CN109325462B - Face recognition living body detection method and device based on iris - Google Patents
Face recognition living body detection method and device based on iris Download PDFInfo
- Publication number
- CN109325462B CN109325462B CN201811183736.1A CN201811183736A CN109325462B CN 109325462 B CN109325462 B CN 109325462B CN 201811183736 A CN201811183736 A CN 201811183736A CN 109325462 B CN109325462 B CN 109325462B
- Authority
- CN
- China
- Prior art keywords
- face
- image
- iris
- database
- motion track
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/19—Sensors therefor
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/40—Spoof detection, e.g. liveness detection
- G06V40/45—Detection of the body part being alive
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/70—Multimodal biometrics, e.g. combining information from different biometric modalities
Abstract
The invention provides a face recognition living body detection method and a device based on iris, wherein the method comprises the steps of generating a random motion track; controlling the follow-up center to move along a random motion track, and shooting an iris image and a face image; comparing the collected human eye saccadic trajectory with the random motion trajectory to perform iris test, performing face search on a video image by adopting a deep learning algorithm, segmenting if a face exists, performing feature points and comparing the extracted feature values with face feature values previously stored in a local face database, and if the comparison succeeds, indicating that the image of the tested person exists in the local face database, and if the comparison fails, indicating that the image of the tested person does not exist, the invention ensures the accuracy of face recognition living body detection, does not need the tested person to perform excessive actions, improves the experience of the tested person, has high speed and accuracy in the applied algorithm, occupies small resources, can be operated on a desktop end or an ARM, and is rich in practical scene.
Description
Technical Field
The invention relates to the technical field of identity recognition, in particular to a face recognition living body detection method and device based on irises.
Background
Along with the improvement of the life quality of people, the networking of living houses, the modernization of office high-rise buildings, the increasingly developed and diversified transportation means and the like, the aspects of people are influenced all the time and moment undoubtedly, so that people worry more about the safety of big to social security and small to family residence; the demands of government departments and public places such as airports, high-speed railway stations, subway stations, customs and the like on people flow control, public security management, potential crime analysis and the like are increasing; the monitoring of the human stream and the identity in large public places such as a stadium, a football stadium or a financial center CBD and the like requires the use of a human face detection and recognition technology.
With the development of a biological recognition technology, a pattern recognition technology and an artificial intelligence technology, the technology in the field of face recognition tends to be mature from the current prior art, and a face detection algorithm in the traditional pattern recognition, such as a face Haar feature classifier and dlib face detection in Opencv; face feature extraction methods such as LBP face recognition and other algorithms can well perform face detection and recognition, but the methods cannot meet the requirements of various complex environments at present such as system miniaturization, high operation efficiency, high recognition accuracy, cheat prevention and other comprehensive factors.
The existing face recognition living body detection mainly adopts the following three schemes: 1. the user performs corresponding actions to complete the in-vivo detection, and the user needs to perform a series of actions such as shaking head, blinking, opening mouth and the like in a matching manner, so that the user experience is low; 2. the video images collected by a single visible light camera are analyzed and judged by utilizing an algorithm, and because the video images collected by the camera are in a two-dimensional space, whether the face facing the camera is a living human face or not is difficult to distinguish no matter what algorithm is used, so that the recognition accuracy is low; 3. the depth of field camera is used for carrying out 3D modeling on the tested person, and whether the tested person is a living human face is judged through the 3D model, but the depth of field camera used in the method is extremely high in cost, a large amount of computer computing capacity needs to be consumed in a 3D modeling algorithm, and the computing speed is extremely low.
Disclosure of Invention
The invention aims to provide a face recognition living body detection method and device based on an iris, and aims to solve the problems of low user experience and low recognition accuracy in face recognition of a person to be detected in the prior art.
The present invention is achieved in this way, and a first aspect of the present invention provides an iris-based face recognition live body detection method, including:
generating a random motion track;
controlling the follow-up center to move along the random motion track, and shooting an iris image and a face image;
acquiring a motion track of the pupil center of the tested target according to the iris image, and judging whether the coincidence degree of the motion track of the pupil center and the random motion track is smaller than a preset value;
when the contact ratio is smaller than a preset value, judging that the tested target is a non-living human face;
when the contact ratio is not smaller than a preset value, judging whether the face image has a face or not;
when the face image does not have a face, judging that the tested target is a non-living body face;
when the face image has a face, segmenting the face image, extracting characteristic values from characteristic points in the segmented face image, and comparing the extracted characteristic values with face characteristic values in a face database;
when the comparison is successful, judging that the face image exists in the face database;
and when the comparison fails, judging that the face image does not exist in the face database.
