CN106599785A - Method and device for building human body 3D feature identity information database - Google Patents

Method and device for building human body 3D feature identity information database Download PDF

Info

Publication number
CN106599785A
CN106599785A CN201611002201.0A CN201611002201A CN106599785A CN 106599785 A CN106599785 A CN 106599785A CN 201611002201 A CN201611002201 A CN 201611002201A CN 106599785 A CN106599785 A CN 106599785A
Authority
CN
China
Prior art keywords
human
information
characteristic
personal
rgbd
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.)
Granted
Application number
CN201611002201.0A
Other languages
Chinese (zh)
Other versions
CN106599785B (en
Inventor
黄源浩
肖振中
许宏淮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Obi Zhongguang Technology Group Co., Ltd
Original Assignee
Shenzhen Orbbec Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Shenzhen Orbbec Co Ltd filed Critical Shenzhen Orbbec Co Ltd
Priority to CN201611002201.0A priority Critical patent/CN106599785B/en
Publication of CN106599785A publication Critical patent/CN106599785A/en
Application granted granted Critical
Publication of CN106599785B publication Critical patent/CN106599785B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/00201Recognising three-dimensional objects, e.g. using range or tactile information
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/00221Acquiring or recognising human faces, facial parts, facial sketches, facial expressions
    • G06K9/00228Detection; Localisation; Normalisation
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
    • G06K9/00Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
    • G06K9/00973Hardware and software architectures for pattern recognition, e.g. modular organisation
    • G06K9/00979Hardware and software architectures for pattern recognition, e.g. modular organisation structured as a network

Abstract

The invention provides a method and a device for building a human body 3D feature identity information database. The method comprises the following steps: an RGBD human body graph set of a person is acquired, wherein the identity information of the person is known; 3D space distribution feature information for human body feature points of the person is acquired through the RGBD human body graph set; the identity information of the person is marked to the 3D space distribution feature information for the human body feature points of the corresponding person to acquire person information, and the personal information is saved to form a human body 3D identity information database. The device comprises a human body acquisition module, a human body information acquisition module and an information database module. The 3D space distribution feature information for the human body feature points which comprises color information and depth information can be acquired, the method and the device of the invention are not influenced by different seasons, clothing of the person and ambient illumination changes, and the human body recognition accuracy is improved.

