CN111985303A - Human face recognition and human eye light spot living body detection device and method - Google Patents
Human face recognition and human eye light spot living body detection device and method Download PDFInfo
- Publication number
- CN111985303A CN111985303A CN202010625586.6A CN202010625586A CN111985303A CN 111985303 A CN111985303 A CN 111985303A CN 202010625586 A CN202010625586 A CN 202010625586A CN 111985303 A CN111985303 A CN 111985303A
- Authority
- CN
- China
- Prior art keywords
- image
- processing board
- human
- dsp processing
- infrared light
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000001514 detection method Methods 0.000 title claims abstract description 25
- 210000001508 eye Anatomy 0.000 claims abstract description 142
- 238000012545 processing Methods 0.000 claims abstract description 92
- 210000005252 bulbus oculi Anatomy 0.000 claims abstract description 43
- 239000004973 liquid crystal related substance Substances 0.000 claims description 17
- 238000005516 engineering process Methods 0.000 claims description 12
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 230000004424 eye movement Effects 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000001727 in vivo Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 claims description 4
- 230000001815 facial effect Effects 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 2
- 230000001131 transforming effect Effects 0.000 claims description 2
- 230000007547 defect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 210000003128 head Anatomy 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 210000000887 face Anatomy 0.000 description 3
- 239000011324 bead Substances 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000001678 irradiating effect Effects 0.000 description 2
- 210000000697 sensory organ Anatomy 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 208000008589 Obesity Diseases 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 210000004709 eyebrow Anatomy 0.000 description 1
- 230000003760 hair shine Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 235000020824 obesity Nutrition 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 210000001747 pupil Anatomy 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 210000001525 retina Anatomy 0.000 description 1
- 238000011426 transformation method Methods 0.000 description 1
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- 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
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- 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
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
-
- 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
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/18—Eye characteristics, e.g. of the iris
- G06V40/193—Preprocessing; Feature extraction
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Ophthalmology & Optometry (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a human face recognition and human eye facula living body detection device and a method, images collected by two wide-angle cameras are fused in a DSP processing board, a human eye area is extracted from a recognized human face image, the position coordinate of a human eye is calculated, the DSP processing board adjusts the rotation angle of an adjustable camera according to the position coordinate of the human eye, the adjustable camera aims at eyeballs of the human eye to collect images, then the quantity of light beams of infrared light emitted by an infrared light source array is controlled, infrared light beams emitted by the infrared light source array irradiate the eyeballs to generate facula, the adjustable camera aiming at the eye collects facula images on the eyeballs and sends the facula images to the DSP processing board to carry out facula image recognition, the quantity of the facula images on the eyeballs is calculated, when the quantity of the facula images recognized in the DSP processing board is the same as the quantity of the infrared light beams emitted by the infrared light source, it can be judged whether the current face of the user belongs to the living body.
Description
Technical Field
The invention relates to the field of image processing, in particular to a human face recognition and human eye light spot living body detection device and method.
Background
With the development of scientific technology, safe and effective identity authentication has become an urgent need. The biometric-based identity authentication technology is more and more concerned by people due to the reliability thereof, such as an access control system based on face or fingerprint identification, a face or fingerprint payment system, and the like, and the fingerprint identification has copied traces, which easily causes information exposure, so that the face identification gradually becomes a research hotspot in security neighborhoods with the advantages of non-contact, concealment, non-enforcement and low cost.
The accuracy of face recognition is not high, and particularly when order payment is completed, a mobile phone number needs to be input for double authentication. The mobile phone number is public and is easy to know; the convenience is affected by inputting a complicated mobile phone number. On the other hand, facial features are changed due to face-lifting, obesity, aging, etc., and thus recognition is likely to fail and no payment is made. For another example, when there are many customers who are in line for checkout, the face device camera is easy to capture multiple faces at the same time due to a large field angle range, and facial images of other customers may be used as recognition targets. These pose a risk to payment events that have high requirements for security due to face recognition. Moreover, some lawbreakers can deceive the human face acquisition system by using the false face, and the false face is easily recognized as the real face by the system, wherein one main reason is that the outline of the false face is the same as that of the real face, but the human eye facula of the false face cannot be changed, and the human eye facula of the real face can be changed according to different illumination. The account is single only by logging on the face, information leakage and misjudgment are easily caused, the login of the account is realized by combining a plurality of biological characteristics, and unnecessary loss caused by the fact that a fake face is adopted by other people to replace a real face can be effectively prevented.
The existing face recognition system mostly adopts a single camera, the single camera collects images of a two-dimensional plane, but eye movement tracking under the two-dimensional plane needs to keep the positions of the camera and the lens unchanged and the distance is short.
Disclosure of Invention
In order to solve the above problems, the present invention provides a device and a method for detecting human eyes based on human face recognition and human eyes light spot living body, which adopts binocular stereo vision to collect human face images, utilizes super resolution technology to improve the resolution of the human face images, and the binocular stereo vision can provide a basis for determining three-dimensional space position coordinates, overcomes the defects that under a two-dimensional plane, eye movement tracking needs to keep the positions of a head and a lens unchanged and the distance is short, because the color and geometric characteristics of a human eye region and other five sense organ regions have great differences, uses an integral projection method to extract human eyes images, extracts human eyes gray images through the distribution condition of gray curves, controls an infrared light source array to irradiate human eyes to generate light spots on eyeballs, aims at the positions of the human eyes through an adjustable camera to collect the light spots, extracts the light spots in the eyeballs through a Hough transformation method, and obtaining the number of light spots on the eyeball, and when the number of the identified light spots is the same as the number of the light beams emitted by the infrared light source array, indicating that the face of the user is a living face.