The invention provides a face recognition living body detection device based on iris, which comprises: the device comprises an infrared camera, a visible light camera, a follow-up target generator, a motion trail acquisition module, a living iris discrimination algorithm module, a face detection algorithm module and a face database module;
the follow-up target generator is used for generating a random motion track and controlling a follow-up center to move along the random motion track;
the infrared camera is used for shooting an iris image;
the visible light camera is used for shooting a face image;
the motion track acquisition module is used for acquiring a motion track of the pupil center point of the tested target according to the iris image;
the living body iris distinguishing algorithm module is used for judging whether the contact ratio of the motion track of the pupil center point and the random motion track is smaller than a preset value or not, and when the contact ratio is smaller than the preset value, judging that the tested target is a non-living body face;
the face detection algorithm module is used for judging whether the face image has a face or not when the living body iris distinguishing algorithm module judges that the contact ratio is not smaller than a preset value, and judging that the tested target is a non-living body face when the face image does not have the face;
the face recognition algorithm module is used for segmenting the face image when the face detection algorithm module judges that the face image has a face, extracting characteristic values from characteristic points in the segmented face image and comparing the extracted characteristic values with face characteristic values in a face database;
the face database module is used for judging that the face image exists in the face database when the face recognition algorithm module is successfully compared, and judging that the face image does not exist in the face database when the face recognition algorithm module is failed to be compared.
The invention provides a face recognition living body detection method and a device based on iris, wherein the method comprises the steps of generating a random motion track; controlling the follow-up center to move along a random motion track, and shooting an iris image and a face image; comparing the collected human eye saccadic trajectory with the random motion trajectory to perform iris test, performing face search on a video image by adopting a deep learning algorithm, segmenting if a face exists, performing feature points and comparing the extracted feature values with face feature values previously stored in a local face database, if the comparison succeeds, indicating that the image of a tested person exists in the local face database, and if the comparison fails, indicating that the image of the tested person does not exist, the technical scheme of the invention can ensure the accuracy of face recognition living body detection, simultaneously does not need the tested person to perform excessive actions, improves the experience of the tested user, and has the advantages of high speed and accuracy of the applied algorithm, small occupied resources, capability of running on a desktop end or an ARM, and richness in practical scenes.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
FIG. 1 is a flow chart of a method for detecting a living body based on iris recognition according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of an executing device of a face recognition live body detection method based on an iris according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a random movement trajectory in a face recognition live detection method based on irises according to an embodiment of the present invention;
fig. 4 is a flowchart of step S30 in a method for detecting a living body based on iris recognition according to an embodiment of the present invention;
FIG. 5 is a flowchart of a method for live detection based on iris recognition according to an embodiment of the present invention;
fig. 6 is a schematic structural diagram of a face recognition live body detection device based on an iris according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
In order to explain the technical means of the present invention, the following description will be given by way of specific examples.
The embodiment of the invention provides a face recognition living body detection method based on an iris, which comprises the following steps of:
and S10, generating a random motion track.
And S20, controlling the follow-up center to move along the random motion track, and shooting an iris image and a face image.
In steps S10 and S20, as an embodiment, as shown in fig. 2 and 3, the device for implementing the method includes a video image display area 1, a follow-up target generator 2, an infrared camera 3, an infrared fill-in lamp 4 and a visible light camera 5, a random motion track 11 is generated by the follow-up target generator 2, a follow-up center 12 capable of moving randomly is arranged in the follow-up target generator 2, because the movement of the human eyeball has a jump-look characteristic, based on the biological characteristic, a random motion track 13 can be generated by the follow-up target generator through the follow-up center, and the random motion track 13 is recorded and stored in the background, the follow-up target generator 2 generates a central light spot at the follow-up center 12 and moves according to the random motion track 13, a prompt sound is sent to prompt the testee before the follow-up center 12 moves, the eyeball of the testee is required to move along with the follow-up center 12, and simultaneously, the iris image and the face image are shot through the camera, preferably, the iris image is shot through the infrared camera, and the face image is shot through the visible light camera.