Description

The method for building up and equipment of human body 3D characteristic identity information banks
Technical field
The present invention relates to the method for building up field of human body 3D characteristic identity information banks, more particularly to a kind of human body 3D features The method for building up and equipment in identity information storehouse.
Background technology
Information security issue has caused the extensive attention of various circles of society.The main path for ensuring information safety is exactly The identity of information user is accurately differentiated, is determined whether that user obtains whether the authority of information closes by identification result Method, so as to reach the purpose that guarantee information was not leaked and ensured user's legitimate rights and interests.Therefore, reliable identification is very It is important and necessary.
Recognition of face is to carry out a kind of biological identification technology of identification, recognition of face based on the facial feature information of people Technology increasingly receives publicity as a kind of personal identification authentication technique more conveniently, safely.Traditional face recognition technology For 2D recognitions of face, 2D recognitions of face do not have depth information, easily non-several by attitude, expression, illumination and facial makeup etc. The impact of what cosmetic variation, therefore, it is difficult to carrying out accurate recognition of face.
The content of the invention
The present invention provides a kind of method for building up and equipment of human body 3D characteristic identity information banks, can solve the problem that prior art is deposited The low problem of human bioequivalence degree of accuracy.
To solve above-mentioned technical problem, one aspect of the present invention is:A kind of human body 3D features are provided The method for building up in identity information storehouse, the method is comprised the following steps:The personal RGBD human body atlas of collection, wherein, the individual's Identity information is known;The 3d space distribution characteristicss letter of the personal human body feature point is obtained by the RGBD human bodies atlas Breath;The identity information of the individual is identified to into the 3d space distribution characteristicss information of the corresponding human body feature point of the individual To obtain personal information, and the personal information is preserved to form human body 3D characteristic identity information banks.
Wherein, the human body 3D characteristic identities information bank carries out hierarchical classification management to the identity information.
Wherein, the level includes personal attribute's level and group property level.
Wherein, the RGBD human bodies atlas is RGBD human body image sequences;It is described to be obtained by the RGBD human bodies atlas The step of 3d space distribution characteristicss information of the personal human body feature point, also includes:Obtained according to the RGBD human body images sequence Take human body dynamic feature information;The identity information of the individual is identified to into the corresponding human body 3D characteristic informations of the individual To obtain personal information, and it is the step of the personal information is preserved to form human body 3D characteristic identity information banks:By this The identity information of people is identified to the corresponding human body 3D characteristic informations of the individual and human body dynamic feature information to obtain Personal information, and the personal information is preserved to form human body 3D characteristic identity information banks.
Wherein, the identity information of the individual is identified to into the corresponding human body 3D characteristic informations of the individual to obtain Personal information, and by the personal information preserve to form human body 3D characteristic identity information banks the step of after, also include:To institute Stating human body 3D characteristic identities information bank carries out human bioequivalence training.
Wherein, it is described the step of human bioequivalence is trained is carried out to the human body 3D characteristic identities information bank to include:Collection is Know the RGBD human body atlas of the tester of identity information;The people for obtaining the tester is concentrated from the RGBD human figures of the tester The 3d space distribution characteristicss information of body characteristicses point;By the 3d space distribution characteristicss letter of the human body feature point of the tester for obtaining Cease and compare with the 3d space distribution characteristicss information of the human body feature point in the human body 3D characteristic identity information banks;If comparing As a result it is correct, then the 3d space distribution characteristicss of the RGBD human body atlas of the tester, the corresponding human body feature point are believed Breath and the identity information are saved in the human body 3D characteristic identity information banks.
Wherein, the tester include the human body 3D characteristic identity information banks in preserved personal information individual and The human body 3D characteristic identity information banks do not preserve the individual of personal information.
Wherein, the step of gathering personal RGBD human body atlas also includes:The collection personal RGBD face atlas;It is logical Crossing the step of the RGBD human bodies atlas obtains the 3d space distribution characteristicss information of the personal human body feature point also includes:Pass through The RGBD faces atlas obtains the 3d space distribution characteristicss information of the personal human face characteristic point;By the identity of the individual Message identification to the 3d space distribution characteristicss information of the corresponding human body feature point of the individual to obtain personal information, and by institute State the step of personal information is preserved to form human body 3D characteristic identity information banks and be:The identity information is identified to into the face The 3d space distribution characteristicss information of characteristic point and the 3d space distribution characteristicss information of the human body feature point to obtain personal information, And preserve the personal information to form human body 3D characteristic identity information banks.
To solve above-mentioned technical problem, another technical solution used in the present invention is:There is provided and a kind of set up human body The equipment of 3D characteristic identity information banks, the equipment includes human body acquisition module, human body information acquisition module and Knowledge Base Module;People Body acquisition module is used to gather the RGBD human body atlas of individual, wherein, the personal identity information is known;Human body information is obtained Module is connected with the human body acquisition module, for obtaining the 3D of the personal human body feature point by the RGBD human bodies atlas Spatial distribution characteristic information;Knowledge Base Module includes memory module, and the memory module connects with the human body information acquisition module Connect, for the identity information of the individual to be identified to the 3d space distribution characteristicss of the corresponding human body feature point of the individual Information preserves the personal information to form human body 3D characteristic identity information banks to obtain personal information.
Wherein, described information library module also includes management module, and the management module is connected with the memory module, described Management module is used to carry out the identity information hierarchical classification management.
Wherein, the level includes personal attribute's level and group property level.
Wherein, the RGBD human bodies atlas of the human body acquisition module collection is RGBD human body image sequences;The equipment is also Including multidate information acquisition module, it is connected with the human body acquisition module, for according to the RGBD human body images retrieval Human body dynamic feature information;The memory module is also connected with the multidate information acquisition module, for by described in the individual Identity information is identified to the corresponding human body 3D characteristic informations of the individual and human body dynamic feature information to obtain personal information, And preserve the personal information to form human body 3D characteristic identity information banks.
Wherein, the equipment also includes training module, and the training module is believed with the human body acquisition module, the human body Breath acquisition module and the connection of described information library module, for carrying out human bioequivalence instruction to the human body 3D characteristic identities information bank Practice.
Wherein, the training module includes control module and comparing module;The control module gathers mould with the human body Block and the human body information acquisition module connect, for controlling the tester that the human body acquisition module gathers known identities information RGBD human body atlas, and for controlling the human body information acquisition module and concentrating from the RGBD human figures of the tester to obtain Take the 3d space distribution characteristicss information of the human body feature point of the tester;The comparing module is connected with the control module, uses In the 3d space distribution characteristicss information and the human body 3D characteristic identity information of the human body feature point of the tester that will be obtained The 3d space distribution characteristicss information of the human body feature point in storehouse is compared;The memory module is additionally operable to correct in comparison result When, by the RGBD human body atlas of the tester, the 3d space distribution characteristicss information of the corresponding human body feature point and described Identity information is saved in the human body 3D characteristic identity information banks.
Wherein, the tester include the human body 3D characteristic identity information banks in preserved personal information individual and The human body 3D characteristic identity information banks do not preserve the individual of personal information.
Wherein, the equipment also includes face acquisition module and face information acquisition module;Face acquisition module is used to adopt The collection personal RGBD face atlas;Face information acquisition module is connected with the face acquisition module, for by described RGBD face atlas obtains the 3d space distribution characteristicss information of the personal human face characteristic point;The memory module also with the people Face data obtaining module connects, for the identity information to be identified to the 3d space distribution characteristicss information of the human face characteristic point With the 3d space distribution characteristicss information of the human body feature point to obtain personal information, and the personal information is preserved to be formed Human body 3D characteristic identity information banks.
The invention has the beneficial effects as follows:The situation of prior art is different from, the present invention obtains people by RGBD human bodies atlas The 3d space distribution characteristicss information of body characteristicses point, then personal identity information is identified to into the personal corresponding human body feature point 3D Spatial distribution characteristic information is simultaneously preserved together and forms human body 3D characteristic identities information bank for human bioequivalence, due to human body spy The 3d space distribution characteristicss information levied a little includes colouring information and depth information, can set up human skeleton, therefore, the human body Human body information in 3D characteristic identity information banks is more comprehensive, when human bioequivalence is carried out, can recognize more to incite somebody to action accurate, and And, because the human body information in human body 3D characteristic identity information banks is 3D information, therefore, different season, the dress ornament of people and Ambient lighting change etc. will not be impacted to human bioequivalence, therefore the present invention can improve the degree of accuracy of human bioequivalence.
Description of the drawings
Technical scheme in order to be illustrated more clearly that the embodiment of the present invention, below will be to making needed for embodiment description Accompanying drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present invention, for For those of ordinary skill in the art, on the premise of not paying creative work, can be obtaining other according to these accompanying drawings Accompanying drawing.
Fig. 1 is that a kind of flow process of the method for building up of human body 3D characteristic identity information banks that first embodiment of the invention is provided is shown It is intended to;
Fig. 2 is a kind of schematic diagram of human body 3D characteristic identity information banks provided in an embodiment of the present invention;
Fig. 3 is that a kind of flow process of the method for building up of human body 3D characteristic identity information banks that second embodiment of the invention is provided is shown It is intended to;
Fig. 4 is a kind of hierarchical classification pipe of identity information of human body 3D characteristic identity information banks provided in an embodiment of the present invention The schematic diagram of reason;
Fig. 5 is that a kind of flow process of the method for building up of human body 3D characteristic identity information banks that third embodiment of the invention is provided is shown It is intended to;
Fig. 6 is the schematic flow sheet of step S34 in Fig. 5;
Fig. 7 be a kind of human body 3D characteristic identity information banks provided in an embodiment of the present invention be identified train when signal Figure;
Fig. 8 be another kind of human body 3D characteristic identity information banks provided in an embodiment of the present invention be identified train when signal Figure;
Fig. 9 be another human body 3D characteristic identity information banks provided in an embodiment of the present invention be identified train when signal Figure;
Figure 10 is a kind of flow process of the method for building up of human body 3D characteristic identity information banks that fourth embodiment of the invention is provided Schematic diagram;
Figure 11 is a kind of structure of equipment for setting up human body 3D characteristic identity information banks that first embodiment of the invention is provided Schematic diagram;
Figure 12 is a kind of structure of equipment for setting up human body 3D characteristic identity information banks that second embodiment of the invention is provided Schematic diagram;
Figure 13 is a kind of structure of equipment for setting up human body 3D characteristic identity information banks that third embodiment of the invention is provided Schematic diagram;
Figure 14 is a kind of structure of equipment for setting up human body 3D characteristic identity information banks that fourth embodiment of the invention is provided Schematic diagram;
Figure 15 is a kind of entity of equipment for setting up human body 3D characteristic identity information banks that fourth embodiment of the invention is provided The structural representation of device.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete Site preparation is described, it is clear that described embodiment is only a part of embodiment of the present invention, rather than the embodiment of whole.It is based on Embodiment in the present invention, it is all other that those of ordinary skill in the art are obtained under the premise of creative work is not made Embodiment, belongs to the scope of protection of the invention.
Fig. 1 is referred to, Fig. 1 is a kind of foundation side of human body 3D characteristic identity information banks that first embodiment of the invention is provided The schematic flow sheet of method.
The method for building up of the human body 3D characteristic identity information banks of the present embodiment is comprised the following steps:
S11:The personal RGBD human body atlas of collection, wherein, the personal identity information is known.
In step S11, the collection of RGBD human body atlas can be carried out by Kinect sensor, and RGBD human body atlas includes The colouring information (RGB) and depth information (Depth) of human body, compares traditional 2D images and increased depth information, the RGBD people Body atlas can include many personal RGBD human body atlas, and the RGBD human figures of same person concentrate can include that the human body is multiple Multiple RGBD images of angle.
Personally identifiable information can include personal name, sex, age, nationality, native place, contact method, work unit, The personal basic information of department, unit address etc..
In certain embodiments, when occurring how personal in camera lens, then many personal RGBD human body images are gathered.
S12:The 3d space distribution characteristicss information of the personal human body feature point is obtained by RGBD human body atlas.
Specifically, step S12 includes:
S121:Human body feature point is gathered by RGBD human body images.
Specifically, the present embodiment carries out the collection of human body feature point by gathering human body, wherein, human body bag Include:One or more in trunk, extremity and head.
The acquisition method of characteristic point can be various, for example, by the face such as the eyes of handmarking's face, nose, face The characteristic points such as cheek, lower jaw and its edge, it is also possible to which the human face characteristic point labeling method of compatible RGB (2D) is determining the feature of face Point, can also automatic marker characteristic point.
For example, automatic marker characteristic point is needed through three steps:
(1) human body segmentation.The method that the present embodiment is combined using inter-frame difference and background subtraction split-phase is splitting motion Human body, chooses in advance the frame in RGBD images as background frames, sets up the Gauss model of each pixel, then with inter-frame difference Method carries out difference processing to adjacent two field pictures, distinguishes region (the region bag for changing in the current frame of background dot and change Include and appear area and moving object), then the corresponding region of region of variation and background frames is carried out into models fitting and distinguishes to appear area And moving object, shade is finally removed in moving object, so by without the moving meshes of shade out.Context update When inter-frame difference is determined as the point of background, then be updated with certain rule;Background subtraction timesharing is determined as appearing area Point, then background frames are updated with larger turnover rate, the corresponding region of moving object is not updated.The method can be managed relatively The segmentation object thought.
(2) contours extract and analysis.After image after binaryzation is obtained, calculated using some classical rim detection Method is obtaining profile.For example with Canny algorithms, Canny edge detection operators fully reflect the number of optimal edge detector Characteristic is learned, for different types of edge, good signal to noise ratio is respectively provided with, excellent positioning performance is produced many to single edge The low probability of individual response and the maximum suppression ability to false edge response, are obtained behind light stream segmentation field using partitioning algorithm, Comprising our moving targets of concern all in these cut zone.Therefore, Canny will be utilized in these cut zone Operator extraction edge, on the one hand can significantly limit ambient interferences, on the other hand can effectively improve the speed of operation.
(3) the automatic labelling of articulare.Obtain being partitioned into moving target by calculus of finite differences, Canny edge detection operators are carried After contouring, by MaylorK.LeungandYee-HongYang 2D belt patterns (RibbonModel) to human body target Further analysis.Human body front is divided into different regions by the model, for example, with 5 U-shaped regions human body is constructed, this 5 U-shaped region represents respectively the head and extremity of human body.
So, by the body end points of 5 U-shapes of searching, so that it may the approximate location of body is determined, in the profile for having extracted On the basis of, compress to extract the information of needs by vector outline, retain the feature of topmost human limb, by human body Profile is compressed into a fixed shape, for example so that profile is inverted U-shaped with fixed 8 end points and 5 U-shaped points and 3 Point, so obvious feature can conveniently calculate profile.The distance algorithm that adjacent end points on profile may be used herein carrys out compression wheel Exterior feature, by iterative processing so that 8 end points of profile boil down to.
Automatic marking just can be carried out to characteristic point using following algorithm after compression profile is obtained:
(1) the body end points of U-shape is determined.Some reference length M is set, the vector more than M can consider that it is body A part for body profile, then ignores less than it.Begin look for from certain point according to the profile after vector quantization, find one more than M Vector be designated as Mi, find Next vector and be designated as Mj, compare the angle of Mi to Mj, if angle within a certain range (0~ 90 °), (noticing that here angle is just, to represent that it is convex), then it is assumed that they are U end points, record the two vectors, find one Individual U end points.So until finding out 5 U end points.
(2) end points of three inverted u-shaped is determined.Same step (1), as long as angle condition is just being changed to negative.
(3) head, handss, the position of foot are readily available according to U and the end points of U.According to the physiology shape of body, so that it may To determine each articulare, using arm and body angle portions, head and leg angle portions, trunk can be respectively determined Width and length;Then trunk ratio 0.75,0.3 is accounted for respectively using neck, waist position, ancon is located at the midpoint of shoulder and handss, Knee is located at the midpoint of waist and foot.So each characteristic point approximate location can be defined and.
S122:Human body 3D grids are set up according to human body feature point.
S123:The eigenvalue of human body feature point is measured according to human body 3D grids and the 3d space distribution of human body feature point is calculated Characteristic information.
Eigenvalue in step S123 include height, brachium, shoulder breadth, hand size and head size in one or It is multiple.The spatial positional information of each human body feature point can be calculated by human body 3D grids, such that it is able to calculate each individual Topological relation between body characteristicses point, it is hereby achieved that the body shape information of solid, empty with the 3D for obtaining human body feature point Between distribution characteristicss information.When so that the later stage carries out human bioequivalence, can be entered by the 3d space distribution characteristicss information of human body Row identification.
S13:The personal identity information is identified to into the 3d space distribution characteristicss letter of the personal corresponding human body feature point Breath preserves personal information to form human body 3D characteristic identity information banks to obtain personal information.
Personal information in the present embodiment includes the identity information of individual and the 3d space distribution characteristicss of human body feature point Information, identity information is identified to after the 3d space distribution characteristicss information of human body feature point, when the later stage carries out recognition of face, is known It is clipped to the 3d space distribution characteristicss information identical people of human body feature point, you can the 3d space distribution for obtaining the human body feature point is special Reference ceases corresponding identity information.In certain embodiments, personal information includes the 3d space of identity information and human body feature point Distribution characteristicss information and corresponding RGBD atlas, as shown in Fig. 2 Fig. 2 is a kind of human body 3D features provided in an embodiment of the present invention The schematic diagram in identity information storehouse.
Prior art is different from, the present invention obtains the 3d space distribution characteristicss letter of human body feature point by RGBD human bodies atlas Breath, then personal identity information is identified to the personal corresponding human body feature point 3d space distribution characteristicss information and is preserved together And human body 3D characteristic identities information bank is formed for human bioequivalence, due to the 3d space distribution characteristicss packet of human body feature point Colouring information and depth information are included, human skeleton can have been set up, therefore, the human body letter in human body 3D characteristic identity information banks Breath more comprehensively, when human bioequivalence is carried out, can recognize more will be accurate, also, due to human body 3D characteristic identities letter Human body information in breath storehouse is 3D information, therefore, different season, the dress ornament of people and ambient lighting change etc. will not be to human body Identification is impacted, therefore the present invention can improve the degree of accuracy of human bioequivalence.
Fig. 3 is referred to, Fig. 3 is a kind of foundation side of human body 3D characteristic identity information banks that second embodiment of the invention is provided The schematic flow sheet of method.
S21:The personal RGBD human body image sequences of collection, wherein, the personal identity information is known
S22:The 3d space distribution characteristicss information and root of the personal human body feature point are obtained by RGBD human body images According to RGBD human body image retrieval human body dynamic feature informations.
Step S22 according to human body by standing, walking, the behavioral characteristics of the behavior such as run, and specific dynamic behavior Process, such as double palm information such as finger crossover process and result, both arms crossover process and result.
Specifically, the present embodiment can detect the movement posture of human body using the continuous RGBD image sequences of dynamic, Increase the attribute project of feature identification, for example:If target is the rigid body articles such as cup, automobile, in continuous RGBD figures, mesh Mark continuously shows as rigid body, with this discrimination objective as rigid body;If animals such as target behaviour, cat, Canis familiaris L.s, according to continuous dynamic RGBD Tracking target, detecting position non-rigid carries out accurate human bioequivalence according still further to technologies such as characteristics of human body's identifications.
In certain embodiments, the feature of the animals such as collection voice, body temperature or people can also be passed through to be identified certification, Prevent from cracking certification identifying system using image, recording etc., improve accuracy of identification.
Obtain human body dynamic feature information, it is necessary first to carry out human motion detection, i.e., in the image sequence for obtaining Determine the process of being defined of movement human, scale size and attitude.The method of human motion detection has various, for example, OGHMs (Orthogonal Gaussian-Hermite Moments) detection method, its ultimate principle is:It is continuous in time by comparing Picture frame between correspondence pixel value intensity of variation judging whether the pixel belongs to foreground moving region.
With f (x, y, t) | and t=0,1,2 ... } one group of image sequence being input into is represented, f (x, y, t) represents the figure of t Picture, x, y represents the coordinate of pixel on image, if Gaussian functions are g (x, σ), Bn (t) is g (x, σ) and Hermite Polynomial product, then n ranks OGHMs be represented by:
Wherein aiDetermined by the standard deviation of Gaussian functions.According to the property of convolution algorithm, n ranks OGHMs can be regarded as It is the convolution of image sequence function all-order derivative sum in time and Gaussian functions.Certain point derivative value is bigger, then table Show that the pixel value changes changed on the position are also bigger, illustrate that the point should belong to moving region block, this is OGHMs Method can detect that moving object provides theoretical foundation.In addition, can be seen that from formula (1), the basic function of OGHMs isThis is formed by the different order derivative linear combinations of Gaussian functions.Because Gaussian function itself has There is the ability of smooth noise, so OGHMs equally effectively filters out the performance of various noises.
And for example, time differencing method, time differencing method (Temporal Difference) is using image continuous in time Several consecutive frames before and after sequence, based on the time difference of pixel, by thresholding the moving region in image are extracted.The side of early stage Method is to obtain moving object using adjacent two frame difference, such as sets FkIt is kth frame gradation of image Value Data, F in image sequencek+1Table Show the two field picture of kth+1 gray value data in image sequence, then the difference image of temporally adjacent two field pictures is defined as:
Wherein T is threshold value.If difference is more than T, illustrate that the grey scale change in the region is larger, that is, need the fortune for detecting Moving-target region.
And for example, optical flow method (Optical Flow), optical flow method be based on it is assumed hereinafter that:The change of gradation of image entirely due to What the motion of target or background caused.That is, the gray scale of target and background is not changed over time.Motion based on optical flow approach is examined Survey, the characteristic for showing as velocity field in the picture is exactly changed over using moving object, estimated according to certain constraints The corresponding light stream of motion is calculated, its advantage is the interframe movement less-restrictive to target, can process larger interframe displacement.
For another example, background subtraction method (Background Subtraction), its ultimate principle is to build a background first Model image, is then made the difference point with current frame image and background two field picture, and by thresholding difference result moving target is detected. Assume that t background two field picture is F0, correspondence current frame image is Ft, then the difference of present frame and background frames be represented by:
Assume that current frame image is more than threshold value with the gray value difference of background two field picture respective pixel, then resulting two-value Corresponding value is 1 in image, that is, assert that the region belongs to moving target.
After human motion attitude is detected, by motion history image (MHI, motion history images) With the expression that kinergety image (MEI, motion energy images) carries out human action attitude.
The expression of human action attitude is carried out using motion history image (MHI) and kinergety image (MEI), wherein MEI reflects region and the intensity that human action attitude occurs, and MHI then reflects to a certain extent human action attitude How to occur and how to change in time.
Bianry image MEI produces as follows:
Wherein:B (x, y, n) is the bianry image sequence for representing human action attitude generation area, and parameter τ represents that human body is moved The persistent period for gesturing.Therefore, MEI describes the region that whole human action attitude occurs.
The generation of MHI is as follows:
Motion history image MH I not only reflect shape, and the distribution and human action attitude that also reflects brightness occurs Direction.In MHI, the brightness value of each pixel is proportional to the persistent movement time of the position movement posture, and, most The pixel brightness values of the nearly movement posture for occurring are maximum, and the change of gray scale embodies the direction of movement posture generation.
The statistics description of movement posture template is set up using Invariant Moment Method.Bending moment is not:M’k=lg | Mk|, wherein:K= 1,2 ..., 7.Characteristic vector is designated as into F=[M '1, M '2... M '7], use F1, F2..., FMCertain human body of M width in representative image storehouse The image of movement posture. to image Fi, its corresponding characteristic vector is designated as Fi=[M 'i1, M 'i2..., M 'i7], so moved by human body The image library that gestures can be obtained by the eigenmatrix F=M ' of M × 7 of the movement postureij, wherein M 'ijIt is FiJth Individual characteristic element. the mean vector and covariance square of the set of eigenvectors of the M width human action pose presentation is thus obtained Battle array, sets up the statistics description of the movement posture template.
S23:The personal identity information is identified to into the personal corresponding human body 3D characteristic informations and human body behavioral characteristics Information preserves personal information to form human body 3D characteristic identity information banks to obtain personal information.
The human body 3D characteristic informations that the present embodiment is obtained from human body RGBD image sequences are not only including human body feature point 3d space distribution characteristicss information, further comprises human body dynamic feature information, increased the attribute project of feature identification, the present embodiment Human bioequivalence is carried out with reference to the 3d space distribution characteristicss information and human body dynamic feature information of human body feature point so that compare identification Attribute project it is more abundant so that human bioequivalence is more accurate.
In one embodiment, human body 3D characteristic identities information bank carries out hierarchical classification management to identity information, wherein, Level includes personal attribute's level and group property level.
For example, human body 3D characteristic identities information bank will extract characteristics of human body in the 3D characteristic informations of the crowd of group property The 3d space distribution characteristicss information of point and the common feature of human body dynamic feature information.
Fig. 4 is referred to, Fig. 4 is a kind of identity information layer of human body 3D characteristic identity information banks provided in an embodiment of the present invention Levelization sorts out the schematic diagram of management.
For example, personal attribute's level includes the letter such as personal name, sex, age, identification card number with uniqueness The set of breath.Group property then includes old man, child, the adolescence of nonuniqueness, or same company personnel, same office building Staff, or man, woman, or colony's hierarchical information such as Aisan, European and African.
As shown in figure 4, individual A and individual B are child, personal C, individual D and individual E are adolescence, also, personal C is the staff of office building " 1 ", and personal D and individual E is the staff of office building " 2 ".
It is understood that the level of above-mentioned identity information divides a kind of level dividing mode in only the present embodiment, Other embodiments can also have other dividing modes.
Hierarchical classification management can make recognition of face convenient, quick.
For example, the common feature for extracting child according to the division of level is a, and the common feature of adult is b.For example Part game item is only disclosed child in recreation ground, when into the playground, needs to carry out identification, now, The information such as the personal skeleton of acquisition and gait can be passed through and judge whether people to be measured is child.For example, when carrying out identity to personal A During identification, identifying system obtains the RGBD human body image sequences of individual A, so as to obtain the 3d space of the human body feature point of personal A Distribution characteristicss information and individual A walk the human body dynamic feature information of attitude, from the 3d space of the human body feature point of the personal A Distribution characteristicss information and individual A walk attitude human body dynamic feature information in carry and can get feature a, thus individual can be judged A is child, so allowing individual A to pass through.When identification is carried out to personal C, identifying system obtains the RGBD human bodies of individual C Image sequence, from the 3d space distribution characteristicss information of personal C human body feature points and the human body dynamic feature information of attitude on foot Feature a cannot be extracted, thus judges that individual C is not child, so not allowing personal C to pass through.Identification process only needs to part Information identifies people to be measured whether child, the identity information without the need for identifying the specific personal attribute of people to be measured.