In order to achieve the above object, the present invention provides a living body detection device and method based on face recognition and human eye light spots, which is realized as follows:
a living body detection device and method based on face recognition and human eye facula comprises a shell, a liquid crystal display screen, an infrared light source array, an adjustable camera, an image acquisition card, a DSP processing board, an OBD interface and a wide-angle camera, wherein the liquid crystal display screen is embedded on the lower surface of the shell and used for displaying a face image of a user, account information can be input on the liquid crystal display screen by the user, the infrared light source array, the adjustable camera, the wide-angle camera and the OBD interface are embedded on the upper surface of the shell, the image acquisition card and the DSP processing board are installed in the shell, wherein the two same wide-angle cameras are adopted for acquiring the face image of the user, images acquired by the two wide-angle cameras are transmitted to the image acquisition card for processing and then transmitted to the DSP processing board, the image processing is carried out in the DSP processing board, and the images acquired by the two same wide-angle cameras, the method has the advantages that the effect of acquiring images through binocular stereo vision is achieved, the position coordinates of the three-dimensional space where the eyes are located are calculated in the DSP processing board, infrared light beams are emitted to the eyes by the infrared light source array, light spots are generated on the eyeballs, the DSP processing board adjusts the position of the adjustable camera according to the position coordinates where the eyes are located, the high-definition camera is aligned to the positions of the eyes, light spot images on the eyeballs are acquired, the acquired image information is transmitted to the image acquisition card and then transmitted to the DSP processing board, light spot image identification is carried out in the DSP processing board, the DSP processing board sends the identified face image information and light spot image information to the OBD interface, the acquired face image is sent to the liquid crystal display screen to be displayed, and a user can receive user registration information and face images and light spot image information corresponding to the user registration information at the OBD interface.
The adjustable camera comprises a universal speed reducing motor and a high-definition camera, wherein the high-definition camera is installed on a rotating shaft of the universal speed reducing motor, and the rotation direction of the universal speed reducing motor is controlled by a DSP processing board, so that the rotation angle of the high-definition camera is adjusted, the high-definition camera can aim at eyeballs of human eyes for collection, and light spots on the eyeballs are collected.
The method for detecting the light spot living body through the face recognition and the human eye comprises the following steps: firstly, carrying out image acquisition on the face of a user through two identical wide-angle cameras, sending the acquired images to a DSP processing board for face recognition, wherein the two wide-angle cameras are adopted for simultaneously acquiring the face, so that the images acquired by the two wide-angle cameras need to be fused in the DSP processing board, a human eye area is extracted from the recognized face image, the position coordinates of human eyes are calculated in the DSP processing board, the DSP processing board adjusts the rotating angle of the adjustable cameras according to the position coordinates of the human eyes, so that the adjustable cameras aim at the eyeballs of the human eyes for image acquisition, then the DSP processing board controls the quantity of light beams of infrared light emitted by an infrared light source array, infrared light beams emitted by the infrared light source array irradiate the eyeballs to generate light spots, the adjustable cameras aiming at the eyes acquire light spot images on the eyeballs and send the light spot images to the DSP processing board for spot image recognition, and calculating the number of light spots on the eyeball, and judging that the current face of the user belongs to a living body when the number of light spot images identified in the DSP processing board is the same as the number of infrared light beams emitted by the infrared light source array.
The DSP processing board of the invention carries out gray processing on the face image collected by the wide-angle camera to reduce the calculation complexity, then carries out noise reduction processing on the image by adopting a two-dimensional median filtering algorithm to eliminate the influence of noise on the image quality, establishes the relation between a plurality of images and a single image by image registration and V system image fusion technology, obtains a final super-resolution image by a sub-pixel convolution neural network algorithm, namely reconstructs a high-resolution image in the collected low-resolution face image, requires that the distance between eyes and the camera cannot be too far in the process of eye movement tracking, otherwise reduces the pixel value of the eye part to influence the extraction of key point characteristics of the eyes, so that carries out pre-processing on the collected image by adopting the super-resolution technology and obtains three-dimensional geometrical information of the face by adopting a binocular technology to calculate the three-dimensional space coordinates of the eyes, because the color geometric characteristics and the like of the human eye region and other facial features are greatly different, human eyes are extracted by a pixel integral projection method, namely, a gray level image of the human eyes is extracted by adopting the distribution condition of a gray level curve, and then the human eye region is positioned, wherein the human eye positioning calculation method comprises the following steps:
setting a face picture with M multiplied by N pixels, locating the eye region, I (x, y) represents the pixel value at the image point (x, y), x1，x2Is a value in the horizontal direction and is defined as x1,x2]Integral projection function H [ x ] of1,x2]Wherein x is1,x2∈[0,M]。
Define y in the same way1，y2Is a value in the vertical direction ofy1,y2]Integral projection function V [ y ] of1,y2]Wherein y is1,y2∈[0,N].
The integral projection method obtained by the formulas (1) and (2) is obtained by simply adding corresponding element values, the horizontal and vertical coordinate positions of human eyes are determined by utilizing the maximum value of edge projection of the top half of the human face image in the vertical and horizontal directions so as to position the positions of the human eyes, the actual pixel value of each pixel point is replaced by utilizing the mean value of the pixel values of the neighborhood of each pixel point in a small area, and then integral operation is carried out, so that the influence of noise can be effectively overcome.