And S30, acquiring a motion track of the pupil center of the tested target according to the iris image, judging whether the coincidence degree of the motion track of the pupil center and the random motion track is smaller than a preset value, executing S40 when the coincidence degree is smaller than the preset value, and executing S50 when the coincidence degree is not smaller than the preset value.
In step S30, specifically, as an embodiment, the infrared camera synchronously collects iris video images of the eye jump of the person to be measured while the servo center moves, processes the collected iris video images, calculates the motion trajectory of the pupil center coordinate in the video images, compares the coincidence degree of the servo center motion trajectory and the motion trajectory of the pupil center coordinate, determines whether the iris is a living body by determining whether the similarity of the comparison exceeds a predetermined value, for example, the motion trajectory of the pupil center coordinate may be scaled to the same size as the motion trajectory of the servo center, sets a coordinate system for the motion trajectory of the pupil center coordinate and the motion trajectory of the servo center, obtains the coordinates of a predetermined number of points on the motion trajectory of the pupil center coordinate, and determines whether the coordinates coincide with the coordinates of points on the motion trajectory of the servo center, when the ratio of the number of the coincident coordinate points is greater than the predetermined value, for example, 80%, the iris image is determined to be derived from a living iris.
As an embodiment, as shown in fig. 4, the acquiring a motion trajectory of a pupil center point of a tested object according to the iris image in step S30 includes:
and S301, processing each frame of iris image, and acquiring the position of the pupil center of the tested target in each frame of iris image.
In step S301, the processing of each frame of iris image includes:
and positioning, normalizing and enhancing the iris in each frame of iris image.
Specifically, the iris positioning includes positioning an inner edge and an outer edge of the iris, the inner edge is a boundary of the pupil, the outer edge is a boundary of the iris and the sclera, and centers of the iris and the pupil are not overlapped in general, so that the two edges need to be positioned, and after the inner edge and the outer edge are positioned, the radius and the center of the inner circle and the outer circle can be fitted by utilizing an edge detection operator to perform normalization and image enhancement.
And S302, connecting the positions of the pupil center points to generate a motion track of the pupil center point.
In step S302, the centers of the acquired images of each frame are connected to form a motion trajectory of the pupil center.
The embodiment can accurately acquire the motion trail of the pupil center point.
It should be noted that, in the technical scheme, infrared cameras with different focal lengths may also be adopted, and a single infrared camera and two infrared cameras with different focal lengths are different in that iris images acquired by the two infrared cameras contain more information, wherein the depth of field of the iris images is richer, and the layering sense is stronger. The iris images collected by the two infrared cameras are more beneficial to further processing, wherein the positioning, normalization and image enhancement of the iris are also adopted, meanwhile, the characteristics of the iris image can be extracted from the iris, iris recognition and comparison can be carried out on a detected target, meanwhile, the double recognition of the detected target is carried out by matching with a face recognition algorithm, and the accuracy of the whole set of face recognition living body detection method based on the iris is improved.
As an implementation mode, the infrared wavelength used by the infrared camera is between 700-900 nm, the energy is extremely small, the infrared camera can only penetrate through the gaps of atomic molecules but not the interiors of the atomic molecules, the energy is lower than the international safety standard, the eye is hardly damaged, even if 10 ten thousand times of radiation is used, the radiation is smaller than the radiation of a 1-minute telephone to a human body, the illumination intensity tends to the natural light intensity, and the infrared camera can be automatically closed after the infrared LED indicator lamp is turned on for more than 9 seconds, so that the eyes of a user are protected.
When needing to be explained, the visible light camera module synchronously collects the face video image while the infrared camera collects the iris video image, so that the collected iris video image and the face video image are ensured to come from the same person.
As an implementation mode, converting an acquired color image into a gray-scale image, and denoising the image; acquiring a subregion with the lowest gray value, segmenting the image into a binary image according to the lowest gray value, roughly positioning the center position of a pupil in the binary image by utilizing all points (x k, y k) smaller than a threshold value, wherein k is 1 … N, N is the number of the points with the gray value smaller than the lowest gray value in a histogram, and setting a rectangular region taking the pupil as the center as a region of interest ROI according to the roughly determined center of the pupil; setting an initial threshold value by using a maximum inter-class variance method, then evaluating the segmented ROI image, if the light spots meeting the conditions cannot be searched in the image segmented by using the current threshold value, automatically adjusting the threshold value, re-segmenting the image, searching again, and directly searching out all the light spots meeting the conditions; determining the central coordinates of the five light spots by using a centroid method; matching the light spots with the light sources based on the geometric relationship; replacing the gray level of a point set of a light spot region by an optimal adaptive threshold value in the light spot extraction process by using a cornea reflection light spot region obtained in the light spot extraction process, then carrying out binarization on an image by adopting a classical threshold value segmentation technology, and segmenting the image according to the optimal adaptive threshold value in the light spot extraction process to obtain an approximate region of a pupil; extracting an edge point set of the pupil; and (3) obtaining the central coordinates of the pupil by ellipse fitting: firstly, carrying out ellipse fitting on candidate points by using a least square ellipse fitting method, and removing points which are too far away from the center of an ellipse in the candidate edge points; the fitting is performed in a loop until a stable ellipse center position is obtained, which is the center coordinate of the pupil.