Again for example, the identification system used when switching TV, the system only need to recognize people to be measured be adult or Child, so as to open different authorities according to different crowds, to watch different TV programme.When identification system is only obtained The lower limb for getting personal B are walked the RGBD human body image sequences of attitude, the RGBD human figures of attitude of being walked by the lower limb of personal B As sequence can obtain the human body feature point 3d space distribution characteristicss information of the lower limb of personal B and on foot the human body dynamic spy of attitude Reference ceases, and from the human body feature point 3d space distribution characteristicss information of the lower limb of personal B and on foot the human body behavioral characteristics of attitude are believed Feature a can be extracted in breath, therefore judges individual B for child, the authority of TV then automatic opening child.When identification system System gets the lower limb of personal D and walks the RGBD human body image sequences of attitude, the RGBD people of attitude of being walked by the lower limb of personal D Body image sequence can obtain the human body feature point 3d space distribution characteristicss information of the lower limb of personal D and on foot the human body of attitude is moved State characteristic information, the human body dynamic from the human body feature point 3d space distribution characteristicss information of the lower limb of personal D and attitude on foot is special Feature b can be extracted in reference breath, therefore judges that individual D is to be grown up, the TV authority that then automatic opening is grown up.
Above-mentioned identification process can be by the 3d space distribution characteristicss information of less, part human body feature point or people Body dynamic feature information can determine whether that people to be measured is adult or child, and need not obtain more information judging people to be measured is Who.
Again for example, in the gate control system of office building, the gate control system of an office building " 1 ", the gate control system only needs to sentence Whether disconnected identified person is the group property of the inner staff of the office building " 1 ", and need not judge the identified person name, The personal attribute at age etc..Therefore, when through office building gate control system, the human body RGBD images that gate control system gets May be not fine enough, for example, the RGBD image sequences when lower limb for only obtaining personal C are walked about, or only obtain one Overall RGBD images ... the gate control system of people D can call the body of group property level in human body 3D characteristic identity information banks The identification of part information, for example, the human body that gate control system obtains the lower limb of personal C by the image sequence of the lower limb of individual C is special 3d space distribution characteristicss information a little is levied, while obtaining the dynamic feature information of the attitude of walking of personal C, although cannot judge Whom the personal C is, but it can be seen from the information preserved in human body 3D characteristic identity information banks, the action in the office building " 1 " The 3d space distributed intelligence and identical that the human body feature point with identical lower limb is there are in personnel is walked the dynamic of attitude The individual of characteristic information, hence allows to personal C into office building " 1 ".
Or, the gate control system of office building " 2 " passes through the overall image of individual D, can obtain the overall human bodies of personal D Characteristic point 3d space distribution characteristicss information, so as to obtain the human skeleton of personal D, protects according in human body 3D characteristic identity information banks The information deposited understands exist in the office building " 2 " with the overall human body feature point 3d space distribution characteristicss information (people of identical Body skeleton) individual, without the need for judging that whom people to be measured is, then can allow the people to be measured enter office building " 2 ".
If individual D enters office building " 1 ", the gate control system of office building " 1 " obtains the general image sequence of personal D, from And the personal behavioral characteristics letter of the 3d space distribution characteristicss information of acquisition individual D overall human body feature point and on foot attitude Breath, it can be seen from the information preserved in body 3D characteristic identity information banks, does not exist in the office building " 1 " and has identical entirety Human body feature point 3d space distribution characteristicss information and the individual of the personal dynamic feature information of attitude on foot, thus do not permit Perhaps individual D enters office building " 1 ".
Fig. 5 is referred to, Fig. 5 is a kind of foundation side of human body 3D characteristic identity information banks that third embodiment of the invention is provided The schematic flow sheet of method.
S31:The personal RGBD human body atlas of collection, wherein, the personal identity information is known.
S32:The 3d space distribution characteristicss information of the personal human body feature point is obtained by RGBD human body atlas.
S33:The personal identity information is identified to into the 3d space distribution characteristicss letter of the personal corresponding human body feature point Breath preserves personal information to form human body 3D characteristic identity information banks to obtain personal information.
S34:Human bioequivalence training is carried out to human body 3D characteristic identities information bank.
The present embodiment is the step of increased S34 with the difference of first embodiment, to human body 3D characteristic identity information banks Carrying out human bioequivalence training can improve the abundant degree of information resources in human body 3D characteristic identity information banks, so as to improve human body The degree of accuracy of identification.
As shown in fig. 6, Fig. 6 is the schematic flow sheet of step S34 in Fig. 5.Specifically, step S34 includes:
S341:The RGBD human body atlas of the tester of collection known identities information.
In step S341, tester includes having preserved the personal and people of personal information in human body 3D characteristic identity information banks Body 3D characteristic identity information banks do not preserve the individual of personal information.Wherein, RGBD human bodies atlas can be multiple of interruption The set of RGBD human body images.
Wherein, the identity information of tester is, it is known that the identity information section that may refer to tester is known or all Know, for example, it may be the identity information of personal attribute's level of tester is all or part of, it is known that and group property level Identity information is unknown, or identity information section of group property level of tester or all, it is known that and individual Genus Homo The identity information of property level is unknown, or the identity information and group property level that can also be personal attribute's level of tester Identity it is new all known.
S342:The 3d space distribution characteristicss of the human body feature point for obtaining the tester are concentrated from the RGBD human figures of tester Information.
In step S342 the step of the acquisition methods and first embodiment of the 3d space distribution characteristicss information of human body feature point The method of S12 is identical, will not be described here.
S343:The 3d space distribution characteristicss information of the human body feature point of the tester for obtaining and human body 3D characteristic identities are believed The 3d space distribution characteristicss information of the human body feature point in breath storehouse is compared.
For example, in the 3d space distribution characteristicss information of the human body feature point of tester and human body 3D characteristic identity information banks The 3d space distribution characteristicss information of the human body feature point of someone is compared, and show that the 3D of the human body feature point of the tester is empty Between personal X in distribution characteristicss information and human body 3D characteristic identity information banks human body feature point 3d space distribution characteristicss information Similarity reach predetermined threshold value, then may determine that in the tester i.e. human body 3D characteristic identity information banks preserve individual X, if not reaching predetermined threshold value, judges the tester individual that not human body 3D characteristic identity information banks are preserved.
For example, when tester be human body 3D characteristic identity information banks in preserved personal information it is personal when, such as Fruit comparison result is that the tester is corresponding with the personal information preserved in human body 3D characteristic identity information banks, then show to compare knot Fruit is correct, into step S344;If comparison result represents the personal information of the tester in human body 3D characteristic identity information banks In do not preserve, then show comparison result mistake, it is therefore desirable to the tester in human body 3D characteristic identity information banks The information of preservation carries out correcting and further enriching the personal information.
Again for example, when tester be do not preserve in human body 3D characteristic identity information banks personal information it is personal when, if Comparison result is that the information of the tester does not exist in human body 3D characteristic identity information banks, then show that comparison result is correct, Into step S344, to collect the personal information of the tester;If comparison result is the artificial human body 3D feature bodies of the test Someone in part information bank, then show comparison result mistake, needs to the someone in human body 3D characteristic identity information banks Information corrected, and further enrich the personal information, meanwhile, also the personal information of tester is saved in into human body 3D In characteristic identity information bank, to improve the abundant degree of human body 3D characteristic identity information base information resources.
Specifically, the matching of human body dynamic feature information can be in the following manner:Weighed by Mahalanobis distances Similarity between the new movement posture being input into of amount and the known movement posture template for having stored, as long as calculate Mahalanobis distances can be assumed that it is that the match is successful within the threshold range of regulation, if the movement posture of matching is more than One, then that of chosen distance minimum is used as successfully matching.The computing formula of Mahalanobis distances is as follows:
γ2=(f- μx)Tc-1(f-μx)
Wherein, γ is Mahalanobis distances, and f is that the invariant moment features of improved human action pose presentation are vectorial, μx It is the mean vector of the set of eigenvectors trained, c is the covariance matrix of the set of eigenvectors trained.
S344:By the RGBD human body atlas of tester, the 3d space distribution characteristicss information of corresponding human body feature point and body Part information is saved in human body 3D characteristic identity information banks.
The 3d space distribution characteristicss of the RGBD human body atlas of the tester collected in recognition training, human body feature point are believed Breath and identity information are saved in human body 3D characteristic identity information banks in corresponding personal information so that human body 3D characteristic identities are believed Information resources in breath storehouse are more enriched, so as to be beneficial to improve the later stage carry out recognition of face when degree of accuracy.
For example, in the human body 3D characteristic identity information banks tentatively set up, by artificial method by 500 people Part identity message identification to the corresponding RGBD human bodies atlas of 500 people and the 3d space distribution characteristicss of human body feature point In information, in being stored in human body 3D characteristic identity information banks.During recognition training, 5000 or 50000 people of collection are even The RGBD human body atlas of more people is identified training, and an at least partly person is also identified on the RGBD human body atlas of tester Part information, preserves again the personal information of a large amount of testers, if tester is that human body 3D characteristic identity information banks are original, Can continue to supplement the personal RGBD atlas, the 3d space distribution characteristicss information of human body feature point and identity information.
As shown in fig. 7, Fig. 7 is a kind of human body 3D characteristic identity information banks provided in an embodiment of the present invention being identified training When schematic diagram.Human body 3D characteristic identity information banks Central Plains preserves the RGBD human body atlas of personal G, by the RGBD human figures Collection obtain human body feature point 3d space distribution characteristicss information and the identity information with regard to personal attribute, when be identified instruction When white silk, the RGBD human body atlas of the personal G of collection may includes the RGBD human body images of more perspective, from the RGBD people The 3d space distribution characteristicss information of more human body feature points can be obtained in body image, also, during recognition training, in personal G RGBD human figure's centralised identities its work unit, work unit place building etc. group property identity information, therefore, During recognition training, it is same person to recognize the tester with the personal G of face body 3D characteristic identity information banks Central Plains preservation, Therefore, RGBD human body atlas, the 3d space distribution characteristicss information of human body feature point and the mark that individual G is gathered during recognition training The identity information of the group property of knowledge is saved in human body 3D characteristic identity information banks in the personal information of individual G so that people The personal information of individual G is more enriched in body 3D characteristic identity information banks.
Again for example, RGBD human figures image set and the people of acquisition of personal F is saved in human body 3D characteristic identity information banks The 3d space distribution characteristicss information of body characteristicses point, and the information such as manual identification name, work unit, the office building of personal F, but It is sex and age level that the limited RGBD human body images collected by this cannot judge personal F, is identified what is trained When, the RGBD human body image sequences of personal F are obtained, obtain more characteristics of human body from the RGBD human body image sequences The 3d space distribution characteristicss information of point and human body dynamic feature information is obtained, thus can be from newly-increased human body feature point The man for extracting in 3d space distribution characteristicss information and human body dynamic feature information and preserving in human body 3D characteristic identity information banks The common feature of the common feature of property or women, adolescence and old people, for example, has extracted being total to for women in the present embodiment Property feature and adolescence common feature, believe such that it is able to obtain the sex of individual F, the identity of age level by recognition training Breath, the identity information is saved in human body 3D characteristic identity information banks in the personal information of individual F, as shown in figure 8, Fig. 8 is Another kind of human body 3D characteristic identity information banks provided in an embodiment of the present invention are identified schematic diagram when training.
Again for example, any personal information of personal H is not preserved in human body 3D characteristic identities information bank, instruction is identified When white silk, the RGBD human body atlas of personal H is collected, and concentrate the 3D for obtaining human body feature point empty from the RGBD human figures Between distribution characteristicss information, also, RGBD human figure's centralised identities at least part of identity information of personal H, as shown in figure 9, Fig. 9 be another human body 3D characteristic identity information banks provided in an embodiment of the present invention be identified train when schematic diagram.Identification In training process, comparison result is that individual H does not have preservation in human body 3D characteristic identities information bank, so as to by the individual of personal H Information, including the 3d space distribution characteristicss information and identity information of RGBD human body atlas, human body feature point is maintained in human body 3D In characteristic identity information bank, so as to set up the archives of personal H in human body 3D characteristic identity information banks.