The scheme for realizing the human eye light spot living body detection comprises the following steps: the DSP processing board controls the quantity of light beams of infrared light emitted by the infrared light source array, the emitted infrared light irradiates eyeballs of human eyes to form light spots on the eyeballs, the quantity of the light spots is the quantity of the infrared light beams irradiating the eyeballs, the DSP processing board adjusts the rotation angle of the adjustable camera according to the identified position coordinates of the human eyes to enable the adjustable camera to aim at the human eyes to shoot, and sends the collected light spot images to the DSP processing board to identify the light spot images, the light spot images are firstly grayed to be converted into gray level images, the system calculation amount is reduced, then the gray level images are filtered by a two-dimensional median filtering algorithm, then the Hoff transformation is adopted to detect the light spots of the human eyes, a rectangular coordinate system is transformed into a parameter space, in order to find out a straight line in a rectangular coordinate plane, the number of collinear coordinates needs to be determined, if the number of the collinear coordinates exceeds a threshold value, a certain straight line can, then determining the number of collinear coordinates, firstly dividing the polar coordinate space (rho, theta) into a certain number of counting grids, storing a counting variable in the counting grids, calculating theta by traversing each point (x, y) on the rectangular coordinate plane and rotating the point (x, y) for one circle, putting the point (x, y) values which are calculated to be the same rho into the grids which can be counted, and then counting the numberThe grids are increased by one bit, after the traversal of all the points (x, y) is completed, the counters in all the counting grids of the polar coordinate space (rho, theta) are sorted from large to small, the collinear point of the polar coordinate space (rho, theta) is arranged in the largest counting grid, and finally the collinear point is mapped back to the rectangular coordinate plane, so that the detection process of a straight line is completed, and in the rectangular coordinate system of the plane, the equation of a circle is (x-a)2+(y-b)2＝r2According to the transformation of the linear equation, the (X, Y) space of the circle can be transformed into the parameter space (r, θ), and then the rectangular coordinate system equation can become:
x＝a+rcos(θ) (3)
y＝b+rsin(θ) (4)
then, the center of the circle is set to be (a) at any point on the detected imagei,bi) Then, with the radius r known, rotating by 2 π, the coordinates (x) of each point on the circle are obtainedi,yi) And similarly, when the coordinates of the circle center are reversely deduced under the condition of knowing each point on the circle, under the condition of traversing each point, the counting variable value obtained by the circle center point (a) and the circle center point (b) is the largest, so that the image and the number of the light spots are identified, and when the number of the identified light spots is the same as the number of the light beams of the infrared light emitted by the infrared light source array, the light spots of the human eyes are judged to be living bodies, otherwise, the light spots are false bodies.
Because the invention adopts the human eye facula living body detection to judge whether the human face of the user is the living body structure, the following beneficial effects can be obtained:
the super-resolution technology is utilized to improve the resolution of the face image, the binocular stereo vision technology is adopted to obtain the three-dimensional geometric information of the face, a foundation is provided for determining the position coordinates of the three-dimensional space, and the defects that the positions of a head and a lens are required to be kept unchanged and the distance is short in the eye movement tracking under a two-dimensional plane are overcome.
The method comprises the steps of extracting human eye images by using an integral projection method, extracting the human eye gray level images through the distribution condition of a gray level curve, controlling an infrared light source array to irradiate human eyes to generate light spots on eyeballs, carrying out light spot acquisition on the positions of the human eyes through an adjustable camera, extracting the light spots in the eyeballs through a Hough transform method, obtaining the number of the light spots on the eyeballs, and when the number of the identified light spots is the same as the number of light beams emitted by the infrared light source array, indicating that the human face of a user is a living human face, and effectively preventing other people from adopting false faces to replace real faces to cause unnecessary loss.
Drawings
FIG. 1 is a schematic structural diagram of a human face recognition and human eye light spot living body detection device and method according to the present invention;
FIG. 2 is a schematic structural diagram of an adjustable camera based on a human face recognition and human eye light spot living body detection device and method of the present invention;
FIG. 3 is a flow chart of a method for face recognition and eye spot in vivo detection based on the face recognition and eye spot in vivo detection device and method of the present invention;
FIG. 4 is a flow chart of a human face recognition and human eye positioning scheme based on a human face recognition and human eye light spot living body detection device and method of the present invention;
FIG. 5 is a schematic diagram of human eye light spot in vivo detection based on human face recognition and human eye light spot in vivo detection apparatus and method of the present invention;
fig. 6 is a working principle diagram of a human face recognition and human eye light spot living body detection device and method.
The main elements are indicated by symbols.
| |
1 | Liquid |
2 |
| Infrared |
3 | |
4 |
| |
5 | DSP processing board | 6 |
| |
7 | Wide- |
8 |
| Universal |
9 | High- |
10 |
Detailed Description
The present invention will be described in further detail with reference to the following examples and drawings.
Referring to fig. 1 to 6, a living body detection device and method based on face recognition and human eye facula in the invention is shown, which includes a housing 1, a liquid crystal display 2, an infrared light source array 3, an adjustable camera 4, an image acquisition card 5, a DSP processing board 6, an OBD interface 7, and a wide-angle camera 8.