And S40, when the contact ratio is smaller than a preset value, judging that the tested target is a non-living human face.
And S50, when the contact ratio is not smaller than a preset value, judging whether the face image has a face, when the face image does not have the face, executing a step S40, and when the face image has the face, executing a step S70.
And S60, when the face image has a face, segmenting the face image, extracting characteristic values from characteristic points in the segmented face image, and comparing the extracted characteristic values with face characteristic values in a face database.
In step S60, different databases of the face database store the face image feature value, the path of the face image, and the face image feature value, the path of the face image, and the face image correspond to each other.
And S70, judging that the face image exists in the face database when the comparison is successful.
And S80, when the comparison fails, judging that the face image does not exist in the face database.
Further, in step S50, the determining whether the face image has a face includes:
and judging whether the face image has a face or not through an MTCNN face detection algorithm.
Specifically, after receiving the trigger condition of the living iris discrimination algorithm, the MTCNN face detection algorithm is used to start face search on the video image acquired by the visible light camera module, and if no face is found in the acquired video image, it indicates that the target does not exist, and the detection is finished. The deep learning algorithm MTCNN face detection algorithm is mainly divided into three steps: firstly, multi-scale transformation is carried out on an image, an image pyramid is obtained, multi-scale information of the image is obtained, and then face detection is achieved.
Further, in step S60, the extracting feature values from the feature points in the segmented face image includes:
and extracting characteristic values from characteristic points in the segmented face image through a LightenedCNN face recognition algorithm.
Specifically, after a triggering condition of a face detection algorithm is received, a LightenedCNN face recognition algorithm is adopted to extract face characteristic points of each frame of segmented video image, the extracted characteristic values are compared with face characteristic values stored in a local face database, and the system considers that the value after comparison calculation is lower than 80 points as failure, namely, no visible light camera exists in the local face database to acquire a person in the video image, and the value is equal to or higher than 80 points as success, namely, the visible light camera exists in the local face database to acquire a person in the video image.
The face verification by LightenedCNN is divided into three types: one is to train CNN to extract features using the task of face classification and then judge whether it is the same person or not with a classifier. The second is to directly optimize the verification loss. And the third is to carry out face recognition and verification tasks simultaneously, the Lightened CNN face recognition algorithm is a light CNN, and the light CNN has better effect, simplified network structure, optimized time and space and can be operated on embedded equipment and mobile equipment.
According to the embodiment of the invention, the motion of the follow-up center along the random motion track is controlled, and the infrared camera and the visible light camera system are adopted to respectively acquire video images of the iris and the face; according to the characteristic that human eyes can move along with a follow-up center in a follow-up target generator, comparing a human eye saccadic track acquired by an infrared camera with a random motion track to carry out iris test, carrying out face search on a video image by adopting a deep learning algorithm after the iris test is finished, segmenting if the face exists, and carrying out processing such as feature point extraction on the segmented face; the extracted characteristic value is compared with the characteristic value of the face stored in the local face database, if the comparison is successful, the fact that the person in the video image acquired by the visible light camera exists in the local face database is shown, and if the comparison is failed, the fact that the person in the video image acquired by the visible light camera does not exist in the local face data is shown, the embodiment of the invention can ensure the accuracy of face recognition living body detection when being applied to different scenes.
Further, as another embodiment, the iris image may be recognized, as shown in fig. 5, the capturing the iris image and the face image in step S10 further includes:
and S101, performing windowing processing on the iris image, and performing wavelet packet decomposition on each window to obtain a sub-band image.
And S102, screening the sub-band images to obtain sub-band images in the iris feature set.