In some other embodiment, the present embodiment " can also be carried out human body to human body 3D characteristic identities information bank The step of recognition training ", increases in above-mentioned second embodiment, and when carrying out human bioequivalence training, collection is people to be measured RGBD human body image sequences, the 3d space distribution characteristicss information and people according to the RGBD human body image retrieval human body feature points Body dynamic feature information, by the 3d space distribution characteristicss information and human body dynamic feature information of human body feature point and human body 3D features The 3d space distribution characteristicss information of the human body feature point in identity information storehouse is compared with human body dynamic feature information, will be tested The RGBD human body image sequences of people, the 3d space distribution characteristicss information of corresponding human body feature point, human body dynamic feature information and Identity information is saved in human body 3D characteristic identity information banks.
As shown in Figure 10, Figure 10 is a kind of foundation of human body 3D characteristic identity information banks that fourth embodiment of the invention is provided The schematic flow sheet of method.
S41:The personal RGBD human bodies atlas of collection and RGBD face atlas, wherein, the personal identity information is known.
S42:The 3d space distribution characteristicss information of the personal human body feature point is obtained by RGBD human body atlas;Pass through RGBD face atlas obtains the 3d space distribution characteristicss information of the personal human face characteristic point.
In step S42, the acquisition methods of the 3d space distribution characteristicss information of human body feature point can be with any of the above-described enforcement Example is identical, will not be described here.
The acquisition methods of the 3d space distribution characteristicss information of human face characteristic point are comprised the following steps:
(1) by RGBD man face image acquiring human face characteristic points.
After RGBD facial images are obtained, on the RGBD facial images, by gathering face unit feature is usually carried out The collection of point, wherein, face element includes:Eyebrow, eyes, nose, face, cheek and lower Palestine and China one or more.
The acquisition methods of characteristic point can be various, for example, by the face such as the eyes of handmarking's face, nose, face The characteristic points such as cheek, lower jaw and its edge, it is also possible to which the human face characteristic point labeling method of compatible RGB (2D) is determining the feature of face Point.
For example, the localization method of face key feature points:Choose face 9 characteristic points, these characteristic points point Cloth has angle invariability, respectively 2 eyeball central points, 4 canthus points, the midpoint in two nostrils and 2 corners of the mouth points.In this base Other characteristic point positions of each organ characteristic of the face relevant with identification and extension can be readily available on plinth, be used for into one The recognizer of step.
When face characteristic extraction is carried out, due to the marginal information of local effectively cannot be organized, traditional side Edge detective operators can not reliably extract the feature (profile of eyes or mouth) of face, but from human visual system, fully The positioning of face key feature points is carried out using the feature of edge and angle point, then can greatly improve its reliability.
Susan operators are wherein selected to be used to extract the edge and Corner Feature of regional area.According to the spy of Susan operators Property, it both can be used to detect edge, and can be used for extracting angle point.Therefore compared with the edge detection operator such as Sobel, Canny For, Susan operators are more suitable for carrying out the extraction of the features such as face eye and face, especially to canthus point and corners of the mouth point from Dynamic positioning.
The following is the introduction of Susan operators:
With a circular shuttering traversing graph picture, if the gray value of other any pixels and template center's pixel (core) in template The difference of gray value be less than certain threshold value, be considered as the point and core and there is the gray value of identical (or close), meet such condition Pixel composition region be referred to as the similar area of core value (Univalue Segment Assimilating Nucleus, USAN). It is the base of SUSAN criterions that each pixel in image is associated with the regional area with close gray value.
It is to scan whole image with circular shuttering during concrete detection, compares the ash of each pixel and center pixel in template Angle value, and given threshold value is differentiating whether the pixel belongs to USAN regions, such as following formula:
In formula, c (r, r0) to belong to the discriminant function of the pixel in USAN regions, I (r in template0) it is template center's pixel The gray value of (core), I (r) is the gray value of other any pixels in template, and t is gray scale difference thresholding.It affects to detect angle point Number.T reduces, and more fine changes in image is obtained, so as to provide relatively large number of amount detection.Thresholding t must root The factors such as the contrast and noise according to image determine.Then the USAN area sizes of certain point can be expressed from the next in image:
Wherein, g is geometric threshold, affects the angle point shape for detecting, and g is less, and the angle point for detecting is more sharp.(1)t,g Determination thresholding g determine output angle point USAN regions maximum, as long as the pixel that is, in image have the USAN less than g Region, the point is just judged as angle point.The size of g not only determines the number that angle point can be extracted from image, and such as front institute State, it also determines the acuity of detected angle point.So once it is determined that the quality (acuity) of required angle point, G can just take a changeless value.Thresholding t represents the minimum contrast that can be detected angle point, is also the noise that can ignore Maximum tolerance.It essentially dictates the feature quantity that can be extracted, and t is less, and spy can be extracted from the lower image of contrast Levy, and the feature extracted is also more.Therefore for different contrast and the image of noise situations, different t values should be taken. SUSAN operators have the advantages that one it is prominent, be exactly that, to local insensitive for noise, anti-noise ability is strong.This is because it is not relied on The result of early stage image segmentation, and avoid gradient calculation, in addition, USAN regions are that had with template center pixel by template The pixel of similar gray-value adds up and obtains, and this is actually an integral process, for Gaussian noise has good inhibiting effect.
(1) eyeball and canthus are automatically positioned.During being automatically positioned of eyeball and canthus, initially with normalized mode The method Primary Location face of plate matching.The general area of face is determined in whole facial image.Common human eye positioning Algorithm is positioned according to the valley point property of eyes, and herein then using by the symmetrical of the search of valley point and direction projection and eyeball The method that property combines, using the dependency between two accuracy of eye location can be improved.Upper left to face area Gradient map integral projection is carried out with upper right portion, and the rectangular histogram to integral projection is normalized, first according to floor projection Valley point determine approximate location of the eyes in y directions, then allow x to change in the larger context, find the paddy in this region Point, will detect o'clock as the eyeball central point of two.
On the basis of two eyeball positions are obtained, ocular is processed, initially with self-adaption binaryzation method Determine threshold value, obtain the automatic binary image of ocular, then in conjunction with Susan operators, examined using edge and angle point The algorithm of survey is accurately positioned interior tail of the eye point in ocular.
Through the ocular edge image that above-mentioned algorithm is obtained, on this basis angle is carried out to the boundary curve in image Point is extracted can obtain accurate two ophthalmic external eyes corner location.
(2) nose characteristic of field point is automatically positioned.The key feature points of face nasal area are defined as into two nostrils center The midpoint of line, i.e. muffle central point.The position of face muffle central point is relatively stable, and for facial image normalizing The effect of datum mark is may also function as when changing pretreatment.
Based on two eyeball positions for finding, the position in two nostrils is determined using the method for area grayscale integral projection Put.
The strip region of two eye pupil hole widths is intercepted first, Y-direction integral projection is carried out, and then drop shadow curve is carried out point Analysis.Can see, along drop shadow curve from the search downwards of the Y-coordinate of eyeball position height, the position for finding first valley point is (logical Cross adjustment and select appropriate peak valley Δ value, ignore the middle burr impact produced because of factors such as face's cicatrix or glasses), Using this valley point as naris position Y-coordinate datum mark;Second step is chosen with two eyeball X-coordinate as width, in the Y-coordinate of nostril Lower δ pixels (for example, choose the region that δ=[nostril Y-coordinate-eyeball Y-coordinate] × 0.06) is height and carry out X-direction integration and throw Shadow, is then analyzed to drop shadow curve, and used as central point, respectively to the left and right both sides are carried out the X-coordinate using two eye pupil hole midpoints Search, first for finding valley point is the X-coordinate of the central point in left and right nostril.The midpoint in two nostrils is calculated as in muffle Point, obtains the accurate location of muffle central point, and delimits nasal area.
(3) corners of the mouth is automatically positioned.Because the difference of human face expression may cause the large variation of face shape, and Face region is easier to be disturbed by factors such as beards, thus mouth feature point extract accuracy for identification affect compared with Greatly.Affect relative variability less because the position of corners of the mouth point is expressed one's feelings etc., the position of angle point is more accurate, so taking mouth region Important Characteristic Points be two corners of the mouth points positioning mode.
On the basis of eyes region and nose characteristic of field point is determined, first with the method for area grayscale integral projection It is determined that from first valley point of the following Y-coordinate drop shadow curve in nostril (in the same manner, need to be eliminated by appropriate peak valley Δ value due to The burr that the factors such as beard, nevuss trace are produced affects) as the Y-coordinate position of face;Then face region is selected, to area image Processed using Susan operators, after obtaining mouth edge figure;Angle point grid is finally carried out, two corners of the mouths just can be obtained Exact position.
(2) face colour 3D grids are set up according to human face characteristic point.
(3) eigenvalue of human face characteristic point is measured according to face colour 3D grids and calculates the 3d space point of human face characteristic point Cloth characteristic information.
Specifically, the feature point pairs associated eigenvalue that can be directed to face characteristic by colouring information is measured, should Eigenvalue is face characteristic including in position, distance, shape, size, angle, radian and curvature in 2D planes Plant or various tolerance, additionally, also including the tolerance to color, brightness, texture etc..For example according to iris central pixel point to Surrounding extends, and obtains whole location of pixels of eyes, the shape of eyes, the inclination radian at canthus, eye color etc..
Color combining information and depth information, then can calculate the annexation between characteristic point, and the annexation can Being the topological connection relation and space geometry distance between characteristic point, or it can also be the dynamic of the various combinations of characteristic point Connection relation information etc..
The plane letter of each element including face itself can be obtained according to the tolerance of face colour 3D grids and calculating The local message of the spatial relation of the characteristic point on breath and each element, and the spatial relation between each element Global Information.Local message and Global Information respectively from local and on the whole reflection lie in information on face RGBD figures and Structural relation.
By the analysis to eigenvalue and annexation, thus the face shape information of solid can be obtained, so as to obtain The 3d space distribution characteristicss information of each characteristic point of face, when making the later stage carry out recognition of face, the 3D that can pass through face is empty Between distribution characteristicss information be identified.
For example, can be using finite element method to the topological connection relation and sky between eigenvalue, characteristic point Between geometric distance be analyzed to obtain the 3d space distribution characteristicss information of human face characteristic point.
The formative method of face curve and surface can be described as with mathematical model:
Required deformation curve
Or curved surface
It is the solution of following extreme-value problem
Wherein,It is the energy functional of curve and surface, it reflects to a certain extent the deformation characteristicses of curve and surface, Give curve and surface physical characteristics.F1, f2, f3, f4 are the functions with regard to variable in (),It is parameter definition The border in domain, Γ ' is the curve in Surface Parameters domain, (μ0, v0) be certain parameter value in parameter field, condition (1) be interpolating on sides about Beam, condition (2) is boundary continuity constraint, and condition (3) is the constraint of characteristic curve in curved surface, and condition (4) is in curve and surface Point constraint.In the application, energy functionalTake into following form:
Curve:
Curved surface:
Wherein, α, β, γ represent respectively curve stretching, object for appreciation go, coefficient of torsion, α ij and β ij be respectively curved surface (μ, v) Place removes coefficient very much partially along μ, the drawing in v directions with object for appreciation.
It is both full as can be seen that Deformable curve and surface modeling method is same, process all kinds of constraints in phase from mathematical model Foot Partial controll, in turn ensure that overall wide suitable.Using variation principle, solving above-mentioned extreme-value problem can be converted into solution such as lower section Journey:
Here δ represents first variation.Formula (5) is a differential equation, due to the equation it is more complicated, it is difficult to obtain essence Really analysis is tied, therefore is liberated using numerical value.For example, using finite element method.
Finite Element Method is regarded as first selecting suitable Interpolation as needed, then solves combination parameter, therefore institute The solution for obtaining is not only conitnuous forms, and the grid that pre-treatment is generated also is laid a good foundation for finite element analyses.
Similarity measurement between cognitive phase, unknown facial image and known face template is given by:
In formula:CiXjThe feature of face, i in the feature and face database of face respectively to be identified1,i2,j1,j2,k1,k2For 3D grid vertex features.Section 1 in formula is that machine selects corresponding local feature X in two vector fieldsjAnd CiSimilarity degree, Binomial is then to calculate local location relation and matching order, it can be seen that, when best match i.e. least energy function Match somebody with somebody.
Further, it is also possible to using wavelet transformation texture analysiss method to the Dynamic link library relation between eigenvalue and characteristic point It is analyzed, to obtain the 3d space distribution characteristicss information of characteristic point.
Specifically, Dynamic link library relation is the Dynamic link library relation of various features point combination.Wavelet transformation be the time and The local conversion of frequency, it has the feature of multiresolution analysis, and all has sign signal local feature in time-domain and frequency-domain Ability.The present embodiment is by the way that wavelet transformation texture analysiss are through the extraction to textural characteristics, classification and analytical procedure and combine Face characteristic value and Dynamic link library relation information, specifically include colouring information and depth information, final to obtain three-dimensional face Shape information, finally analyzes again from face shape information and extracts the lower people's shape of face with invariance of face slight expression change Shape information, carries out encoding face shape model parameter, and the model parameter can be used as the geometric properties of face, so as to obtain face The 3d space distribution characteristicss information of characteristic point.
Face 2D features letter has been also compatible with the acquisition methods of the face 3D characteristic informations that some other embodiment is provided The acquisition of breath, the acquisition methods of face 2D characteristic informations can be the conventional various methods in this area.In those embodiments, While obtaining face 3D characteristic informations, face 2D characteristic informations can also be obtained, to carry out the knowledge of 3D and 2D to face simultaneously Not, so as to further improving the degree of accuracy of recognition of face.
For example, the basis of 3 D wavelet transformation is as follows:
Wherein,
AJ1SPACE V is arrived for function f (x, y, z)3 J1Projection operator,
QnFor Hx,Hy,Hz Gx,Gy,GzCombination;
Order matrix H=(HM, k), G=(GM, k), wherein, Hx,Hy,HzRespectively Represent that H is applied to three dimensional signal x, on y, z direction, Gx,Gy,GzRepresent that G is applied to three dimensional signal x respectively, on y, z direction.
Cognitive phase, by unknown facial image wavelet transformation after, take its low frequency low resolution subgraph and be mapped to face space, Characteristic coefficient will be obtained, it is possible to use between Euclidean distance characteristic coefficient more to be sorted and everyone characteristic coefficient away from From with reference to PCA algorithms, according to formula:
In formula, K is the people most matched with unknown face, and N is data base's number, and Y is mapped to by eigenface for unknown face The m dimensional vectors obtained on the subspace of formation, YkIt is mapped on the subspace formed by eigenface for known face in data base The m dimensional vectors for obtaining.
It is to be appreciated that in another embodiment, it is based on the 3D recognitions of face of 2-d wavelet feature that can also adopt Method is identified, it is necessary first to carry out 2-d wavelet feature extraction, and 2-d wavelet basic function g (x, y) is defined as
gmn(x, y)=a-nmG (x ', y ') a > 1, m, n ∈ Z
Wherein, σ is the size of Gauss window, and the filter function of a self similarity can pass through function gmn(x, y) to g (x, Y) carry out suitably expansion and rotation to obtain.Based on superior function, the wavelet character of image I (x, y) can be defined as
Facial image 2-d wavelet extraction algorithm realizes that step is as follows:
(1) the small echo sign with regard to face is obtained by wavelet analysises, converts the individual features in original image I (x, y) For wavelet-based attribute vector F (F ∈ Rm)。
(2) using Fractional power polynomial models (FPP) model k (x, y)=(xy)d(0 < d < 1) makes m dimension wavelet characters space RmProject to higher n-dimensional space RnIn.
(3) based on the linear judgment analysis algorithm (KFDA) of core, in RnMatrix S between class is set up in spacebWith matrix S in classw
Calculate SwNormal orthogonal characteristic vector α1, α2..., αn
(4) extract facial image and significantly differentiate characteristic vector.Another P1=(α1, α2..., αq), wherein, α1, α2..., αqIt is SwCorresponding q eigenvalue is positive characteristic vector, q=rank (Sw).CalculateIt is maximum special corresponding to L Characteristic vector β of value indicative1, β2..., βL, (L≤c-1), wherein,C is face point The quantity of class.Significantly differentiate characteristic vector, fregular=BTP1 TY wherein, v ∈ Rn;B=(β1, β2..., βl)。
(5) the inapparent differentiation characteristic vector of facial image is extracted.CalculateCorresponding to the feature of an eigenvalue of maximum Vectorial γ1, γ2..., γL, (L≤c-1).Make P2=(αq+1, αq+2..., αm), then inapparent differentiation characteristic vector
It is as follows the step of including in the 3D recognition of face stages:
(1) front face is detected, positions face characteristic crucial in a front face and a facial image The contour feature point of point, such as face, left eye and right eye, mouth and nose etc..
(2) three-dimensional people is rebuild by the two-dimensional Gabor characteristic vector and a conventional 3D face database of said extracted Face model.In order to rebuild a three-dimensional face model, using ORL (Olivetti Research Laboratory) single face three Dimension face database, including 100 facial images for detecting.Each faceform has nearly 70000 tops in data base Point.Determine a Feature Conversion matrix P, in original three-dimensional face identification method, the matrix is typically by subspace analysis side The subspace analysis projection matrix that method is obtained, by the covariance matrix of sample the characteristic vector of front m eigenvalue of maximum is corresponded to Composition.The small echo for extracting is differentiated into characteristic vector of the characteristic vector corresponding to m eigenvalue of maximum, main Feature Conversion square is constituted Battle array P ', this feature transition matrix has higher robustness than original eigenmatrix P to factors such as illumination, attitude and expressions, i.e., The feature of representative is more accurately and stable.
(3) newly-generated faceform is processed using template matching and linear discriminant analysiss (FLDA) method, is carried Difference and class inherited in the class of delivery type, further optimize last recognition result.
S43:By identity information be identified to human face characteristic point 3d space distribution characteristicss information and human body feature point 3D it is empty Between distribution characteristicss information to obtain personal information, and personal information is preserved to form human body 3D characteristic identity information banks.
The human body 3D characteristic informations that the present embodiment is obtained include the 3d space distribution characteristicss letter of overall human body feature point The 3d space distribution characteristicss information of the human face characteristic point of breath and local, can be from entirety and local feature when human bioequivalence To be identified, the attribute project of human bioequivalence is increased, improve the degree of accuracy of human bioequivalence.
In some other embodiment, RGB facial images can also be obtained 2D information and the human bodies such as face complexion, texture The 3d space distribution characteristicss information of characteristic point and the 3d space distribution characteristicss information of human face characteristic point combine, and further increase and know Other attribute project, improves identification accuracy.
Figure 11 is referred to, Figure 11 is that a kind of of first embodiment of the invention offer sets up human body 3D characteristic identity information banks The structural representation of equipment.
The equipment for setting up human body 3D characteristic identity information banks of the present embodiment is obtained including human body acquisition module 10, human body information Delivery block 11 and Knowledge Base Module 12.
Specifically, human body acquisition module 10 is used to gather the RGBD human body atlas of individual, its
In, the personal identity information is known.
Human body information acquisition module 11 is connected with human body acquisition module 10, for obtaining the individual by RGBD human body atlas Human body feature point 3d space distribution characteristicss information.
Knowledge Base Module 12 includes memory module 121, and memory module 121 is connected with human body information acquisition module 11, is used for The personal identity information is identified to into the 3d space distribution characteristicss information of the personal corresponding human body feature point to obtain individual Information, and personal information is preserved to form human body 3D characteristic identity information banks.
Figure 12 is referred to, Figure 12 is that a kind of of second embodiment of the invention offer sets up human body 3D characteristic identity information banks The structural representation of equipment.
The equipment for setting up human body 3D characteristic identity information banks of the present embodiment is obtained including human body acquisition module 20, human body information Delivery block 21, Knowledge Base Module 22 and multidate information acquisition module 23.
Wherein, human body acquisition module 20 is used to gather the RGBD human body image sequences of individual, wherein, the personal identity letter Cease for known.
Human body information acquisition module 21 is connected with human body acquisition module 20, for obtaining the individual by RGBD human body atlas Human body feature point 3d space distribution characteristicss information.
Multidate information acquisition module 23 is connected with human body acquisition module 20, for according to RGBD human body image retrieval people Body dynamic feature information.
Knowledge Base Module 22 includes memory module 221 and management module 222.
Memory module 221 is connected with human body information acquisition module 21 and multidate information acquisition module 23, for by this The identity information of people is identified to the personal corresponding human body 3D characteristic informations and human body dynamic feature information to obtain personal information, And preserve personal information to form human body 3D characteristic identity information banks.
Management module 222 is connected with memory module 221, and management module 222 is used to carry out hierarchical classification to identity information Management.Level includes personal attribute's level and group property level.
Figure 13 is referred to, Figure 13 is that a kind of of third embodiment of the invention offer sets up human body 3D characteristic identity information banks The structural representation of equipment.
The equipment for setting up human body 3D characteristic identity information banks of the present embodiment is obtained including human body acquisition module 30, human body information Delivery block 31, Knowledge Base Module 32 and training module 33.
Specifically, human body acquisition module 30 is used to gather the RGBD human body atlas of individual, wherein, the personal identity letter Cease for known.
Human body information acquisition module 31 is connected with human body acquisition module 30, for obtaining the individual by RGBD human body atlas Human body feature point 3d space distribution characteristicss information.
Knowledge Base Module 32 includes memory module 321, and memory module 321 is connected with human body information acquisition module 31, is used for The personal identity information is identified to into the 3d space distribution characteristicss information of the personal corresponding human body feature point to obtain individual Information, and personal information is preserved to form human body 3D characteristic identity information banks.
Training module 33 is connected with human body acquisition module 30, human body information acquisition module 31 and Knowledge Base Module 32, uses In carrying out human bioequivalence training to human body 3D characteristic identities information bank.
Specifically, training module 33 includes control module 331 and comparing module 332, the human body of control module more than 331 collection mould Block 30 and human body information acquisition module 31 connect, for controlling the tester of human body acquisition module collection known identities information RGBD human body atlas, and obtain the tester from the RGBD human figures concentration of tester for controlling human body information acquisition module Human body feature point 3d space distribution characteristicss information.Wherein, tester includes being preserved in human body 3D characteristic identity information banks The individual and human body 3D characteristic identity information banks that have personal information do not preserve the individual of personal information.
Comparing module 332 is connected with control module 331 and memory module 321, and the human body of the tester for obtaining is special Levy the 3d space distribution characteristicss of the human body feature point in 3d space distribution characteristicss information and human body 3D characteristic identity information banks a little Information is compared.
Memory module 321 is additionally operable to when comparison result is correct, and the RGBD human body atlas of tester, corresponding human body is special Levy 3d space distribution characteristicss information a little and identity information is saved in human body 3D characteristic identity information banks.
Figure 14 is referred to, Figure 14 is that a kind of of fourth embodiment of the invention offer sets up human body 3D characteristic identity information banks The structural representation of equipment.
The equipment for setting up human body 3D characteristic identity information banks of the present embodiment is obtained including human body acquisition module 40, human body information Delivery block 41, Knowledge Base Module 42, face acquisition module 43 and face information acquisition module 44.
Specifically, human body acquisition module 40 is used to gather the RGBD human body atlas of individual, wherein, the personal identity information For known.
Human body information acquisition module 41 is connected with human body acquisition module 40, for obtaining the individual by RGBD human body atlas Human body feature point 3d space distribution characteristicss information.
Face acquisition module 43 is used to gather the RGBD face atlas of individual.
Face information acquisition module 44 is connected with face acquisition module 43, for obtaining the individual by RGBD face atlas Human face characteristic point 3d space distribution characteristicss information.
Knowledge Base Module 42 includes memory module 421, and memory module 421 is believed with human body information acquisition module 41 and face Breath acquisition module 44 connect, for the identity information is identified to the human face characteristic point 3d space distribution characteristicss information and The 3d space distribution characteristicss information of the human body feature point preserves the personal information to form people to obtain personal information Body 3D characteristic identity information banks.
Figure 15 is referred to, Figure 15 is that a kind of of fourth embodiment of the invention offer sets up human body 3D characteristic identity information banks The structural representation of the entity apparatus of equipment.The device of present embodiment can perform the step in said method, related content The detailed description in said method is referred to, be will not be described here.
The intelligent electronic device includes the memorizer 52 that processor 51 is coupled with processor 51.
Memorizer 52 be used for storage program area, arrange program and RGBD human body images, RGBD human body image sequences, RGBD facial images, the 3d space distribution characteristicss information of human body feature point, the 3d space distribution characteristicss information of human face characteristic point, people One or more in the information such as body dynamic feature information ....
Processor 51 is used to gather the RGBD human body atlas of individual, wherein, the personal identity information is known;Pass through RGBD human body atlas obtains the 3d space distribution characteristicss information of the personal human body feature point;By the personal identity information mark To the 3d space distribution characteristicss information of the personal corresponding human body feature point to obtain personal information, and by personal information preserve with Form human body 3D characteristic identity information banks.
Processor 51 is additionally operable to human body 3D characteristic identities information bank and hierarchical classification management is carried out to identity information.Wherein, Level includes personal attribute's level and group property level.
Processor 51 is additionally operable to obtain RGBD human body image sequences;It is special according to RGBD human body image retrievals human body dynamic Reference ceases.
Processor 51 is additionally operable to carry out human bioequivalence training to human body 3D characteristic identities information bank.
Processor 51 is additionally operable to the RGBD human body atlas of the tester for gathering known identities information;From the RGBD people of tester The 3d space distribution characteristicss information of the human body feature point of the tester is obtained in body atlas;By the characteristics of human body of the tester for obtaining The 3d space distribution characteristicss letter of the human body feature point in the 3d space distribution characteristicss information and human body 3D characteristic identity information banks of point Breath is compared.
Processor 51 is additionally operable to gather personal RGBD face atlas;The personal face is obtained by RGBD face atlas The 3d space distribution characteristicss information of characteristic point;By the 3d space distribution characteristicss information of human face characteristic point be identified with identity information The 3d space distribution characteristicss information of human body feature point preserves together to form human body 3D characteristic identities letter as personal information Breath storehouse.In several embodiments provided by the present invention, it should be understood that disclosed apparatus and method, it can be passed through Its mode is realized.For example, equipment embodiment described above is only schematic, and for example, module or unit are drawn Point, only a kind of division of logic function can have other dividing mode when actually realizing, such as multiple units or component can To combine or be desirably integrated into another system, or some features can be ignored, or not perform.It is another, it is shown or beg for By coupling each other or direct-coupling or communication connection can be INDIRECT COUPLING by some interfaces, device or unit Or communication connection, can be electrical, mechanical or other forms.
Can be as the unit of separating component explanation or may not be physically separate, show as unit Part can be or may not be physical location, you can with positioned at a place, or can also be distributed to multiple networks On unit.Some or all of unit therein can according to the actual needs be selected to realize the mesh of present embodiment scheme 's.
In addition, each functional unit in each embodiment of the invention can be integrated in a processing unit, also may be used Being that unit is individually physically present, it is also possible to which two or more units are integrated in a unit.It is above-mentioned integrated Unit both can be realized in the form of hardware, it would however also be possible to employ the form of SFU software functional unit is realized.
If integrated unit is realized and as independent production marketing or when using using in the form of SFU software functional unit, can In to be stored in a computer read/write memory medium.Based on such understanding, technical scheme substantially or Saying all or part of the part or technical scheme contributed to prior art can be embodied in the form of software product Out, the computer software product is stored in a storage medium, including some instructions are used so that a computer equipment (can be personal computer, server, or network equipment etc.) or processor (processor) perform each enforcement of the present invention The all or part of step of methods.And aforesaid storage medium includes:USB flash disk, portable hard drive, read only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. it is various Can be with the medium of store program codes.
In sum, the present invention is not affected by different seasons, the dress ornament of people and ambient lighting change etc., therefore is improved The degree of accuracy of human bioequivalence.
Embodiments of the present invention are the foregoing is only, the scope of the claims of the present invention is not thereby limited, it is every using this Equivalent structure or equivalent flow conversion that description of the invention and accompanying drawing content are made, or directly or indirectly it is used in other correlations Technical field, is included within the scope of the present invention.

Claims (16)

1. a kind of method for building up of human body 3D characteristic identity information banks, it is characterised in that comprise the following steps:
The personal RGBD human body atlas of collection, wherein, the personal identity information is known;
The 3d space distribution characteristicss information of the personal human body feature point is obtained by the RGBD human bodies atlas;
The identity information of the individual is identified to into the 3d space distribution characteristicss letter of the corresponding human body feature point of the individual Breath preserves the personal information to form human body 3D characteristic identity information banks to obtain personal information.
2. method according to claim 1, it is characterised in that the human body 3D characteristic identities information bank is believed the identity Breath carries out hierarchical classification management.
3. method according to claim 2, it is characterised in that the level includes personal attribute's level and group property layer Level.
4. method according to claim 3, it is characterised in that the RGBD human bodies atlas is RGBD human body image sequences;
It is described the personal human body feature point is obtained by the RGBD human bodies atlas 3d space distribution characteristicss information the step of Also include:According to the RGBD human body images retrieval human body dynamic feature information;
The identity information of the individual is identified to into the corresponding human body 3D characteristic informations of the individual to obtain personal information, And by the personal information preserve to form human body 3D characteristic identity information banks the step of be:By the identity information of the individual The corresponding human body 3D characteristic informations of the individual and human body dynamic feature information are identified to obtain personal information, and will be described Personal information preserves to form human body 3D characteristic identity information banks.
5. method according to claim 1, it is characterised in that the identity information of the individual is identified to into the individual right The human body 3D characteristic informations answered preserve the personal information to form human body 3D characteristic identities to obtain personal information After the step of information bank, also include:
Human bioequivalence training is carried out to the human body 3D characteristic identities information bank.
6. method according to claim 5, it is characterised in that described that pedestrian is entered to the human body 3D characteristic identity information banks The step of body recognition training, includes:
The RGBD human body atlas of the tester of collection known identities information;
The 3d space distribution characteristicss information of the human body feature point for obtaining the tester is concentrated from the RGBD human figures of the tester;
The 3d space distribution characteristicss information of the human body feature point of the tester for obtaining is believed with the human body 3D characteristic identities The 3d space distribution characteristicss information of the human body feature point in breath storehouse is compared;
If comparison result is correct, by the RGBD human body atlas of the tester, the 3d space of the corresponding human body feature point Distribution characteristicss information and the identity information are saved in the human body 3D characteristic identity information banks.
7. method according to claim 6, it is characterised in that the tester includes the human body 3D characteristic identity information The personal and described human body 3D characteristic identity information banks for having preserved personal information in storehouse do not preserve the individual of personal information.
8. method according to claim 1, it is characterised in that also include the step of collection personal RGBD human body atlas: The collection personal RGBD face atlas;
The step of 3d space distribution characteristicss information that the personal human body feature point is obtained by the RGBD human bodies atlas, is also wrapped Include:The 3d space distribution characteristicss information of the personal human face characteristic point is obtained by the RGBD faces atlas;
The identity information of the individual is identified to into the 3d space distribution characteristicss letter of the corresponding human body feature point of the individual Breath is the step of the personal information is preserved to form human body 3D characteristic identity information banks to obtain personal information:By institute State identity information and be identified to the 3d space distribution characteristicss information of the human face characteristic point and the 3d space point of the human body feature point Cloth characteristic information preserves the personal information to form human body 3D characteristic identity information banks to obtain personal information.
9. a kind of equipment for setting up human body 3D characteristic identity information banks, it is characterised in that include:
Human body acquisition module, for gathering the RGBD human body atlas of individual, wherein, the personal identity information is known;
Human body information acquisition module, is connected with the human body acquisition module, for obtaining this by the RGBD human bodies atlas The 3d space distribution characteristicss information of the human body feature point of people;
Knowledge Base Module, including memory module, the memory module is connected with the human body information acquisition module, for by this The 3d space distribution characteristicss information that the identity information of people is identified to the corresponding human body feature point of the individual is individual to obtain People's information, and the personal information is preserved to form human body 3D characteristic identity information banks.
10. equipment according to claim 9, it is characterised in that described information library module also includes management module, the pipe Reason module is connected with the memory module, and the management module is used to carry out the identity information hierarchical classification management.
11. equipment according to claim 10, it is characterised in that the level includes personal attribute's level and group property Level.
12. equipment according to claim 11, it is characterised in that the RGBD human body atlas of the human body acquisition module collection For RGBD human body image sequences;
The equipment also includes multidate information acquisition module, is connected with the human body acquisition module, for according to the RGBD people Body image sequence obtains human body dynamic feature information;
The memory module is also connected with the multidate information acquisition module, for the identity information of the individual to be identified to The corresponding human body 3D characteristic informations of the individual and human body dynamic feature information are believed the individual with obtaining personal information Breath preserves to form human body 3D characteristic identity information banks.
13. equipment according to claim 9, it is characterised in that the equipment also includes training module, the training module It is connected with the human body acquisition module, the human body information acquisition module and described information library module, for the human body 3D characteristic identities information bank carries out human bioequivalence training.
14. equipment according to claim 13, it is characterised in that the training module includes control module and compares mould Block;
The control module is connected with the human body acquisition module and the human body information acquisition module, for controlling the human body The RGBD human body atlas of the tester of acquisition module collection known identities information, and obtain mould for controlling the human body information Block concentrates the 3d space distribution characteristicss information of the human body feature point for obtaining the tester from the RGBD human figures of the tester;
The comparing module is connected with the control module, for will obtain the tester human body feature point 3d space The 3d space distribution characteristicss information of the human body feature point in distribution characteristicss information and the human body 3D characteristic identity information banks is carried out Compare;
The memory module is additionally operable to when comparison result is correct, by the RGBD human body atlas of the tester, corresponding described The 3d space distribution characteristicss information and the identity information of human body feature point is saved in the human body 3D characteristic identity information banks.
15. equipment according to claim 14, it is characterised in that the tester believes including the human body 3D characteristic identities The personal and described human body 3D characteristic identity information banks for having preserved personal information in breath storehouse do not preserve the individual of personal information.
16. equipment according to claim 9, it is characterised in that the equipment also includes:
Face acquisition module, for gathering the RGBD face atlas of the individual;
Face information acquisition module, is connected with the face acquisition module, for obtaining this by the RGBD faces atlas The 3d space distribution characteristicss information of the human face characteristic point of people;
The memory module is also connected with the face information acquisition module, for the identity information to be identified to into the face The 3d space distribution characteristicss information of characteristic point and the 3d space distribution characteristicss information of the human body feature point to obtain personal information, And preserve the personal information to form human body 3D characteristic identity information banks.
CN201611002201.0A 2016-11-14 2016-11-14 Method and equipment for establishing human body 3D characteristic identity information base Active CN106599785B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201611002201.0A CN106599785B (en) 2016-11-14 2016-11-14 Method and equipment for establishing human body 3D characteristic identity information base

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201611002201.0A CN106599785B (en) 2016-11-14 2016-11-14 Method and equipment for establishing human body 3D characteristic identity information base

Publications (2)

Publication Number Publication Date
CN106599785A true CN106599785A (en) 2017-04-26
CN106599785B CN106599785B (en) 2020-06-30

Family

ID=58590298

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201611002201.0A Active CN106599785B (en) 2016-11-14 2016-11-14 Method and equipment for establishing human body 3D characteristic identity information base

Country Status (1)

Country Link
CN (1) CN106599785B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107273804A (en) * 2017-05-18 2017-10-20 东北大学 Pedestrian recognition method based on SVMs and depth characteristic
CN107563359A (en) * 2017-09-29 2018-01-09 重庆市智权之路科技有限公司 Recognition of face temperature analysis generation method is carried out for dense population
CN108009483A (en) * 2017-11-28 2018-05-08 信利光电股份有限公司 A kind of image collecting device, method and intelligent identifying system
CN108416312A (en) * 2018-03-14 2018-08-17 天目爱视(北京)科技有限公司 A kind of biological characteristic 3D data identification methods and system taken pictures based on visible light
CN108537236A (en) * 2018-04-04 2018-09-14 天目爱视(北京)科技有限公司 A kind of polyphaser data control system for identifying
WO2020134478A1 (en) * 2018-12-29 2020-07-02 北京灵汐科技有限公司 Face recognition method, feature extraction model training method and device thereof

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040037450A1 (en) * 2002-08-22 2004-02-26 Bradski Gary R. Method, apparatus and system for using computer vision to identify facial characteristics
CN102164113A (en) * 2010-02-22 2011-08-24 深圳市联通万达科技有限公司 Face recognition login method and system
CN103235943A (en) * 2013-05-13 2013-08-07 苏州福丰科技有限公司 Principal component analysis-based (PCA-based) three-dimensional (3D) face recognition system
CN104091159A (en) * 2014-07-14 2014-10-08 无锡市合鑫川自动化设备有限公司 Human face identification method
CN104573634A (en) * 2014-12-16 2015-04-29 苏州福丰科技有限公司 Three-dimensional face recognition method
KR101556992B1 (en) * 2014-03-13 2015-10-05 손우람 3d scanning system using facial plastic surgery simulation
CN105184280A (en) * 2015-10-10 2015-12-23 东方网力科技股份有限公司 Human body identity identification method and apparatus
CN105447466A (en) * 2015-12-01 2016-03-30 深圳市图灵机器人有限公司 Kinect sensor based identity comprehensive identification method
CN105786016A (en) * 2016-03-31 2016-07-20 深圳奥比中光科技有限公司 Unmanned plane and RGBD image processing method
CN106023363A (en) * 2016-05-12 2016-10-12 重庆佐鸣科技有限公司 Identity verification method and system

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20040037450A1 (en) * 2002-08-22 2004-02-26 Bradski Gary R. Method, apparatus and system for using computer vision to identify facial characteristics
CN102164113A (en) * 2010-02-22 2011-08-24 深圳市联通万达科技有限公司 Face recognition login method and system
CN103235943A (en) * 2013-05-13 2013-08-07 苏州福丰科技有限公司 Principal component analysis-based (PCA-based) three-dimensional (3D) face recognition system
KR101556992B1 (en) * 2014-03-13 2015-10-05 손우람 3d scanning system using facial plastic surgery simulation
CN104091159A (en) * 2014-07-14 2014-10-08 无锡市合鑫川自动化设备有限公司 Human face identification method
CN104573634A (en) * 2014-12-16 2015-04-29 苏州福丰科技有限公司 Three-dimensional face recognition method
CN105184280A (en) * 2015-10-10 2015-12-23 东方网力科技股份有限公司 Human body identity identification method and apparatus
CN105447466A (en) * 2015-12-01 2016-03-30 深圳市图灵机器人有限公司 Kinect sensor based identity comprehensive identification method
CN105786016A (en) * 2016-03-31 2016-07-20 深圳奥比中光科技有限公司 Unmanned plane and RGBD image processing method
CN106023363A (en) * 2016-05-12 2016-10-12 重庆佐鸣科技有限公司 Identity verification method and system

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
SRINIVAS KISHAN ANAPU: "Fusion of RGB and Depth Images for Robust Face Recognition using Close-Range 3D Camera", 《INTERNATIONAL JOURNAL OF INNOVATIVE RESEARCH IN COMPUTER AND COMMUNICATION ENGINEERING》 *
戴萧何: "基于RGB-D数据的人体检测与跟踪", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
李晓捷: "基于粒子滤波器算法的特征标记点匹配", 《计算机工程》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107273804A (en) * 2017-05-18 2017-10-20 东北大学 Pedestrian recognition method based on SVMs and depth characteristic
CN107563359A (en) * 2017-09-29 2018-01-09 重庆市智权之路科技有限公司 Recognition of face temperature analysis generation method is carried out for dense population
CN107563359B (en) * 2017-09-29 2018-09-11 重庆市智权之路科技有限公司 Recognition of face temperature is carried out for dense population and analyzes generation method
CN108009483A (en) * 2017-11-28 2018-05-08 信利光电股份有限公司 A kind of image collecting device, method and intelligent identifying system
CN108416312A (en) * 2018-03-14 2018-08-17 天目爱视(北京)科技有限公司 A kind of biological characteristic 3D data identification methods and system taken pictures based on visible light
CN108416312B (en) * 2018-03-14 2019-04-26 天目爱视(北京)科技有限公司 A kind of biological characteristic 3D data identification method taken pictures based on visible light
CN108537236A (en) * 2018-04-04 2018-09-14 天目爱视(北京)科技有限公司 A kind of polyphaser data control system for identifying
WO2020134478A1 (en) * 2018-12-29 2020-07-02 北京灵汐科技有限公司 Face recognition method, feature extraction model training method and device thereof

Also Published As

Publication number Publication date
CN106599785B (en) 2020-06-30

Similar Documents

Publication Publication Date Title
CN103632132B (en) Face detection and recognition method based on skin color segmentation and template matching
CN106778584B (en) A kind of face age estimation method based on further feature Yu shallow-layer Fusion Features
CN106599785A (en) Method and device for building human body 3D feature identity information database
CN102324025B (en) Human face detection and tracking method based on Gaussian skin color model and feature analysis
CN106682598B (en) Multi-pose face feature point detection method based on cascade regression
CN103942577B (en) Based on the personal identification method for establishing sample database and composite character certainly in video monitoring
CN107168527B (en) The first visual angle gesture identification and exchange method based on region convolutional neural networks
CN106778468B (en) 3D face identification method and equipment
CN104091147B (en) A kind of near-infrared eyes positioning and eye state identification method
CN103049751A (en) Improved weighting region matching high-altitude video pedestrian recognizing method
CN106778474A (en) 3D human body recognition methods and equipment
WO2016110005A1 (en) Gray level and depth information based multi-layer fusion multi-modal face recognition device and method
CN108182397B (en) Multi-pose multi-scale human face verification method
CN105654035B (en) Three-dimensional face identification method and the data processing equipment for applying it
CN107368778A (en) Method for catching, device and the storage device of human face expression
CN108256421A (en) A kind of dynamic gesture sequence real-time identification method, system and device
CN106650574A (en) Face identification method based on PCANet
CN108629336A (en) Face value calculating method based on human face characteristic point identification
CN107330371A (en) Acquisition methods, device and the storage device of the countenance of 3D facial models
Li et al. Robust iris segmentation based on learned boundary detectors
CN106778491B (en) The acquisition methods and equipment of face 3D characteristic information
CN106611158A (en) Method and equipment for obtaining human body 3D characteristic information
CN108563999A (en) A kind of piece identity's recognition methods and device towards low quality video image
CN107480586A (en) Bio-identification photo bogus attack detection method based on human face characteristic point displacement
CN106778489A (en) The method for building up and equipment of face 3D characteristic identity information banks

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
CP01 Change in the name or title of a patent holder

Address after: 518057 Guangdong city of Shenzhen province Nanshan District Hing Road three No. 8 China University of Geosciences research base in building A808

Patentee after: Obi Zhongguang Technology Group Co., Ltd

Address before: 518057 Guangdong city of Shenzhen province Nanshan District Hing Road three No. 8 China University of Geosciences research base in building A808

Patentee before: SHENZHEN ORBBEC Co.,Ltd.

CP01 Change in the name or title of a patent holder