As shown in fig. 1, the DSP processing board 6 is electrically connected to the liquid crystal display 2, the infrared light source array 3, the adjustable camera 4, the image acquisition card 5, the OBD interface 7, and the wide-angle camera 8. The liquid crystal display screen 2 is embedded on the lower surface of the shell 1 and is used for displaying a face image of a user and facilitating the viewing of the user, the user can input account information on the liquid crystal display screen 2 and conveniently correspond the account information and the face one to one, the infrared light source array 3, the adjustable camera 4, the wide-angle camera 8 and the OBD interface 7 are embedded on the upper surface of the shell 1, the image acquisition card 5 and the DSP processing board 6 are installed in the shell 1, wherein two identical wide-angle cameras 8 are adopted to acquire the face image of the user, the wide-angle cameras 8 can acquire the face image in a larger range and can acquire the face image comprehensively without rotating the wide-angle cameras 8, the images acquired by the two wide-angle cameras 8 are transmitted into the image acquisition card 5 for processing and then transmitted into the DSP processing board 6, the image processing is carried out in the DSP processing board 6, and the images acquired by the two identical wide-angle cameras 8 are fused, the effect of acquiring images by binocular stereo vision is achieved, the position coordinates of the three-dimensional space where eyes are located are calculated in the DSP processing board 6, the binocular stereo vision can provide a basis for determining the position coordinates of the three-dimensional space, the defects that the positions of a head and a lens are required to be kept unchanged and the distance is short in tracking of eye movement under a two-dimensional plane are overcome, infrared beams are emitted to eyes by the infrared light source array 3, light spots are generated on eyeballs, the position of the adjustable camera 4 is adjusted by the DSP processing board 6 according to the position coordinates of the eyes, the high-definition camera 10 is aligned to the positions of the eyes, light spot images on the eyeballs are acquired, the acquired image information is transmitted to the image acquisition card 5 and then transmitted to the DSP processing board 6, light spot image identification is performed in the DSP processing board 6, the identified face image information and the identified light spot image information are transmitted to the OBD interface 7 by the DSP processing board 6, and the acquired face image is sent to the liquid crystal display screen 2 for display, and the user can receive user registration information and the face image and light spot image information corresponding to the user registration information at the OBD interface 7, so that the system can be conveniently integrated into an access control system, a face payment system and the like.
The liquid crystal display screen 2 is a touch LCD liquid crystal screen.
As shown in fig. 2, the adjustable camera 4 includes a universal gear motor 9 and a high-definition camera 10, the high-definition camera 10 is installed on a rotating shaft of the universal gear motor 9, the DSP processing board 6 controls the rotation direction of the universal gear motor 9, and then adjusts the rotation angle of the high-definition camera 10, so that the high-definition camera 10 can aim at the eyeball of the human eye to collect light spots on the eyeball. High definition digtal camera 10 with adjustable adopts carries out image acquisition to the eyes light spot, can adjust turned angle according to the user of different heights, improves the practicality of system.
As shown in fig. 3, the flow of the method for detecting living bodies by face recognition and eye light spots includes: firstly, the human face of a user is subjected to image acquisition through two identical wide-angle cameras 8, the acquired images are sent to a DSP processing board 6 for face recognition, because two wide-angle cameras 8 are adopted for simultaneously acquiring the human face, the images acquired by the two wide-angle cameras 8 need to be fused in the DSP processing board 6, the binocular stereo vision three-dimensional positioning of human eyes is realized, the human eye area is extracted from the identified human face image, the position coordinates of the human eyes are calculated in the DSP processing board 6, the DSP processing board 6 adjusts the rotation angle of an adjustable camera 4 according to the position coordinates of the human eyes, the adjustable camera 4 aims at the eyeballs of the human eyes for image acquisition, then the DSP processing board 6 controls the quantity of light beams of infrared light emitted by an infrared light source array 3, and the infrared light source array 3 is composed of a plurality of visible infrared light beads with the wavelength range of 850nm, the infrared light source with the intensity higher than that of the ambient light is used for collecting the face image, the face image and the eye image which are irrelevant to the surrounding environment can be extracted, and the infrared image can only change along with the change of the distance between a person and the camera. Moreover, the infrared light source with the waveband of 850nm has no harm to human eyes, and the normal reaction of the human eyes can be ensured. Infrared light beam that infrared light source array 3 sent shines on the eyeball and produces the facula, aim at the facula image on the eyeball of adjustable camera 4 of eyes and gather, and send to DSP processing board 6 and carry out facula image recognition, and calculate the facula quantity on the eyeball, the facula image number that discerns in DSP processing board 6 is the same with the infrared light beam quantity that infrared light source array 3 sent, can judge that user's present face belongs to the living body, can effectually put other people and adopt fake face to replace real face and cause unnecessary loss.
The infrared light source array 3 is composed of four infrared light source arrays 3 and a plurality of visible infrared lamp beads with 850nm wave bands.
As shown in fig. 4, the DSP processing board 6 performs gray processing on the face image acquired by the wide-angle camera 8 to reduce the computational complexity, then performs noise reduction on the image by using a two-dimensional median filtering algorithm to eliminate the influence of noise on the image quality, establishes the relationship between multiple images and a single image by using image registration and V-system image fusion techniques, and obtains a final super-resolution image by using a sub-pixel convolution neural network algorithm, that is, reconstructs a high-resolution image from the acquired low-resolution face image, and requires that the distance between the eyes and the camera is not too long in the eye movement tracking process, otherwise, the pixel value of the eye part is reduced to influence the extraction of the key point features of the eyes, so the super-resolution technique is used to pre-process the acquired image. The super-resolution processing is carried out by utilizing a plurality of images, the characteristics of a target can be extracted more effectively to obtain better super-resolution, so that the super-resolution algorithm based on the plurality of images can solve the problem that the image resolution is not high when the wide-angle camera 8 is far away from human eyes, the human face three-dimensional geometric information is obtained by a binocular stereo vision technology, the binocular stereo vision can provide a basis for determining the position coordinates of a three-dimensional space, the defects that the positions of a head and a lens are kept unchanged and the distance is short in eye movement tracking under a two-dimensional plane are overcome, the three-dimensional space coordinates of the human eyes are calculated conveniently, the human eyes are extracted by a pixel integral projection method due to the fact that the color geometric characteristics and the like of the human eye area and other five sense organ areas have great differences, namely the gray scale image of the human eyes is extracted by adopting the distribution condition of a gray scale curve, and calculating to obtain an accumulation curve of the two pixel values. And then determining the position of the human eye through the distribution of peaks and troughs on the curve. Generally, integral projection is performed on a face image, two maximum value points appear under the condition of vertical projection, the positions of human eyes correspond to the two maximum value points, 4 wave crests are obtained from horizontal projection, the wave crests correspond to eyebrows, eyes, a nose and a mouth, the human eyes are identified according to the characteristics of pixel integration, and then the human eye region is positioned, wherein the human eye positioning calculation method comprises the following steps:
setting a face picture with M multiplied by N pixels, locating the eye region, I (x, y) represents the pixel value at the image point (x, y), x1，x2Is a value in the horizontal direction and is defined as x1,x2]Integral projection function H [ x ] of1,x2]Wherein x is1,x2∈[0,M]。
Define y in the same way1，y2Is taken as a value in the vertical direction in [ y1,y2]Integral projection function V [ y ] of1,y2]Wherein y is1,y2∈[0,N].