And S103, carrying out singular value decomposition on the sub-band image in the iris feature set to obtain a feature vector.
And S104, comparing the characteristic vector with the iris characteristic vector in the iris database.
And S105, when the comparison is successful, judging that the iris image exists in the iris database.
And S106, when the comparison fails, judging that the iris image does not exist in the iris database.
Specifically, in the above steps S101 to S106, the iris recognition and comparison for the iris of the target to be detected is to extract the features of the iris image, the window division processing is performed on the iris image, wavelet packet decomposition is performed for each window, on this basis, the subband image of each window is subjected to the screening processing, then singular value decomposition is performed for the subband image in the iris feature set to extract the features, finally, the decomposed feature vector is compressed according to the properties of the singular value features to be used as the final recognition feature vector, and then, the comparison recognition is performed according to the feature vector to determine whether the subband image is in the iris database.
Further, as another embodiment, the comparing the extracted feature value with the face feature value in the face database previously includes:
after adding and deleting the face images into the face database, loading the modified face database file in an internal memory;
the comparing the extracted feature value with the face feature value in the face database further comprises:
and comparing the extracted characteristic value with the face characteristic value in the face database in the memory.
Specifically, the face database module is a local database, face characteristic values stored in the database are added and deleted through a specific reading and writing interface, a face database file is loaded into a memory at the beginning of equipment operation, and when a face recognition algorithm compares the face characteristic values, the data loaded into the memory is directly compared, but not the face database file per se; when the face is added to the face database by calling the interface, the data in the memory is modified firstly, the local database file is updated when the equipment is closed to operate, and meanwhile, the latest local database file is loaded into the memory when the equipment operates again next time.
Another embodiment of the present invention provides an iris-based face recognition live body detection apparatus, as shown in fig. 6, the iris-based face recognition live body detection apparatus including: the system comprises an infrared camera 401, a visible light camera 402, a follow-up target generator 403, a motion track acquisition module 404, a living iris discrimination algorithm module 405, a face detection algorithm module 406, a face recognition algorithm module 407 and a face database module 408;
the follow-up target generator 406 is used for generating a random motion track and controlling a follow-up center to move along the random motion track;
the infrared camera 401 is used for shooting an iris image;
the visible light camera 402 is used for shooting a face image;
the motion track acquiring module 404 is configured to acquire a motion track of a pupil center of the target to be tested according to the iris image;
the living iris identification algorithm module 405 is configured to determine whether a coincidence degree of the motion trajectory of the pupil center point and the random motion trajectory is smaller than a preset value, and when the coincidence degree is smaller than the preset value, determine that the tested target is a non-living human face;
the face detection algorithm module 406 is configured to determine whether a face exists in the face image when the living iris discrimination algorithm module determines that the contact ratio is not smaller than a preset value, determine that the target to be tested is a non-living face when the face image does not have a face, and segment the face image when the face detection algorithm module 406 determines that the face exists in the face image;
the face recognition algorithm module 407 is configured to extract feature values from feature points in the segmented face image, and compare the extracted feature values with face feature values in a face database;
the face database module 408 is configured to determine that the face image exists in the face database when the comparison of the face recognition algorithm module 407 is successful, and determine that the face image does not exist in the face database when the comparison of the face recognition algorithm module is failed.
The motion trajectory acquisition module 404 is specifically configured to:
processing each frame of iris image, and acquiring the position of the pupil center point of the tested target in each frame of iris image;
and connecting the positions of the pupil center points to generate a motion trail of the pupil center point.
The motion trail obtaining module 404 is further configured to locate, normalize, and enhance the iris in each frame of iris image.
The technical scheme of the invention provides a face recognition living body detection method and a face recognition living body detection device based on an iris, which can solve the problem that a recorded video or a shot picture cannot be recognized for cheating when a single camera carries out face recognition at present, and the recognition speed is higher than that of the existing multi-angle multi-camera 3D modeling living body recognition scheme; the algorithm applied by the system is high in speed and accuracy, and simultaneously, the occupied resources are very small, so that the system can be operated at a desktop end and can also be operated on an ARM (advanced RISC machine), and therefore, the practical applicable scenes are very rich; the human face detection method and the system have the advantages that the visible light camera synchronously acquires the human face video images when the human eyes are close to the specified infrared camera acquisition area, and the human face detection algorithm analyzes the human face video images acquired by the visible light camera when the living iris discrimination algorithm determines that the eyeballs have saccadic vision, so that the video images acquired by the infrared camera and the visible light camera are from the same person, and the accuracy of human face identification living body detection can be guaranteed while the scheme is applied to different scenes.
