identification unit is used for judging whether the identified object is a living body in at least one mode to obtain at least one judgment result and obtaining an identification result according to the at least one judgment result;
a fingerprint recognition unit for performing fingerprint recognition on the image in a case where the recognition result shows that the recognized object is a living body.
[ fingerprint Collection and comparison method ]
One of the differences between the above-described non-contact fingerprint identification method and the existing contact fingerprint identification method in the present disclosure is that a plurality of fingerprints can be simultaneously collected and identified. However, in this case, since each fingerprint needs to be compared with a large number of fingerprints stored in the library, the comparison speed is reduced, and the user experience is impaired. Therefore, in view of the above problems, the present disclosure discloses a fingerprint collection comparison method to solve the technical problem.
Example 1
Next, a flowchart of a fingerprint collection method to which the fingerprint collection comparison method of the present disclosure is applied is explained. As shown in fig. 5, the fingerprint acquisition method of the present embodiment includes:
s111, an image acquisition step, namely acquiring at least one image containing fingers;
here, the content of acquiring the finger image is specifically described in the section of the non-contact fingerprint identification method, and only the difference point is taken as the center for description, and the rest of the content can be referred to the description in the section. In addition, in addition to the manner of taking a photograph to obtain an image of a finger, an image that has been taken of the finger may be obtained from the outside in the present embodiment, which will be described in detail below.
S112, an incidental information determination step of determining incidental information of the finger; wherein the incidental information includes at least one of: finger position information, left and right hand information, mirror image information and finger number information.
Here, the incidental information determining step S112 further includes a finger position information determining step:
dividing at least one finger included in the image into separate regions, respectively; determining the central position of each area; determining the sequence of the regions clockwise or counterclockwise; the finger position information is determined according to the sequence. In one or more embodiments, as shown in fig. 15, fingers including a first joint in an image are identified, a finger frame region of each finger is determined, where 4 finger frames are taken as an example, center positions of the 4 finger frames are found, and finger position information is determined by combining left-hand information and right-hand information according to an order of determining the frames clockwise/counterclockwise relative to the center.
S113, a combination setting step of setting at least one combination mode based on the finger and the additional information;
here, the meaning of different combinations is explained, where, for example, fa, fb, fc, fd... fj represents the picture order at the time of submission followed by a number representing the corresponding finger of the fingerprint picture, e.g., fa2, fb3 represents one submission, respectively finger 2 and 3; for example, fa, fb, fc represents submitting three fingerprints at a time, without finger bits.
Given a finger position single finger: single finger submission at a time, with finger positions, such as fa 3;
given the finger position multi-finger: multiple fingers are submitted at a time and all need to be provided with fingers, such as fa2, fb3, fa2, fb7, fa2, fb3, fc4 and fd 5;
no single finger to assign a finger: a single finger submitted at a time, and without a finger position, such as fa;
not assigning a finger to both fingers: multiple fingers are submitted at a time without finger positions, such as fa, fb (continuous finger), fa, fb, fc, fd (continuous finger);
left and right hand with fingers (variable number): multiple fingers are submitted at a time, without finger positions, are connected, and left-hand and right-hand information is given, such as fa, fb (connected finger, left hand), fa, fb, fc, fd (connected finger, right hand).
In one or more embodiments, the combination setting step assumes a possible fingerprint-to-finger correspondence based on existing information, e.g., given 3 fingerprints abc, with accompanying information "is a continuous finger", assuming abc-left index finger, middle finger, ring finger.
S114, a data storage step, namely extracting fingerprint characteristics of the finger and storing the fingerprint characteristics and the additional information as fingerprint data in at least one combination mode;
in one or more embodiments, the method further includes a normalization step of normalizing the acquired images of the finger respectively to acquire fingerprint information;
referring to fig. 6, fig. 6 is a flowchart illustrating processing of images acquired according to different modes according to an embodiment of the disclosure. And in the step of normalization, the density of the lines is adjusted to the preset density of the lines according to the preset frequency to obtain the normalized fingerprint image. As can be seen from fig. 6, compared to the image obtained by non-contact shooting, the image obtained by contact shooting does not need to acquire the finger position, the finger contour, and the finger pitch line information, but directly extracts the fingerprint frequency/fingerprint line density information for adjustment. In other words, the finger contour, such as the length of the finger knuckle, the area of the contour line, etc., belongs to the coarse adjustment reference, and the fingerprint ridge density belongs to the fine adjustment reference. When the area of some images is relatively small, for example, when finger shape and finger pitch line information are not included, the image obtained by non-contact shooting can also be adjusted by using only the density of the lines as a reference.
In one or more embodiments, the method further includes a scale expansion step of performing scale expansion on the acquired fingerprint information at least once to acquire at least one scale of fingerprint samples.
In one or more embodiments, the method further comprises an image unfolding step of unfolding the image of the finger to perform analog conversion and obtain at least one piece of fingerprint data of the finger.
According to the fingerprint identification method, under the condition that auxiliary hardware except for an image acquisition device such as a mobile phone is not relied on, the sizes of fingerprint images shot by the image acquisition device or acquired by a contact type acquisition instrument are adjusted, so that the fingerprint images acquired by different batches of the same finger are all zoomed and adjusted to be close to the same size, the relative distance and the angle error of fingerprint characteristic lines are kept not to exceed a certain threshold value, such as 5%, and further the subsequent processing of a fingerprint comparison system is supported.
Example 2
Next, a flowchart of a fingerprint comparison method to which the fingerprint collection comparison method of the present disclosure is applied is explained. As shown in fig. 7, the fingerprint comparison method of the present embodiment includes:
s131, an image acquisition step, namely acquiring at least one image containing fingers;
here, the content of acquiring the finger image is specifically described in the section of the non-contact fingerprint identification method, and only the difference point is taken as the center for description, and the rest of the content can be referred to the description in the section.
S132, an incidental information determination step of determining incidental information of the finger;
wherein the incidental information includes at least one of: finger position information, left and right hand information, mirror image information and finger number information. The finger position information, the left hand information, the right hand information, the mirror image information and the finger quantity information at least comprise one of the following information: giving a finger position single finger, giving a finger position multi-finger, not giving a finger position single finger, not giving a finger position connecting finger, and connecting the left hand and the right hand with the finger.
In one or more embodiments, the incidental information determining step S132 further includes a finger position information determining step of: dividing at least one finger included in the image into separate regions, respectively; determining the central position of each area; determining the sequence of the regions clockwise or counterclockwise; the finger position information is determined according to the sequence. In one or more embodiments, fingers including a first joint in the image are identified, a finger frame area of each finger is determined respectively, here, taking 4 finger frames as an example, the center positions of the 4 finger frames are found, the frames are determined according to the sequence of clockwise/counterclockwise directions relative to the center, and finger position information is determined by combining left-hand and right-hand information.