The integral projection method obtained by the formulas (1) and (2) is obtained by simply adding corresponding element values, the horizontal and vertical coordinate positions of human eyes are determined by utilizing the maximum value of edge projection of the top half of the human face image in the vertical and horizontal directions so as to position the positions of the human eyes, the actual pixel value of each pixel point is replaced by utilizing the mean value of the pixel values of the neighborhood of each pixel point in a small area, and then integral operation is carried out, so that the influence of noise can be effectively overcome.
As shown in fig. 5, the scheme for implementing the living body detection of human eye light spots includes: the DSP processing board 6 controls the quantity of beams of infrared light emitted by the infrared light source array 3, the emitted infrared light irradiates eyeballs of human eyes to form light spots on the eyeballs, and the infrared light source of 850nm is utilized to irradiate to generate obvious light spots due to different absorption degrees and refraction degrees of retina and iris to infrared light rays with relatively short distance. Therefore, even if the eyes slightly change, the positions of light spots are influenced, the number of the light spots is the number of the infrared light beams irradiating the eyeballs, the DSP processing board 6 adjusts the rotating angle of the adjustable camera 4 according to the recognized coordinates of the positions of the eyes of the user to enable the adjustable camera to aim at the eyes of the user to shoot, the light spot images on the eyeballs can be conveniently collected, and the collected light spot images are sent to the DSP processing boardAnd 6, identifying the light spot image, firstly carrying out gray processing on the light spot image to convert the light spot image into a gray image, reducing the system calculation amount, then carrying out filtering processing on the gray image by adopting a two-dimensional median filtering algorithm, then detecting the light spot of the human eye by adopting Hough transform, wherein the main content of the Hough transform is obtained by space transform, the circle can be well detected by utilizing the Hough transform, and the solving of the circle center by the Hough transform also becomes an important method for determining the center coordinate of the pupil because the light spot on the eyeball is similar to the circle. Transforming the rectangular coordinate system to a parameter space, determining how many collinear coordinates are needed to find out a straight line in a rectangular coordinate plane, if the number of collinear coordinates exceeds a threshold value, judging that a certain straight line is found, then determining the number of collinear coordinates, firstly dividing the polar coordinate space (rho, theta) into a certain number of counting grids, storing a counting variable in the counting grids, calculating theta by traversing each point (x, y) on the rectangular coordinate plane and rotating the counting grid for one circle, putting the calculated point (x, y) with the same rho into the grids which can be counted, then increasing the counting grids by one bit, after traversing all the points (x, y), sequencing counters in all the counting grids of the polar coordinate space (rho, theta) from large to small, and sequencing the largest counting grid is the polar coordinate space (rho, theta) collinear point, and finally mapping the collinear point back to a rectangular coordinate plane in which the equation of a circle is (x-a)2+(y-b)2＝r2According to the transformation of the linear equation, the (X, Y) space of the circle can be transformed into the parameter space (r, θ), and then the rectangular coordinate system equation can become:
x＝a+rcos(θ) (3)
y＝b+rsin(θ) (4)
then, the center of the circle is set to be (a) at any point on the detected imagei,bi) Then, with the radius r known, rotating by 2 π, the coordinates (x) of each point on the circle are obtainedi,yi) Similarly, when the coordinates of the center of a circle are known, the counting variables obtained from the center points (a, b) are obtained when each point is traversedThe value is maximum, and then the light spot image and the number are identified, when the number of the identified light spots is the same as the number of the light beams of the infrared light emitted by the infrared light source array 3, the light spots of the human eyes are judged to be living bodies, otherwise, the light spots are false bodies.
The working principle and the working process of the invention are as follows:
as shown in fig. 6, firstly, two identical wide-angle cameras 8 are used for collecting images of the face of a user, the images collected by the two identical wide-angle cameras 8 are transmitted to an image collecting card 5 for processing and then transmitted to a DSP processing board 6, the images are processed in the DSP processing board 6, the images collected by the two identical wide-angle cameras 8 are fused to achieve the effect of collecting images in binocular stereo vision, the three-dimensional space position coordinates of eyes are calculated in the DSP processing board 6, the DSP processing board 6 controls an infrared light source array 3 to emit infrared light beams to the eyes, light spots are generated on the eyes, the DSP processing board 6 adjusts the steering direction of a universal speed reducing motor 9 according to the position coordinates of the eyes, the rotation angle of a high-definition camera 10 is adjusted, the high-definition camera 10 can be aligned to the eyes of the eyes for collecting, and light spot images on the eyes are collected, the collected image information is transmitted to an image collecting card 5 and then to a DSP processing board 6, facula image recognition is carried out in the DSP processing board 6, the identified face image information and the facula image information are sent to an OBD interface 7 by the DSP processing board 6, the collected face image is sent to a liquid crystal display screen 2 to be displayed, and a user can receive user registration information and the corresponding face image and facula image information at the OBD interface 7.