The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; such modifications and substitutions do not substantially depart from the spirit and scope of the embodiments of the present invention, and are intended to be included within the scope of the present invention.
Claims (10)
1. An iris-based face recognition live body detection method is characterized by comprising the following steps:
generating a random motion track;
controlling the follow-up center to move along the random motion track, shooting an iris image through an infrared camera, and shooting a face image through a visible light camera;
acquiring a motion track of the pupil center of the tested target according to the iris image, and judging whether the coincidence degree of the motion track of the pupil center and the random motion track is smaller than a preset value;
when the contact ratio is smaller than a preset value, judging that the tested target is a non-living human face;
when the contact ratio is not smaller than a preset value, judging whether the face image has a face or not;
when the face image does not have a face, judging that the tested target is a non-living body face;
when the face image has a face, segmenting the face image, extracting characteristic values from characteristic points in the segmented face image, and comparing the extracted characteristic values with face characteristic values in a face database;
when the comparison is successful, judging that the face image exists in the face database;
and when the comparison fails, judging that the face image does not exist in the face database.
2. The living body detection method for face recognition according to claim 1, wherein the obtaining of the motion trajectory of the pupil center of the tested object according to the iris image comprises:
processing each frame of iris image, and acquiring the position of the pupil center point of the tested target in each frame of iris image;
and connecting the positions of the pupil center points to generate a motion trail of the pupil center point.
3. The face recognition live body detection method according to claim 2, wherein the processing of each frame of iris image comprises:
and positioning, normalizing and enhancing the iris in each frame of iris image.
4. The face recognition live body detection method according to claim 1, wherein the taking of the iris image and the face image further comprises:
performing windowing processing on the iris image, and performing wavelet packet decomposition on each window to obtain a sub-band image;
screening the sub-band images to obtain sub-band images in an iris feature set;
performing singular value decomposition on the sub-band image in the iris feature set to obtain a feature vector;
comparing the feature vector with iris feature vectors in an iris database;
when the comparison is successful, judging that the iris image exists in the iris database;
and when the comparison fails, judging that the iris image does not exist in the iris database.
5. The living body detection method for face recognition according to claim 1, wherein the determining whether the face image has a face comprises:
and judging whether the face image has a face or not through an MTCNN face detection algorithm.
6. The face recognition live body detection method according to claim 1, wherein the extracting feature values from feature points in the segmented face image comprises:
and extracting characteristic values from characteristic points in the segmented face image through a LightenedCNN face recognition algorithm.
7. The face recognition live body detection method according to claim 1, wherein the comparing the extracted feature value with the face feature value in the face database further comprises:
after adding and deleting the face images into the face database, loading the modified face database file in an internal memory;
the comparing the extracted feature value with the face feature value in the face database further comprises:
and comparing the extracted characteristic value with the face characteristic value in the face database in the memory.
8. An iris-based face recognition live body detection apparatus, characterized by comprising: the device comprises an infrared camera, a visible light camera, a follow-up target generator, a motion trail acquisition module, a living iris discrimination algorithm module, a face detection algorithm module and a face database module;
the follow-up target generator is used for generating a random motion track and controlling a follow-up center to move along the random motion track;
the infrared camera is used for shooting an iris image;
the visible light camera is used for shooting a face image;
the motion track acquisition module is used for acquiring a motion track of the pupil center point of the tested target according to the iris image;
the living body iris distinguishing algorithm module is used for judging whether the contact ratio of the motion track of the pupil center point and the random motion track is smaller than a preset value or not, and when the contact ratio is smaller than the preset value, judging that the tested target is a non-living body face;
the face detection algorithm module is used for judging whether the face image has a face or not when the living body iris distinguishing algorithm module judges that the contact ratio is not smaller than a preset value, judging that the tested target is a non-living body face when the face image does not have the face, and segmenting the face image when the face detection algorithm module judges that the face image has the face;
the face recognition algorithm module is used for extracting characteristic values from characteristic points in the segmented face image and comparing the extracted characteristic values with face characteristic values in a face database;
the face database module is used for judging that the face image exists in the face database when the face recognition algorithm module is successfully compared, and judging that the face image does not exist in the face database when the face recognition algorithm module is failed to be compared.