S133, a combination setting step of setting at least one combination mode based on the finger and the incidental information;
in one or more embodiments, for example, based on existing information, a possible fingerprint-to-finger correspondence is first assumed, e.g., given 3 fingerprints abc, with accompanying information "is a continuous finger", assuming abc-left index finger, middle finger, ring finger.
In one or more embodiments, for example, based on the information of whether there is a finger or not and whether there is a finger position or not, all the hypotheses that have been obtained currently are expanded without changing other conditions, such as abc-left index finger, middle finger, and ring finger, and can be expanded into the following three hypotheses:
abc-left hand ring finger, middle finger, index finger
Abc-left hand middle, ring, little finger
Abc-left little, ring, middle finger
In one or more embodiments, all hypotheses that have been currently obtained, e.g., existing hypotheses abc-left-hand ring finger, middle finger, index finger, are expanded without specifying the left and right hand, e.g., based on existing left and right hand information, to: abc-right hand ring finger, middle finger, index finger.
In one or more embodiments, all assumptions that have been currently obtained are extended without changing other conditions, e.g., based on whether information is already mirrored. For example, suppose abc-right hand ring finger, middle finger, index finger. In the case that whether the fingerprint data is mirrored or not is not specified, the following is extended: abc-right hand ring finger, middle finger, index finger. The finger position is assumed to be unchanged, but all fingerprint data is mirrored left and right.
In one or more embodiments, the method further includes a normalization step of normalizing the acquired images of the finger respectively to acquire fingerprint information; the method for acquiring the image of the finger is a shooting acquisition method, and the normalization step comprises the following steps: judging whether the shooting mode is adopted; if yes, extracting one of finger position information, finger contour information and finger joint line information of the fingerprint data; extracting fingerprint frequency information according to one of the finger position information, the finger outline information and the finger node line information; and normalizing the fingerprint data according to the fingerprint frequency information. Here, the finger contour and knuckle line recognition are provided by a deep learning model from which the approximate size of the finger that needs to be scaled can be estimated.
In one or more embodiments, a method of extracting fingerprint frequency information includes: further local scaling, which depends on the calculation of fingerprint frequency, and the specific method is to take a local fingerprint image inside a sliding window; the minimum translation invariant distance is calculated by utilizing the translation invariance between approximately parallel lines of the fingerprint image, so that the fingerprint frequency is estimated.
In one or more embodiments, the normalizing step further includes normalizing the identification card fingerprint image and the fingerprint image, and for the identification card fingerprint image, the area is relatively small, the identification card fingerprint image generally does not contain the finger shape and the finger pitch line information, and the normalization can be performed by adopting the fingerprint line density; for the fingerprint image collected by the contact type collector, the frequency of the lines can be measured, so that the compatibility of the image is kept for normalization processing.
Here, the content of the normalization process is specifically described in the above section, and specific reference may be made to the description in the above section.
In one or more embodiments, the method further includes a scale expansion step, in which the obtained fingerprint information is subjected to scale expansion at least once to obtain fingerprint samples with at least one scale, so that each finger corresponds to a plurality of fingerprint pictures in each person, and the number of the fingerprint pictures is respectively influenced and determined by a scale strategy and an expansion scheme;
in one or more embodiments, the method further includes an image unfolding step of unfolding the image of the finger for analog conversion, obtaining at least one piece of fingerprint data of the finger, performing an unfolding algorithm on the fingerprint image obtained in the form of each photo, simulating the captured fingerprint into a planar fingerprint, and obtaining a plurality of fingerprints by using different unfolding parameters, wherein the number is determined by the number of fingerprints and the given information.
S134, a fingerprint comparison step, namely extracting fingerprint characteristics of the finger and comparing the fingerprint characteristics with the additional information with the existing fingerprint data according to at least one combination mode;
in one or more embodiments, for example, the fingerprint data is acquired, for example, by at least one of the following acquisition modes:
fingerprint collection appearance: the acquisition instruments have various forms, and can finish acquisition of single-finger, multi-finger, single-hand, four-finger, two-hand thumb and the like by one-time acquisition; the technology of the acquisition instrument can be roughly divided into a contact acquisition instrument and a non-contact acquisition instrument;
directly shooting to obtain a fingerprint: for example, the mobile device is generally used for shooting, and the method can also be generalized to shooting fingers to obtain data by using the mobile device/fixed device;
fingerprint data reading: such as an identification card reader; ID cards such as social security cards and drivers' licenses.
In one or more embodiments, for example, at least one of the following information of the fingerprint data being compared is the same or different: the acquisition mode information, the pixel point information, and the data format information may also include other different information, for example: for the fingerprint, the fingerprint is generally output by a fingerprint acquisition instrument or is read from a medium for storing the fingerprint, and is mostly a black and white picture, and different prints of the fingerprint are mainly distinguished in several dimensions such as DPI, pixel, color, gray level and the like; the finger photo image is generally obtained by shooting, and can be provided by a non-contact fingerprint acquisition instrument; for the fingerprint features, the fingerprint feature file may be read from a medium storing the fingerprint features, or the fingerprint features may be extracted after the fingerprints are collected by some fingerprint collecting devices.
S135, a result determining step, namely determining a final comparison result according to comparison results of different combination modes;
in one or more embodiments, the result determination step S135 further includes,
and a score obtaining step, namely obtaining the highest score obtained by comparing the fingerprint characteristics of at least one finger with the existing fingerprint data in at least one combined mode as the comparison score of the finger.
And a score summarizing step of summarizing the comparison scores of different fingers, and calculating the weighted square sum of the comparison scores of the multiple fingers in at least one combined mode to be used as a final comparison score.
And taking the maximum value in the final alignment scores of at least one combination mode as a final alignment result.
In order to implement the technical solution of the fingerprint collection and comparison method in the embodiments of the present disclosure, an embodiment of the present disclosure provides a fingerprint collection and comparison device, which can be applied to various electronic terminal devices, as shown in fig. 12, and is characterized by including,
an image acquisition module 191 for acquiring at least one image containing a finger;
an incidental information determination module 192 for determining incidental information of the finger; wherein the incidental information includes at least one of: finger position information, left and right hand information, mirror image information and finger number information.
In one or more embodiments, the incidental information determination module 192 further comprises a finger position information determination module for dividing at least one finger included in the image into separate areas, respectively; determining the central position of each area; the detailed implementation can be referred to the contents explained in the above method.
A combination setting module 193 for setting at least one combination method based on the finger and the incidental information; the explanation of the combination method is the same as that of the above method, and is not repeated here.
The data storage module 194 is used for extracting fingerprint characteristics of the finger and storing the fingerprint characteristics and the additional information as fingerprint data in at least one combination mode;
in one or more embodiments, the fingerprint identification method further includes a normalization module, configured to normalize the acquired images of the finger to acquire fingerprint information, where the manner of acquiring the images of the finger is a shooting acquisition manner, and the normalization module is configured to: judging whether the shooting mode is adopted; if yes, extracting one of finger position information, finger contour information and finger joint line information of the fingerprint data; extracting fingerprint frequency information according to one of the finger position information, the finger outline information and the finger node line information; normalizing the fingerprint data according to the fingerprint frequency information;
in one or more embodiments, the normalization module is configured to extract fingerprint frequency information according to the fingerprint frequency information, and includes: scaling a fingerprint image of the fingerprint data; the minimum translation invariant distance is calculated using the translation invariance between approximately parallel lines of the fingerprint image.
In one or more embodiments, the normalization module is further configured to perform normalization processing on the identification card fingerprint image and the fingerprint image, and for the identification card fingerprint image, the area is relatively small, the identification card fingerprint image generally does not contain information of the shape and the finger pitch line, and the normalization may be performed by adopting the density of the fingerprint line; for the fingerprint image collected by the contact type collector, the frequency of the lines can be measured, so that the compatibility of the image is kept for normalization processing.
The normalization module is used for implementing the functions in the above method, and will not be described in detail here.
In one or more embodiments, the system further includes a scale expansion module, configured to perform scale expansion on the acquired fingerprint information at least once to acquire a fingerprint sample with at least one scale, so that each finger corresponds to multiple fingerprint pictures at the same time, and the number of the fingerprint pictures is respectively influenced and determined by a scale strategy and an expansion scheme;
in one or more embodiments, the fingerprint image processing device further includes an image unfolding module, configured to unfold the image of the finger for analog conversion, obtain at least one piece of fingerprint data of the finger, perform an unfolding algorithm on the fingerprint image obtained in the form of each photo, simulate the captured fingerprint into a flat fingerprint, and obtain a plurality of fingerprints using different unfolding parameters, where the number is determined by the number of fingerprints and the amount of given information.
The fingerprint comparison module 195 is used for extracting fingerprint characteristics of the finger and comparing the fingerprint characteristics with the additional information with the existing fingerprint data according to at least one combination mode;
a result determining module 196, configured to determine a final comparison result according to the comparison results of different combination manners;
in one or more embodiments, the result determination module further comprises,
the score acquisition module is used for acquiring the highest score obtained by comparing the fingerprint characteristics of at least one finger with the existing fingerprint data in at least one combined mode as the comparison score of the finger;
the score summarizing module is used for summarizing the comparison scores of different fingers and calculating the weighted square sum of the comparison scores of the multiple fingers in at least one combined mode to be used as a final comparison score;
and the result determining module takes the maximum value in the final comparison scores of at least one combination mode as a final comparison result.
In one or more embodiments, the modules may be integrated or separately implemented as a fingerprint acquisition device or a fingerprint comparison device, which is not limited.
[ fingerprint region detection method ]
In addition, as described above, in the non-contact fingerprint identification method of the present disclosure, since the entire finger of a plurality of fingers is photographed, how to quickly detect a fingerprint area is important for improving the fingerprint identification speed. Therefore, in view of the above problems, the present disclosure discloses a set of fingerprint region detection methods to address this technical problem.
Fig. 8 illustrates a flowchart of a fingerprint area detection method according to an embodiment of the present disclosure. The fingerprint area detection method of one embodiment of the present disclosure mainly includes the following steps:
s141, an image acquisition step, wherein the hand of the object is shot in a non-contact mode, and an image of at least one finger containing a finger joint is acquired;
here, the content of acquiring the finger image is specifically described in the section of the non-contact fingerprint identification method, and only the difference point is taken as the center for description, and the rest of the content can be referred to the description in the section.
S142, a direction adjusting step, namely adjusting the direction of the finger to enable the finger to face a preset direction;
in one or more embodiments, the direction adjusting step comprises: identifying end points of knuckles in the image; determining the direction of a knuckle line according to the end point of the knuckle; and adjusting the image according to the direction of the knuckle line, so that the knuckle line in the adjusted image faces to the direction vertical to the preset direction. For example, a picture of a finger head including a first knuckle in an image is recognized, a first knuckle end point is recognized on the picture, a knuckle line direction is determined according to the end point of the first knuckle, and the picture is adjusted according to the knuckle line direction, so that the knuckle line direction in the adjusted picture is kept in a horizontal direction.
In one or more embodiments, the direction adjusting step may further include: dividing at least one finger containing a finger joint in the image into separate areas respectively; determining a direction of the divided individual regions; the direction of the individual regions is adjusted so that the finger is directed in a preset direction, for example, four fingers are respectively divided into four individual regions, the direction of the pointing direction of the finger of the divided individual regions is determined, and the direction of the pointing direction of the finger of each individual region is directed in a vertically upward direction.
In one or more embodiments, the direction adjusting step may further include: determining the direction of the outer edge of at least one finger in the image containing the knuckle; the direction of the outer edge is adjusted to adjust the direction of the finger so that the finger is directed in a preset direction, for example, the outer edges of four fingers are determined respectively, the direction in which the formed outer edge image is directed is determined, and the direction in which each outer edge is directed in a vertically upward direction.
S143, a background processing step, namely performing preset processing on the background in the image;
in one or more embodiments, the background processing step comprises: identifying the edges of the fingers and the background on the image and forming polygons of the edges; and acquiring an image inside the polygon as a foreground, and removing a background outside the polygon.
In one or more embodiments, the background processing step may further include: dividing at least one finger containing a finger joint in the image into separate areas respectively; acquiring a finger image in the individual region; removing the background outside the finger image in the individual areas, for example, dividing four fingers into four individual areas respectively, acquiring the finger image in each individual area, determining the background outside the finger image of the divided individual areas, and removing the background outside the finger image of each individual area.
S144, determining the finger position, namely determining the finger position information of the finger;
in one or more embodiments, the finger position determining step comprises: dividing at least one finger containing a finger joint in the image into separate areas respectively; determining the central position of the area; determining the sequence of the regions clockwise or counterclockwise; the finger position information is determined according to the sequence. For example, a finger picture including a first joint in the image is recognized, a finger frame is formed for each finger, where the finger frame may be 4 finger frames or 5 finger frames, where 4 finger frames are taken as an example, the center positions of the 4 finger frames are found, the order of the finger frames is determined in a clockwise/counterclockwise direction with respect to the center, and the finger position information is determined by combining left-hand and right-hand information.
S145, an area detection step, namely detecting the fingerprint area of the finger at least according to one of the preset direction of the finger, the preset processing background and the finger position information of the finger.
In one or more embodiments, the region detecting step further comprises tracking the finger.
In one or more embodiments, the method further comprises a scale obtaining step of obtaining scale information corresponding to the physical world by using the width of the finger.
In order to implement the technical solution of the fingerprint area detection method in the embodiment of the present disclosure, an embodiment of the present disclosure provides a fingerprint area detection apparatus, which may be applied to various electronic terminal devices, as shown in fig. 13, and is characterized by including,
an image acquisition module 201 for non-contact shooting a hand of a subject and acquiring an image of at least one finger including a finger joint;
a direction adjusting module 202, configured to adjust a direction of the finger so that the finger faces a preset direction; for example, the method is used for recognizing a picture of a finger head including a first knuckle in an image, recognizing an end point of the first knuckle on the picture, determining a knuckle line direction according to the end point of the first knuckle, and adjusting the picture according to the knuckle line direction, so that the knuckle line direction in the adjusted picture is kept in a horizontal direction.
In one or more embodiments, the direction adjustment module 202 is further configured to: dividing at least one finger containing a finger joint in the image into separate areas respectively; determining a direction of the divided individual regions; the direction of the individual regions is adjusted so that the finger is directed in a preset direction, for example, four fingers are respectively divided into four individual regions, the direction of the pointing direction of the finger of the divided individual regions is determined, and the direction of the pointing direction of the finger of each individual region is directed in a vertically upward direction.
In one or more embodiments, the direction adjustment module 202 is further configured to: determining the direction of the outer edge of at least one finger in the image containing the knuckle; the direction of the outer edge is adjusted to adjust the direction of the finger so that the finger is directed in a preset direction, for example, the outer edges of four fingers are determined respectively, the direction in which the formed outer edge image is directed is determined, and the direction in which each outer edge is directed in a vertically upward direction.
The background processing module 203 is used for performing preset processing on a background in the image;
in one or more embodiments, the background processing module 203 is to: identifying the edges of the fingers and the background on the image and forming polygons of the edges; and acquiring an image inside the polygon as a foreground, and removing a background outside the polygon.
In one or more embodiments, the background processing module 203 is further configured to: dividing at least one finger containing a finger joint in the image into separate areas respectively; acquiring a finger image in the individual region; removing the background outside the finger image in the individual areas, for example, dividing four fingers into four individual areas respectively, acquiring the finger image in each individual area, determining the background outside the finger image of the divided individual areas, and removing the background outside the finger image of each individual area.
A finger position determining module 204, configured to determine finger position information of a finger; for example, a finger picture including a first joint in the image is recognized, a finger frame is formed for each finger, where the finger frame may be 4 finger frames or 5 finger frames, where 4 finger frames are taken as an example, the center positions of the 4 finger frames are found, the order of the finger frames is determined in a clockwise/counterclockwise direction with respect to the center, and the finger position information is determined by combining left-hand and right-hand information.
The area detection module 205 is configured to detect a fingerprint area of a finger according to at least one of a preset direction of the finger, a preset processed background, and finger position information of the finger.
In one or more embodiments, the system further comprises a scale obtaining module, configured to obtain scale information corresponding to the physical world by using the width of the finger;
a finger tracking module 206 for tracking the finger.
In one or more embodiments, the device further includes a region segmentation module, configured to segment at least one finger in the image, which includes the finger joint, into separate regions, respectively.
[ fingerprint normalization method ]
In addition, as described above, in the non-contact fingerprint identification method of the present disclosure, since the various types of fingerprint data, fingerprint images, and the like are processed, identified, collected, stored, and compared, it is also important how to realize normalization of the various types of fingerprint data and fingerprint images. Therefore, in view of the above problem, the present disclosure discloses a set of fingerprint normalization methods to cope with this technical problem.
As shown in fig. 9, fig. 9 is a flowchart of a fingerprint identification method according to an embodiment of the present disclosure, which includes the following steps.
S161, an acquisition step of acquiring an image including a fingerprint.
Specifically, the embodiments of the present disclosure may be implemented by various image capturing devices, including but not limited to a camera, a mobile phone provided with a camera, a capturing device for capturing a fingerprint by using a contact method, and the like. With the above-described image capturing apparatus, the present embodiment can acquire various forms of images including fingerprints. Here, the content of acquiring the fingerprint image is specifically described in the section of the non-contact fingerprint identification method, and only the difference point is taken as the center for description, and the rest of the content can be referred to the description in the section.
And S162, judging whether the image is obtained by non-contact shooting the fingerprint or not.
Specifically, the embodiment of the disclosure may determine whether the image is an image obtained by non-contact shooting the fingerprint according to the attribute of the image; or judging whether the image comprises the outline and/or the knuckle line of the finger or not according to the deep learning model; if so, the image is obtained by carrying out non-contact shooting on the fingerprint; if not, the image is obtained by carrying out contact type acquisition on the fingerprint.
The deep learning is a general term of a type of pattern analysis method, and in terms of specific research content, the deep learning mainly relates to three types of methods: (1) a neural network system based on convolution operations, i.e. a Convolutional Neural Network (CNN). (2) The self-Coding neural network based on the multilayer neurons comprises two categories of self-Coding (Auto encoder) and Sparse Coding (Sparse Coding). (3) And pre-training in a multilayer self-coding neural network mode, and further optimizing a Deep Belief Network (DBN) of the neural network weight by combining the identification information. Through multi-layer processing, after the initial low-layer feature representation is gradually converted into the high-layer feature representation, the complex learning tasks such as classification can be completed by using a simple model. Deep learning forms a more abstract class or feature of high-level representation properties by combining low-level features to discover a distributed feature representation of the data. The computation involved in producing an output from an input can be represented by a flow graph (flow graph): a flow graph is a graph that can represent a computation, where each node represents a basic computation and a computed value, and the results of the computation are applied to the values of the children of that node. Consider a set of computations that can be allowed in each node and possible graph structure and that define a family of functions.
And S163, a fingerprint parameter acquiring step of acquiring the size and the ridge density of the fingerprint when the image is determined to be the image obtained by non-contact type shooting of the fingerprint in the determining step.
Specifically, embodiments of the present disclosure may include the steps of obtaining a length of a finger; or according to the knuckle line model, obtaining two end points of the knuckle line and obtaining the distance between the two end points; or according to the knuckle line model, acquiring two end points of the knuckle line, and according to the finger contour line model, acquiring a plurality of boundary points of the finger edge; smoothly connecting two end points and a plurality of boundary points in sequence to form a contour line of the finger; calculating the area of the contour line to obtain the area; wherein, in the normalizing step, adjusting the size to a preset size comprises: adjusting the length to a preset length; or adjusting the distance to a preset distance; or adjusting the area to a preset area.
Referring to fig. 10, fig. 10 is a schematic diagram of an outline of an embodiment of the disclosure. As shown in fig. 10, the contour lines may be defined by a plurality of boundary points of the finger edge and finger node lines. Wherein, the point A and the point B are two end points of the knuckle line respectively.
S164, a normalization step, namely adjusting the size to a preset size according to a preset standard, and adjusting the line density to a preset line density to obtain a normalized fingerprint image.
Specifically, the embodiment of the present disclosure may include a step of dividing a region where the fingerprint is located into a plurality of sub-regions; respectively acquiring the fingerprint direction and the ridge density of each sub-area in the plurality of sub-areas; respectively obtaining the minimum translation invariant distance of each sub-area according to the translation invariance between the lines in the same fingerprint direction; respectively adjusting the density of the lines of each sub-area according to the minimum translation invariant distance, so that the difference value of the density of the lines of each adjacent sub-area is smaller than a preset difference value; and obtaining the preset streak line density according to the preset frequency.
In addition, the embodiment of the disclosure can also be used for calculating the minimum translation invariant distance by taking a local fingerprint image in a sliding window and utilizing the translation invariance between approximately parallel lines, so as to estimate the fingerprint frequency.
S165, a fingerprint identification step of performing fingerprint identification based on the normalized fingerprint image obtained in the normalization step.
In addition, the embodiment of the disclosure may further include a step of extracting a fingerprint region from the image to exclude a background region in the image; wherein the fingerprint in the fingerprint area has a substantially uniform ridge density. The disclosed embodiment may further include positioning the finger position of the finger by a combination of user input of the left and right hands and detection of the finger position frame by the deep learning model.
In order to implement the technical solution of the fingerprint identification method in the embodiment of the present disclosure, as shown in fig. 14, an embodiment of the present disclosure further provides a fingerprint identification apparatus, which includes an obtaining module 211, configured to obtain an image including a fingerprint; a judging module 213, configured to judge whether the image is an image obtained by non-contact shooting of a fingerprint; a fingerprint parameter obtaining module 215, configured to obtain a size and a ridge density of the fingerprint when the image is determined to be an image obtained by capturing the fingerprint in the determining module; the normalization module 217 is configured to adjust the size to a preset size according to a preset standard, and adjust the ridge density to a preset ridge density to obtain a normalized fingerprint image; and a fingerprint recognition module 219 operable to perform fingerprint recognition based on the normalized fingerprint image obtained in the normalization module. In addition, the embodiment of the present disclosure may further include an extraction module (not shown) configured to extract a fingerprint region from the image to exclude a background region in the image.
It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and may be performed in other orders unless explicitly stated herein. Moreover, at least a portion of the steps in the flow chart of the figure may include multiple sub-steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed alternately or alternately with other steps or at least a portion of the sub-steps or stages of other steps.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and can include the processes of the embodiments of the methods described above when the computer program is executed. The storage medium may be a non-volatile storage medium such as a magnetic disk, an optical disk, a Read Only Memory (ROM), or a Random Access Memory (RAM).
[ non-contact fingerprint acquisition device ]
Referring to fig. 15, fig. 15 schematically illustrates a non-contact fingerprint capturing device fixed on a wall, in which the non-contact capturing in this embodiment includes: the image acquisition device comprises a shell 1, an image acquisition device 2 and a processing device 3, wherein the shell 1 comprises a first component part 11 and a second component part 12 which are mutually connected into an L shape. The image acquisition device 2 is arranged on the shell 1 and is used for shooting a fingerprint image, and the shooting area of the image acquisition device 2 is positioned on one side of the first component 11 close to the second component 12; the processing device 3 is disposed in the housing 1 and is used for performing recognition processing on the image captured by the image capturing device 2.
Specifically, in this embodiment, the housing 1 may be made of a non-light-transmitting material, the shooting direction of the image capturing device may be facing vertically downward, the image capturing device 2 is used for shooting a non-contact fingerprint image, and the side of the second component 12 away from the first component 11 may be used for fixing, for example, on a wall as shown in fig. 15. Since the first component 11 and the second component 12 form an L-shaped structure, and the shooting area of the image capturing device 2 is located on the concave side of the L-shape, the first component 11 and the second component 12 can block most of stray light, thereby reducing the interference of the stray light on capturing the non-contact fingerprint image. In the prior art, when the contact type fingerprint is collected, the finger is in contact with the device for collecting the image, so that the interference of stray light is not needed to be worried, and when the non-contact type fingerprint is collected, the stray light can generate great influence on the shooting of the fingerprint image, so that the shell 1 in the embodiment can at least block the stray light in two directions to ensure the quality of the image collected by the image collecting device 2.
In some alternative embodiments, as shown in fig. 16, the housing 1 further comprises: a third component 13; in this embodiment, the first component 11, the second component 12, and the third component 13 are sequentially connected, the first component 11 and the third component 13 are oppositely disposed at an interval, the first component 11, the second component 12, and the third component 13 surround to form a collecting space, and the shooting area of the image capturing device 2 is located in the collecting space. As shown in fig. 16, the first component 11, the second component 12 and the third component 13 form a concave structure around them, in some embodiments, the shooting direction of the image capturing device 2 faces the third component 13, and at this time, the third component 13 blocks the light directly shot into the image capturing device 2, so that the veiling glare shot into the image capturing device 2 is greatly reduced, and the definition of the image captured by the image capturing device is further improved.
In some embodiments of the present disclosure, with continuing reference to fig. 16, the contactless image capturing apparatus further includes: and the illuminating device 4 is arranged in the shell 1 and used for illuminating when the image acquisition device 2 shoots. Specifically, the lighting device 4 may include one or more light sources, and the finger photographed by the image capturing device 2 is illuminated by the lighting device 4, so that the brightness of the photographed finger can be significantly improved, and the contrast of the fingerprint can be improved.
In some embodiments of the present disclosure, please refer to fig. 17, the illumination device 4 includes: at least two lighting members 41; the illumination components 41 are used for illuminating when the image acquisition device 2 shoots, and the colors of the light rays emitted by at least two illumination components 41 are different, so that the image acquisition device can shoot fingerprint images under different light ray colors. Specifically, when fingerprint collection is carried out, light of different colors can be emitted in sequence, so that fingerprint images under illumination of different colors can be acquired, false fingers can be effectively prevented by adopting the mode, and because the reflectivity of materials such as human bodies and rubber under illumination of different colors is different, the reflectivity of the photographed fingers can be calculated according to the images under illumination of different colors, so that whether the photographed fingers are human fingers or fake false fingers can be judged.
In some embodiments of the present disclosure, referring to fig. 18, the non-contact fingerprint acquisition device in this embodiment further includes: the light path adjusting device 5 is arranged on the side face, close to the shooting area of the image acquisition device 2, of the shell 1, the lens of the image acquisition device 2 faces the light path adjusting device 5, the light path adjusting device 5 is used for changing the shooting direction of the image acquisition device 2, and the image acquisition device 2 shoots the fingerprint image of the shooting area through the light path adjusting device 5. In this embodiment, the reason for setting the optical path adjusting device 5 is that when the distance between the image capturing device 2 and the finger is small, the captured fingerprint image is easy to distort, if the optical path adjusting device 5 is not additionally set outside the image capturing device 2, in order to ensure that the captured fingerprint image is not distorted, it is necessary to ensure that the distance between the image capturing device 2 and the captured finger is large enough, which will result in the increase of the size of the non-contact fingerprint capturing device, which is not beneficial to the miniaturization, and after the optical path adjusting device is adopted, the distance between the image capturing device 2 and the captured finger is increased, the depth of field is increased, thereby ensuring that the captured fingerprint image is clear enough without increasing the size of the whole non-contact fingerprint capturing device.
In some alternative embodiments, with continued reference to fig. 18, the optical path adjustment device 5 includes: a reflective mirror; the reflector is arranged on the side face, close to the shooting area of the image acquisition device 2, of the shell 1, and a preset included angle is formed between the plane where the reflector is located and the shooting direction of the image acquisition device 2. When shooting fingers, the image of the fingers is reflected by the reflector and shot into the lens of the image acquisition device 2, and the shooting area of the image acquisition device 2 can be adjusted by adjusting the preset included angle between the shooting direction of the image acquisition device and the reflector.
In some embodiments of the present disclosure, with continued reference to fig. 16, the contactless fingerprint acquisition device further includes: and the structured light projection device 6 is arranged on the shell 1 and is used for projecting structured light when the image acquisition device 2 shoots so that the image acquisition device 2 shoots a structured light image. The strips with the structured light of the finger image of the user can be captured by adding the structured light projection device 6, when the image is processed by the processing device 3, the finger can be subjected to three-dimensional modeling according to the strips, and then the finger image which is recognized and cut out is unfolded to obtain the finger image which is closer to the pressing.
In some embodiments of the present disclosure, as shown in fig. 16, the contactless fingerprint acquisition apparatus further includes: a time-of-flight device 7; the flight time device 7 is used for emitting infrared light pulses to the object to be photographed and calculating the depth information of the object to be photographed according to the light signals returned by the object to be photographed; the image acquisition device 2 is used for focusing when shooting the fingerprint image of the object to be photographed according to the depth information. Specifically, in this embodiment, a time-of-flight ranging method is adopted, the time-of-flight device 7 can transmit continuous infrared light pulses with specific wavelengths to an object to be photographed (for example, a finger), receive an optical signal transmitted back by the object to be photographed through a sensor on the time-of-flight device 7, calculate the time of flight or the phase difference of light back and forth to obtain three-dimensional depth information of the object to be photographed, in the prior art, the time spent by a camera during focusing is often long, and the user experience is poor, so that a time-of-flight device (TOF module) is added in this embodiment, the distance from an image acquisition device to the finger is measured, and then the distance is directly specified during focusing, so that the effect of rapid.
In some embodiments of the present disclosure, the image capturing apparatus 2 includes: a plurality of cameras, at least two cameras of the plurality of cameras being focused at different positions; thus, images at a plurality of positions can be shot simultaneously, the area where the fingers can be placed is enlarged, the depth of field of the plurality of cameras can be improved, if the depth of field of each camera is kcm, a system with the total depth of field k × ncm can be assembled by using the n cameras, and an image with the highest quality can be selected for each finger through definition during shooting. In some embodiments, the imaging system of the liquid lens of the image capture device can change the focal plane within milliseconds to provide a sharp image regardless of the distance of the object from the camera, making an integrated liquid lens an ideal choice for multiple distance focus shooting scenarios, which can take 250 images per second with the required energy, which consists of a pair of water droplets that shake back and forth when exposed to high frequency sound waves, which in turn changes the focal length of the lens, automatically steps the frame within the focal length range and discards the frame outside the focal length by software. In some embodiments, the image capturing device 2 has a frazier lens, which has a very large depth of field and can shoot very far and very near objects at the same time, and the frazier lens uses a large wide-angle lens for imaging, projects the influence of the large depth of field on a film, and then uses a zoom lens to capture the image on the film to zoom and control the aperture, and the frazier lens realizes the depth of field far beyond the depth of field of a common lens by this design. In some embodiments, the image acquisition apparatus 2 includes: a light field camera. The light field camera can capture information about the light direction of a scene and record data of light beams in all directions, so that the light field camera can focus any depth in a shot picture, and can focus through software according to actual picture requirements in the later period to obtain a clearer picture effect.
The embodiment of the application also provides a non-contact fingerprint acquisition method, and the method in the embodiment comprises the following steps:
(1) and an acquisition step, adopting a non-contact type to shoot a fingerprint image to be identified of the finger.
In the collecting step, the non-contact fingerprint collecting device provided by the present application is used to collect a fingerprint image, in the collecting step, the finger needs to be identified, and the process of identifying the finger can be the finger identifying step in any embodiment of the present application.
(2) And a processing step, namely segmenting each finger in the fingerprint image to be identified and obtaining the fingerprint image corresponding to each finger.
Specifically, one or more fingers are captured in the captured fingerprint image, so that in order to separately process the fingerprint images of the fingers, the fingers in the captured fingerprint image need to be segmented to obtain the fingerprint images of the fingers. The processing step specifically employs a fingerprint acquisition step in any embodiment of the present application.
(3) And a judging step, namely determining whether the condition is met according to the number of the fingers and the fingerprint image corresponding to each finger.
(4) And a storage step of storing the fingerprint images corresponding to the fingers when the conditions are met.
Specifically, in this embodiment, the preset condition is provided, the fingerprint image corresponding to each finger is stored only when the condition is met, and the fingerprint image corresponding to each finger is not stored when the condition is not met, where the preset condition may be, for example, that the sharpness of the acquired fingerprint image meets the requirement, the acquired fingerprint image is determined to be a fingerprint image of a living body, and the acquired fingerprint image is stored when the condition is met, so that the fingerprint image acquisition method can be used for operations such as fingerprint comparison, and abandons storage and reacquires the fingerprint image when the condition is not met.
In some embodiments of the present application, the capturing step may include taking multiple sets of differently focused fingerprint images to be identified. Specifically, each finger of the human body is not in the same plane, the finger has fluctuation, if only one fingerprint image is collected, the definition of different fingers possibly shot is insufficient, and the fingerprint images of each finger can be ensured to be clear enough by collecting a plurality of groups of different fingerprint images to be identified which are focused.
In some embodiments of the present application, the relative pitch of the camera lens is reduced when the fingerprint image to be recognized is photographed, and light is supplemented when the fingerprint image is photographed. Specifically, the relative hole distance and the depth of field of the lens are approximately in an inverse proportion relationship, the depth of field can be improved by reducing the relative hole distance of the camera lens, the depth of field can be improved, large-range non-contact fingerprint collection can be achieved without refocusing, the light inlet quantity can be reduced by reducing the relative hole distance, light supplement is needed during shooting, and therefore the quality of a shot fingerprint image is guaranteed.
In some embodiments of the present application, the fingerprint image to be identified is a structured light image; the storing step further comprises: and when the fingerprint image meets the conditions, a corresponding 3D fingerprint model is established according to the fingerprint image, the fingerprint image is covered on the corresponding 3D fingerprint model, the 3D fingerprint model is unfolded to obtain a corresponding 2D fingerprint image, and the 2D fingerprint image is stored. The non-contact fingerprint collection device who adopts in this embodiment has structured light projection unit, what shoot is structured light image, adopts the structured light technique can obtain 3D image information, improves the precision of 3D formation of image, consequently can acquire the 3D information of finger in this implementation, has recorded 3D information in the 2D image of storage, compares in the information in the fingerprint image of the mode of current fingerprint collection greatly enriched, is favorable to the later stage to improve the fingerprint and compares the precision.