Claims (4)
1. A human face recognition and human eye light spot living body detection device and method are characterized in that: the system comprises a shell, a liquid crystal display screen, an infrared light source array, an adjustable camera, an image acquisition card, a DSP processing board, an OBD interface and a wide-angle camera, wherein the liquid crystal display screen is embedded on the lower surface of the shell and used for displaying a face image of a user, the user can input account information on the liquid crystal display screen, the infrared light source array, the adjustable camera, the wide-angle camera and the OBD interface are embedded on the upper surface of the shell, the image acquisition card and the DSP processing board are arranged in the shell, two identical wide-angle cameras are adopted for acquiring the face image of the user, the images acquired by the two identical wide-angle cameras are transmitted to the image acquisition card for processing and then are transmitted to the DSP processing board for image processing in the DSP processing board, the images acquired by the two identical wide-angle cameras are fused to achieve the effect of binocular stereoscopic vision image acquisition, and the three-dimensional space position coordinates of eyes are calculated in the DSP processing board, the infrared light source array emits infrared light beams to human eyes, light spots are generated on eyeballs, the DSP processing board adjusts the position of an adjustable camera according to position coordinates where the human eyes are located, the high-definition camera is enabled to be aligned to the positions of the human eyes, light spot images on the eyeballs are collected, collected image information is transmitted to an image collecting card and then transmitted to the DSP processing board, light spot image recognition is conducted in the DSP processing board, recognized face image information and recognized light spot image information are sent to an OBD interface through the DSP processing board, collected face images are sent to a liquid crystal display screen to be displayed, and users can receive user registration information and face images and light spot image information corresponding to the user registration information at the OBD interface.
2. The living body detection device and method based on human face recognition and human eye light spots as claimed in claim 1, wherein: the method for detecting the human face and the human eye light spot in vivo comprises the following steps: firstly, carrying out image acquisition on the face of a user through two identical wide-angle cameras, sending the acquired images to a DSP processing board for face recognition, wherein the two wide-angle cameras are adopted for simultaneously acquiring the face, so that the images acquired by the two wide-angle cameras need to be fused in the DSP processing board, a human eye area is extracted from the recognized face image, the position coordinates of human eyes are calculated in the DSP processing board, the DSP processing board adjusts the rotating angle of the adjustable cameras according to the position coordinates of the human eyes, so that the adjustable cameras aim at the eyeballs of the human eyes for image acquisition, then the DSP processing board controls the quantity of light beams of infrared light emitted by an infrared light source array, infrared light beams emitted by the infrared light source array irradiate the eyeballs to generate light spots, the adjustable cameras aiming at the eyes acquire light spot images on the eyeballs and send the light spot images to the DSP processing board for spot image recognition, and calculating the number of light spots on the eyeball, and judging that the current face of the user belongs to a living body when the number of light spot images identified in the DSP processing board is the same as the number of infrared light beams emitted by the infrared light source array.
3. The living body detection device and method based on human face recognition and human eye light spots as claimed in claim 1, wherein: the DSP processing board carries out gray processing on the face image collected by the wide-angle camera to reduce the calculation complexity, then carries out noise reduction processing on the image by adopting a two-dimensional median filtering algorithm to eliminate the influence of noise on the image quality, establishes the relation between a plurality of images and a single image by image registration and V system image fusion technology, obtains a final super-resolution image by adopting a sub-pixel convolution neural network algorithm, namely reconstructs a high-resolution image in the collected low-resolution face image, requires that the distance between eyes and the camera cannot be too far in the eye movement tracking process, otherwise reduces the pixel value of the eye part to influence the extraction of key point characteristics of the eyes, so that the collected image is preprocessed by adopting a super-resolution technology, and obtains three-dimensional geometric information of the face by adopting a binocular technology to calculate the three-dimensional space coordinates of the eyes, because the color geometric characteristics and the like of the human eye region and other facial features are greatly different, human eyes are extracted by a pixel integral projection method, namely, a gray level image of the human eyes is extracted by adopting the distribution condition of a gray level curve, and then the human eye region is positioned, wherein the human eye positioning calculation method comprises the following steps:
setting a face picture with M multiplied by N pixels, locating the eye region, I (x, y) represents the pixel value at the image point (x, y), x1，x2Is a value in the horizontal direction and is defined as x1,x2]Integral projection function H [ x ] of1,x2]Wherein x is1,x2∈[0,M]，
Define y in the same way1，y2Is taken as a value in the vertical direction in [ y1,y2]Integral projection function V [ y ] of1,y2]Wherein y is1,y2∈[0,N]，
The integral projection method obtained by the formulas (1) and (2) is obtained by simply adding corresponding element values, the horizontal and vertical coordinate positions of human eyes are determined by utilizing the maximum value of edge projection of the top half of the human face image in the vertical and horizontal directions so as to position the positions of the human eyes, the actual pixel value of each pixel point is replaced by utilizing the mean value of the pixel values of the neighborhood of each pixel point in a small area, and then integral operation is carried out, so that the influence of noise can be effectively overcome.