9. The face recognition live body detection device of claim 8, wherein the motion trajectory acquisition module is specifically configured to:
processing each frame of iris image, and acquiring the position of the pupil center point of the tested target in each frame of iris image;
and connecting the positions of the pupil center points to generate a motion trail of the pupil center point.
10. The living body detecting device for human face recognition as claimed in claim 9, wherein the motion track obtaining module is further configured to locate, normalize and enhance the iris in each frame of iris image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811183736.1A CN109325462B (en) | 2018-10-11 | 2018-10-11 | Face recognition living body detection method and device based on iris |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811183736.1A CN109325462B (en) | 2018-10-11 | 2018-10-11 | Face recognition living body detection method and device based on iris |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109325462A CN109325462A (en) | 2019-02-12 |
| CN109325462B true CN109325462B (en) | 2021-03-12 |
Family
ID=65261284
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811183736.1A Active CN109325462B (en) | 2018-10-11 | 2018-10-11 | Face recognition living body detection method and device based on iris |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109325462B (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110532933A (en) * | 2019-08-26 | 2019-12-03 | 淮北师范大学 | A kind of living body faces detection head pose returns the acquisition methods and detection method of device |
| CN113255400A (en) * | 2020-02-10 | 2021-08-13 | 深圳市光鉴科技有限公司 | Training and recognition method, system, equipment and medium of living body face recognition model |
| CN111402480A (en) * | 2020-02-29 | 2020-07-10 | 深圳壹账通智能科技有限公司 | Visitor information management method, device, system, equipment and storage medium |
| CN111797677B (en) * | 2020-05-13 | 2024-04-02 | 南京中科道置智能科技有限公司 | Face recognition living body detection method based on face iris recognition and thermal imaging technology |
| CN112164169A (en) * | 2020-09-21 | 2021-01-01 | 深圳前海微众银行股份有限公司 | Access control management method and device, access control equipment and computer readable storage medium |
| CN112287909B (en) * | 2020-12-24 | 2021-09-07 | 四川新网银行股份有限公司 | Double-random in-vivo detection method for randomly generating detection points and interactive elements |
| CN114005019B (en) * | 2021-10-29 | 2023-09-22 | 北京有竹居网络技术有限公司 | Method for identifying flip image and related equipment thereof |
| CN116994325B (en) * | 2023-07-27 | 2024-02-20 | 山东睿芯半导体科技有限公司 | Iris recognition method, chip and terminal |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6954553B2 (en) * | 2000-10-20 | 2005-10-11 | Fujitsu Limited | Personal authentication system using fingerprint information, registration-and-authentication method for the system, determination method for the system, and computer-readable recording medium |
| CN101483652A (en) * | 2009-01-10 | 2009-07-15 | 五邑大学 | Living creature characteristic recognition system |
| JP2010136223A (en) * | 2008-12-05 | 2010-06-17 | Sony Corp | Imaging device and imaging method |
| CN101833646A (en) * | 2009-03-11 | 2010-09-15 | 中国科学院自动化研究所 | Living iris detection method |
| CN104992085A (en) * | 2015-06-13 | 2015-10-21 | 东莞市微模式软件有限公司 | Method and device for human body in-vivo detection based on touch trace tracking |
| CN105184277A (en) * | 2015-09-29 | 2015-12-23 | 杨晴虹 | Living body human face recognition method and device |
| CN105956536A (en) * | 2016-04-26 | 2016-09-21 | 北京森科赛德科技有限公司 | Pretreatment method and device for iris recognition |
| CN106203372A (en) * | 2016-07-19 | 2016-12-07 | 奇酷互联网络科技（深圳）有限公司 | Biopsy methods based on eyes, device and terminal unit |
| CN106339931A (en) * | 2015-12-02 | 2017-01-18 | 广西职业技术学院 | Multifunctional taxation service system |
| CN107463883A (en) * | 2017-07-18 | 2017-12-12 | 广东欧珀移动通信有限公司 | Biometric discrimination method and Related product |
| CN107506721A (en) * | 2017-08-18 | 2017-12-22 | 上海爱优威软件开发有限公司 | A kind of more method for collecting iris and system |
| WO2018164364A1 (en) * | 2017-03-10 | 2018-09-13 | 크루셜텍 (주) | Contactless multiple body part recognition method and multiple body part recognition device, using multiple biometric data |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110157347A1 (en) * | 2009-12-31 | 2011-06-30 | Peter Kalocsai | Unintrusive biometric capture device, system and method for logical access control |