In some embodiments of the present application, the processing step comprises: identifying a first finger joint line of the fingerprint image to be identified, and acquiring a finger head image from the fingerprint image to be identified according to the first finger joint line; and acquiring a finger foreground image from the finger head image as a fingerprint image. Specifically, the finger includes a plurality of knuckles, the knuckle line that is located the most near finger top is first knuckle line, the fingerprint is located the head of finger, the finger head includes first knuckle, consequently only need gather and handle finger head image, need not to gather and handle other parts of finger, gather the data handling capacity that handles and can reduce the later stage to finger head image, improve response speed, it is same, there is not fingerprint information in the background of finger head image, consequently, only need handle finger foreground image. In some embodiments, obtaining a finger foreground image from the finger head image comprises: and identifying the edges of the fingers and the background in the finger head image, and acquiring a finger foreground image from the finger head image according to the edges.
In some embodiments, identifying a first finger joint line of a fingerprint image to be identified comprises: identifying a region where the finger head including a first finger joint is located in the fingerprint image to be identified; identifying a knuckle endpoint for a first knuckle; and determining a knuckle line of the first knuckle according to the knuckle endpoints. Specifically, in this embodiment, the area where the finger head is located is first found in the fingerprint image to be recognized, because the finger head is located at the end of the finger, the area where the finger head is located can be found more quickly, the area where the finger head is located must include the first finger joint line, and because the inside of the finger has a plurality of routing lines such as the finger line, it is difficult to directly find the first finger joint line, and therefore, the finger joint end point of the first finger joint on the contour is found according to the contour of the finger, and then the first finger joint line is found according to the first finger joint end point.
In some embodiments, after identifying the first finger joint line of the fingerprint image to be identified, the method further comprises: and adjusting the finger pointing direction in the finger head image according to the first finger joint line. Specifically, when fingerprint comparison is performed, in order to improve the accuracy and speed of the comparison, the orientations of the fingers should be as close as possible, and therefore, in this embodiment, after the first finger joint line is determined, the position of the finger is adjusted, for example, the pointing direction of the finger is toward the preset direction, and the preset direction may be, for example, directly above the image, and at this time, the pointing direction of the finger may be adjusted in a manner that the first finger joint line is adjusted to be parallel to the horizontal direction in the image.
In some embodiments, the processing step further comprises: determining the relative position relation of each finger in the fingerprint image to be identified; and determining the finger name of each finger according to the relative position relationship and the left-right hand information of the finger. The storing step further comprises: and storing the finger name corresponding to the fingerprint image. Specifically, in the present embodiment, the captured fingerprint image has a plurality of fingers, and the shapes of the left hand and the right hand are significantly different, so that it can be directly determined whether the captured fingerprint image is the left hand or the right hand, on the other hand, the fluctuation of the fingers of the left hand and the right hand are different and have a relative position relationship with each other, for example, the relative position relationship may include the relative position where the head of each finger is located, and it is determined whether each finger is specifically the index finger, the middle finger, the ring finger, or the like according to the left hand information and the right hand information and the relative position relationship. For example, when five fingers are shot, the thumb is obviously not in line with the positions of the other fingers, the finger name of each finger can be clearly determined according to the relative position of the thumb and the other fingers, whether the finger is the left hand or the right hand is judged according to the position of the thumb, and the finger name of each finger is determined according to the sequence of the rest four fingers.
In some embodiments, the determining step further comprises: and determining whether the finger is a living object according to the fingerprint image, and determining that the finger does not meet the condition if the finger is not the living object. As for the living body identification method, a method recorded in the living body identification step in any embodiment of the present application may be adopted, which is not described herein again.
For the embodiments of the apparatus, since they correspond substantially to the method embodiments, reference may be made to the partial description of the method embodiments for relevant points. The above-described apparatus embodiments are merely illustrative, in that modules illustrated as separate modules may or may not be separate. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
It should be understood that although each block in the block diagrams of the figures may represent a module, a portion of which comprises one or more executable instructions for implementing the specified logical function(s), the blocks are not necessarily executed sequentially. Each module and functional unit in the device embodiments in the present disclosure may be integrated into one processing module, or each unit may exist alone physically, or two or more modules or functional units are integrated into one module. The integrated modules can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a separate product, may also be stored in a computer readable storage medium. The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc.
[ electronic apparatus ]
Referring now to fig. 19, a schematic diagram of an electronic device (e.g., the terminal device or the server of fig. 1) 500 suitable for implementing embodiments of the present disclosure is shown. The terminal device in the embodiment of the present disclosure may be various terminal devices in the above system. The electronic device shown in fig. 19 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present disclosure.
As shown in fig. 19, the electronic device 500 may include a processing means (e.g., central processing unit, graphics processor, etc.) 501 for controlling the overall operation of the electronic device. The processing device may include one or more processors to execute instructions to perform all or a portion of the steps of the method described above. Further, the processing device 501 may also include one or more modules for processing interactions with other devices.
The sensor means 503 for sensing the prescribed measured information and converting it into a usable output signal according to a certain rule may comprise one or more sensors. For example, it may include an acceleration sensor, a gyro sensor, a magnetic sensor, a pressure sensor or a temperature sensor, etc. for detecting changes in the on/off state, relative positioning, acceleration/deceleration, temperature, humidity, light, etc. of the electronic device.
The processing means 501, the storage means 502 and the sensor means 503 are connected to each other by a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.
The multimedia device 506 may include an input device such as a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, etc. for receiving an input signal from a user, and the various input devices may cooperate with various sensors of the sensor device 503 to perform, for example, a gesture operation input, an image recognition input, a distance detection input, etc.; the multimedia device 506 may also include output devices such as a Liquid Crystal Display (LCD), speakers, vibrators, and the like.
The power supply device 507, which is used to provide power to various devices in the electronic equipment, may include a power management system, one or more power supplies, and components to distribute power to other devices.
The communication device 508 may allow the electronic apparatus 500 to communicate with other apparatuses wirelessly or by wire to exchange data.
Each of the above devices may also be connected to the I/O interface 505 to enable applications of the electronic device 500.
While fig. 19 illustrates an electronic device having various means, it is to be understood that not all illustrated means are required to be implemented or provided. More or fewer devices may alternatively be implemented or provided.
In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program carried on a non-transitory computer readable medium, the computer program containing program code for performing the method illustrated by the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network via the communication means, or may be installed from a storage means. The computer program, when executed by a processing device, performs the above-described functions defined in the methods of the embodiments of the present disclosure.
In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
It is noted that the computer readable medium described above in this disclosure can be a computer readable signal medium or a computer readable storage medium or any combination of the two. In contrast, in the present disclosure, a computer readable signal medium may comprise a propagated data signal with computer readable program code embodied therein, either in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, optical cables, RF (radio frequency), etc., or any suitable combination of the foregoing.
The computer readable medium may be embodied in the electronic device; or may exist separately without being assembled into the electronic device.
Computer program code for carrying out operations for the present disclosure may be written in any combination of one or more programming languages, including but not limited to an object oriented programming language such as Java, Smalltalk, C + +, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network or connection may be made to an external computer (for example, through the Internet using an Internet service provider).
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The units described in the embodiments of the present disclosure may be implemented by software or hardware. Where the name of an element does not in some cases constitute a limitation on the element itself.
The functions described herein above may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), Application Specific Standard Products (ASSPs), systems on a chip (SOCs), Complex Programmable Logic Devices (CPLDs), and the like.
The foregoing description is only exemplary of the preferred embodiments of the disclosure and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the disclosure herein is not limited to the particular combination of features described above, but also encompasses other embodiments in which any combination of the features described above or their equivalents does not depart from the spirit of the disclosure. For example, the above features and (but not limited to) the features disclosed in this disclosure having similar functions are replaced with each other to form the technical solution.
Further, while operations are depicted in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Likewise, while several specific implementation details are included in the above discussion, these should not be construed as limitations on the scope of the disclosure. Certain features that are described in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable subcombination.
Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.
Claims (9)
1.A non-contact fingerprint acquisition device, comprising:
a housing including a first component and a second component connected to each other in an L-shape;
the image acquisition device is arranged on the shell and used for shooting a fingerprint image, and a shooting area of the image acquisition device is positioned on one side of the first component, which is close to the second component;
and the processing device is arranged in the shell and used for identifying and processing the image shot by the image acquisition device.
2. The non-contact fingerprint acquisition device according to claim 1,
the housing further includes: a third component;
the first component, the second component and the third component are sequentially connected, the first component and the third component are oppositely arranged at intervals, and the first component, the second component and the third component surround to form an acquisition space;
the shooting area of the image acquisition device is positioned in the acquisition space.
3. The contactless fingerprint acquisition apparatus of claim 1, further comprising:
and the illuminating device is arranged in the shell and used for illuminating when the image acquisition device shoots.
4. The non-contact fingerprint acquisition device according to claim 3,
the lighting device includes: at least two lighting components;
the illumination component is used for illuminating when the image acquisition device shoots, and the colors of the light rays emitted by the at least two illumination components are different, so that the image acquisition device can shoot fingerprint images under different light ray colors.
5. The fingerprint acquisition device according to any one of claims 1 to 4, further comprising:
the light path adjusting device is arranged on the side face, close to the shooting area of the image acquisition device, of the shell, the lens of the image acquisition device faces the light path adjusting device, the light path adjusting device is used for changing the shooting direction of the image acquisition device, and the image acquisition device shoots the fingerprint image of the shooting area through the light path adjusting device.
6. The non-contact fingerprint acquisition device according to claim 5,
the optical path adjusting device includes: a reflective mirror;
the reflector is arranged on the side face, close to the shooting area of the image acquisition device, of the shell, and a preset included angle is formed between the plane where the reflector is located and the shooting direction of the image acquisition device.
7. The contactless fingerprint acquisition apparatus of claim 1, further comprising:
and the structured light projection device is arranged on the shell and used for projecting structured light when the image acquisition device shoots so as to enable the image acquisition device to shoot a structured light image.
8. The contactless fingerprint acquisition apparatus of claim 1, further comprising: a time-of-flight device;
the flight time device is used for transmitting infrared light pulses to an object to be photographed and calculating the depth information of the object to be photographed according to light signals returned by the object to be photographed;
the image acquisition device is used for focusing when the fingerprint image of the object to be photographed is shot according to the depth information.
9. The non-contact fingerprint acquisition device according to claim 1,
the image acquisition device includes: a plurality of cameras, at least two cameras of the plurality of cameras being focused at different positions;
or the image acquisition device is provided with a liquid lens;
or the image acquisition device is provided with a Frazier lens;
alternatively, the image acquisition apparatus includes: a light field camera.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202022198405.4U CN212569821U (en) | 2020-09-30 | 2020-09-30 | Non-contact fingerprint acquisition device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202022198405.4U CN212569821U (en) | 2020-09-30 | 2020-09-30 | Non-contact fingerprint acquisition device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN212569821U true CN212569821U (en) | 2021-02-19 |
Family
ID=74622257
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202022198405.4U Active CN212569821U (en) | 2020-09-30 | 2020-09-30 | Non-contact fingerprint acquisition device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN212569821U (en) |
-
2020
- 2020-09-30 CN CN202022198405.4U patent/CN212569821U/en active Active
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112232155B (en) | Non-contact fingerprint identification method and device, terminal and storage medium | |
CN112232163B (en) | Fingerprint acquisition method and device, fingerprint comparison method and device, and equipment | |
CN112016525A (en) | Non-contact fingerprint acquisition method and device | |
KR102587193B1 (en) | System and method for performing fingerprint-based user authentication using images captured using a mobile device | |
CN110326001B (en) | System and method for performing fingerprint-based user authentication using images captured with a mobile device | |
US11263432B2 (en) | Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices | |
US10339362B2 (en) | Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices | |
US20190095701A1 (en) | Living-body detection method, device and storage medium | |
CN112232159B (en) | Fingerprint identification method, device, terminal and storage medium | |
CN112232157B (en) | Fingerprint area detection method, device, equipment and storage medium | |
WO2022068931A1 (en) | Non-contact fingerprint recognition method and apparatus, terminal, and storage medium | |
CN110059579A (en) | For the method and apparatus of test alive, electronic equipment and storage medium | |
CN112232152B (en) | Non-contact fingerprint identification method and device, terminal and storage medium | |
Vasavi et al. | Novel Multimodal Biometric Feature Extraction for Precise Human Identification. | |
CN212569821U (en) | Non-contact fingerprint acquisition device | |
CN116982093A (en) | Presence attack detection | |
Dixit et al. | SIFRS: Spoof Invariant Facial Recognition System (A Helping Hand for Visual Impaired People) | |
Jadhav et al. | Information extraction for touch free biometric authentication system from large databases using deep learning algorithm | |
Nguyen | Face Recognition and Face Spoofing Detection Using 3D Model | |
CN114694265A (en) | Living body detection method, device and system | |
CN115331315A (en) | Living body detection method and living body detection device | |
Nechyporenko ОV | AUTOMATIC HUMAN RECOGNITION SYSTEMS USING BIOMETRIC COMPUTER-INTEGRATED TECHNOLOGIES |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20230816 Address after: Room 1507-1512, 13th Floor, No. 27 Zhichun Road, Haidian District, Beijing, 100083 Patentee after: Beijing jianmozi Technology Co.,Ltd. Address before: 100016 1 / F, building 5, yard 5, Jiangtai Road, Chaoyang District, Beijing Patentee before: Moqi Technology (Beijing) Co.,Ltd. |