4. The living body detection device and method based on human face recognition and human eye light spots as claimed in claim 1, wherein: the scheme for realizing human eye light spot living body detection is as follows: the DSP processing board controls the quantity of the beams of the infrared light emitted by the infrared light source array, the emitted infrared light irradiates the eyeballs of human eyes, forming light spots on the eyeball, wherein the number of the light spots is the number of the infrared light beams irradiated on the eyeball, adjusting the rotation angle of the adjustable camera by the DSP processing board according to the recognized position coordinates of the human eyes so as to lead the adjustable camera to be aligned with the human eyes for shooting, and the collected light spot image is sent to a DSP processing board for light spot image identification, the light spot image is firstly grayed, and is converted into a gray scale image, the system calculation amount is reduced, then filtering the gray level image by adopting a two-dimensional median filtering algorithm, detecting human eye light spots by adopting Hough transform, transforming a rectangular coordinate system into a parameter space, in order to find out a straight line in the rectangular coordinate plane, it is necessary to determine how many collinear coordinates exist, and if the number of collinear coordinates exceeds a threshold value, it can be determined that the collinear coordinates are found.A certain straight line, then determining the number of collinear coordinates, firstly dividing the polar coordinate space (rho, theta) into a certain number of counting grids, the counting grid stores a counting variable, the counting variable is rotated for one circle by traversing each point (x, y) on the rectangular coordinate plane to calculate theta, the point (x, y) values which are calculated to be the same rho are put into the grids which can be counted, then the counting grid is increased by one bit, when the traversal of all the points (x, y) is completed, sorting counters in all counting grids of the polar coordinate space (rho, theta) from large to small, wherein the largest counting grid is the collinear point of the polar coordinate space (rho, theta), and finally mapping the collinear point to the rectangular coordinate plane, at this time, the detection process of a straight line is completed, and the equation of the circle in the plane rectangular coordinate system is (x-a).2+(y-b)2＝r2According to the transformation of the linear equation, the (X, Y) space of the circle can be transformed into the parameter space (r, θ), and then the rectangular coordinate system equation can become:
x＝a+rcos(θ) (3)
y＝b+rsin(θ) (4)
then, the center of the circle is set to be (a) at any point on the detected imagei,bi) Then, with the radius r known, rotating by 2 π, the coordinates (x) of each point on the circle are obtainedi,yi) And similarly, when the coordinates of the circle center are reversely deduced under the condition of knowing each point on the circle, under the condition of traversing each point, the counting variable value obtained by the circle center point (a) and the circle center point (b) is the largest, so that the image and the number of the light spots are identified, and when the number of the identified light spots is the same as the number of the light beams of the infrared light emitted by the infrared light source array, the light spots of the human eyes are judged to be living bodies, otherwise, the light spots are false bodies.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010625586.6A CN111985303A (en) | 2020-07-01 | 2020-07-01 | Human face recognition and human eye light spot living body detection device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010625586.6A CN111985303A (en) | 2020-07-01 | 2020-07-01 | Human face recognition and human eye light spot living body detection device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111985303A true CN111985303A (en) | 2020-11-24 |
Family
ID=73438447
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010625586.6A Pending CN111985303A (en) | 2020-07-01 | 2020-07-01 | Human face recognition and human eye light spot living body detection device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111985303A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113596410A (en) * | 2021-08-24 | 2021-11-02 | 清华大学深圳国际研究生院 | Target monitoring recognition and tracking camera device and method |
| CN113609973A (en) * | 2021-08-04 | 2021-11-05 | 河南华辰智控技术有限公司 | Social security platform wind control management system based on biological recognition technology |
| CN115482595A (en) * | 2022-09-27 | 2022-12-16 | 北京邮电大学 | Specific character visual sense counterfeiting detection and identification method based on semantic segmentation |
| CN116156708A (en) * | 2023-04-23 | 2023-05-23 | 深圳市帝狼光电有限公司 | Desk lamp illumination control method and device, electronic equipment and storage medium |
| CN117058749A (en) * | 2023-08-17 | 2023-11-14 | 深圳市华弘智谷科技有限公司 | Multi-camera perspective method and device, intelligent glasses and storage medium |
| CN117119307A (en) * | 2023-10-23 | 2023-11-24 | 珠海九松科技有限公司 | Video interaction method based on deep learning |
| CN117152670A (en) * | 2023-10-31 | 2023-12-01 | 江西拓世智能科技股份有限公司 | Behavior recognition method and system based on artificial intelligence |
| CN117496584A (en) * | 2024-01-02 | 2024-02-02 | 南昌虚拟现实研究院股份有限公司 | Eyeball tracking light spot detection method and device based on deep learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1842296A (en) * | 2004-08-03 | 2006-10-04 | 松下电器产业株式会社 | Living body determination device, authentication device using the device, and living body determination method |
| CN101819626A (en) * | 2009-02-26 | 2010-09-01 | 何玉青 | Image fusion-based iris spot elimination method |
| CN102749991A (en) * | 2012-04-12 | 2012-10-24 | 广东百泰科技有限公司 | Non-contact free space eye-gaze tracking method suitable for man-machine interaction |
| CN106218409A (en) * | 2016-07-20 | 2016-12-14 | 长安大学 | A kind of can the bore hole 3D automobile instrument display packing of tracing of human eye and device |
| CN107506705A (en) * | 2017-08-11 | 2017-12-22 | 西安工业大学 | A kind of pupil Purkinje image eye tracking is with watching extracting method attentively |
| CN108256379A (en) * | 2016-12-29 | 2018-07-06 | 广州映博智能科技有限公司 | A kind of eyes posture identification method based on Pupil diameter |
| CN109657531A (en) * | 2018-09-18 | 2019-04-19 | 深圳先牛信息技术有限公司 | A kind of human face in-vivo detection method and detection device based on hot spot on eyeball |
-
2020
- 2020-07-01 CN CN202010625586.6A patent/CN111985303A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1842296A (en) * | 2004-08-03 | 2006-10-04 | 松下电器产业株式会社 | Living body determination device, authentication device using the device, and living body determination method |
| US20080152198A1 (en) * | 2004-08-03 | 2008-06-26 | Shinichi Tsukahara | Biometric Identification Device, Authentication Device Using Same, and Biometric Identification Method |
| CN101819626A (en) * | 2009-02-26 | 2010-09-01 | 何玉青 | Image fusion-based iris spot elimination method |
| CN102749991A (en) * | 2012-04-12 | 2012-10-24 | 广东百泰科技有限公司 | Non-contact free space eye-gaze tracking method suitable for man-machine interaction |
| CN106218409A (en) * | 2016-07-20 | 2016-12-14 | 长安大学 | A kind of can the bore hole 3D automobile instrument display packing of tracing of human eye and device |
| CN108256379A (en) * | 2016-12-29 | 2018-07-06 | 广州映博智能科技有限公司 | A kind of eyes posture identification method based on Pupil diameter |
| CN107506705A (en) * | 2017-08-11 | 2017-12-22 | 西安工业大学 | A kind of pupil Purkinje image eye tracking is with watching extracting method attentively |
| CN109657531A (en) * | 2018-09-18 | 2019-04-19 | 深圳先牛信息技术有限公司 | A kind of human face in-vivo detection method and detection device based on hot spot on eyeball |
Non-Patent Citations (2)
| Title |
|---|
| 翁爽: "基于双目图像的视差估计方法研究及实现", 中国优秀硕士学位论文全文数据库 信息科技辑, pages 41 - 45 * |
| 莫玲;麦仁标;马博;肖苏华;: "人眼状态检测与识别系统研究", 轻工科技, no. 02, pages 77 - 80 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113609973B (en) * | 2021-08-04 | 2024-02-20 | 河南华辰智控技术有限公司 | Social security platform wind control management system based on biological recognition technology |
| CN113609973A (en) * | 2021-08-04 | 2021-11-05 | 河南华辰智控技术有限公司 | Social security platform wind control management system based on biological recognition technology |
| CN113596410A (en) * | 2021-08-24 | 2021-11-02 | 清华大学深圳国际研究生院 | Target monitoring recognition and tracking camera device and method |
| CN115482595A (en) * | 2022-09-27 | 2022-12-16 | 北京邮电大学 | Specific character visual sense counterfeiting detection and identification method based on semantic segmentation |
| CN115482595B (en) * | 2022-09-27 | 2023-04-07 | 北京邮电大学 | Specific character visual sense counterfeiting detection and identification method based on semantic segmentation |
| CN116156708A (en) * | 2023-04-23 | 2023-05-23 | 深圳市帝狼光电有限公司 | Desk lamp illumination control method and device, electronic equipment and storage medium |
| CN116156708B (en) * | 2023-04-23 | 2023-08-08 | 深圳市帝狼光电有限公司 | Desk lamp illumination control method and device, electronic equipment and storage medium |
| CN117058749A (en) * | 2023-08-17 | 2023-11-14 | 深圳市华弘智谷科技有限公司 | Multi-camera perspective method and device, intelligent glasses and storage medium |
| CN117058749B (en) * | 2023-08-17 | 2024-06-07 | 深圳市华弘智谷科技有限公司 | Multi-camera perspective method and device, intelligent glasses and storage medium |
| CN117119307A (en) * | 2023-10-23 | 2023-11-24 | 珠海九松科技有限公司 | Video interaction method based on deep learning |
| CN117119307B (en) * | 2023-10-23 | 2024-03-08 | 珠海九松科技有限公司 | Video interaction method based on deep learning |
| CN117152670A (en) * | 2023-10-31 | 2023-12-01 | 江西拓世智能科技股份有限公司 | Behavior recognition method and system based on artificial intelligence |
| CN117496584A (en) * | 2024-01-02 | 2024-02-02 | 南昌虚拟现实研究院股份有限公司 | Eyeball tracking light spot detection method and device based on deep learning |
| CN117496584B (en) * | 2024-01-02 | 2024-04-09 | 南昌虚拟现实研究院股份有限公司 | Eyeball tracking light spot detection method and device based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111985303A (en) | Human face recognition and human eye light spot living body detection device and method | |
| CN108427503B (en) | Human eye tracking method and human eye tracking device | |
| US20100110374A1 (en) | Apparatus and method for two eye imaging for iris identification | |
| US11829523B2 (en) | Systems and methods for anatomy-constrained gaze estimation | |
| CN108764058B (en) | Double-camera face in-vivo detection method based on thermal imaging effect | |
| CN101999900B (en) | Living body detecting method and system applied to human face recognition | |
| RU2431190C2 (en) | Facial prominence recognition method and device | |
| CN102831392B (en) | Device for remote iris tracking and acquisition, and method thereof | |
| CN112232109B (en) | Living body face detection method and system | |
| US20150227774A1 (en) | On-the-go touchless fingerprint scanner | |
| US8818048B2 (en) | System and method for cancelable iris recognition | |
| CN110232389A (en) | A kind of stereoscopic vision air navigation aid based on green crop feature extraction invariance | |
| WO2014025448A1 (en) | Spoof detection for biometric authentication | |
| US20220148218A1 (en) | System and method for eye tracking | |
| HUT76950A (en) | Automated, non-invasive iris recognition system and method | |
| JP3586456B2 (en) | Personal authentication method and personal authentication device | |
| CN111738241B (en) | Pupil detection method and device based on double cameras | |
| CN109255282B (en) | Biological identification method, device and system | |
| CN107273812B (en) | A kind of living body iris method for anti-counterfeit for authentication | |
| CN110069962A (en) | A kind of biopsy method and system | |
| Bastias et al. | A method for 3D iris reconstruction from multiple 2D near-infrared images | |
| JP2003271932A (en) | Sight line direction detector | |
| US20210192205A1 (en) | Binding of selfie face image to iris images for biometric identity enrollment | |
| JP7078386B2 (en) | Image processing equipment | |
| Cha et al. | Calibration-free gaze zone estimation using convolutional neural network |
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 |