| CN107451547A (en) * | 2017-07-17 | 2017-12-08 | 广东欧珀移动通信有限公司 | Identify the method and Related product of live body |
-
2018
- 2018-10-11 CN CN201811183736.1A patent/CN109325462B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6954553B2 (en) * | 2000-10-20 | 2005-10-11 | Fujitsu Limited | Personal authentication system using fingerprint information, registration-and-authentication method for the system, determination method for the system, and computer-readable recording medium |
| JP2010136223A (en) * | 2008-12-05 | 2010-06-17 | Sony Corp | Imaging device and imaging method |
| CN101483652A (en) * | 2009-01-10 | 2009-07-15 | 五邑大学 | Living creature characteristic recognition system |
| CN101833646A (en) * | 2009-03-11 | 2010-09-15 | 中国科学院自动化研究所 | Living iris detection method |
| CN104992085A (en) * | 2015-06-13 | 2015-10-21 | 东莞市微模式软件有限公司 | Method and device for human body in-vivo detection based on touch trace tracking |
| CN105184277A (en) * | 2015-09-29 | 2015-12-23 | 杨晴虹 | Living body human face recognition method and device |
| CN106339931A (en) * | 2015-12-02 | 2017-01-18 | 广西职业技术学院 | Multifunctional taxation service system |
| CN105956536A (en) * | 2016-04-26 | 2016-09-21 | 北京森科赛德科技有限公司 | Pretreatment method and device for iris recognition |
| CN106203372A (en) * | 2016-07-19 | 2016-12-07 | 奇酷互联网络科技（深圳）有限公司 | Biopsy methods based on eyes, device and terminal unit |
| WO2018164364A1 (en) * | 2017-03-10 | 2018-09-13 | 크루셜텍 (주) | Contactless multiple body part recognition method and multiple body part recognition device, using multiple biometric data |
| CN107463883A (en) * | 2017-07-18 | 2017-12-12 | 广东欧珀移动通信有限公司 | Biometric discrimination method and Related product |
| CN107506721A (en) * | 2017-08-18 | 2017-12-22 | 上海爱优威软件开发有限公司 | A kind of more method for collecting iris and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109325462A (en) | 2019-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109325462B (en) | Face recognition living body detection method and device based on iris | |
| Shao et al. | Deep convolutional dynamic texture learning with adaptive channel-discriminability for 3D mask face anti-spoofing | |
| CN109558810B (en) | Target person identification method based on part segmentation and fusion | |
| US7620216B2 (en) | Method of tracking a human eye in a video image | |
| JP4743823B2 (en) | Image processing apparatus, imaging apparatus, and image processing method | |
| Abiyev et al. | Personal iris recognition using neural network | |
| Boehnen et al. | A fast multi-modal approach to facial feature detection | |
| KR102554391B1 (en) | Iris recognition based user authentication apparatus and method thereof | |
| CN107590473B (en) | Human face living body detection method, medium and related device | |
| CN109858375A (en) | Living body faces detection method, terminal and computer readable storage medium | |
| Woodard et al. | Appearance-based periocular features in the context of face and non-ideal iris recognition | |
| Lee et al. | An automated video-based system for iris recognition | |
| CN110069965A (en) | A kind of robot personal identification method | |
| Das et al. | Fuzzy logic based selera recognition | |
| KR20210075886A (en) | Image-based facial expression emotion recognition system using dual deep network and method thereof | |
| CN111860394A (en) | Gesture estimation and gesture detection-based action living body recognition method | |
| CN111862413A (en) | Method and system for realizing epidemic situation resistant non-contact multidimensional identity rapid identification | |
| CN107862298B (en) | Winking living body detection method based on infrared camera device | |
| CN110929570A (en) | Iris rapid positioning device and positioning method thereof | |
| Abiyev et al. | Neural network based biometric personal identification with fast iris segmentation | |
| CN113920591A (en) | Medium-distance and long-distance identity authentication method and device based on multi-mode biological feature recognition | |
| Yuan et al. | Ear detection based on CenterNet | |
| KR20120049605A (en) | Apparatus and method for detecting center of pupil | |
| CN108197593B (en) | Multi-size facial expression recognition method and device based on three-point positioning method | |
| Nigam et al. | Iris classification based on its quality |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |