CN107229925A - Conversed using ear recognition - Google Patents
Conversed using ear recognition Download PDFInfo
- Publication number
- CN107229925A CN107229925A CN201710567421.6A CN201710567421A CN107229925A CN 107229925 A CN107229925 A CN 107229925A CN 201710567421 A CN201710567421 A CN 201710567421A CN 107229925 A CN107229925 A CN 107229925A
- Authority
- CN
- China
- Prior art keywords
- image
- face
- ear
- identification
- recognition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04M—TELEPHONIC COMMUNICATION
- H04M1/00—Substation equipment, e.g. for use by subscribers
- H04M1/72—Mobile telephones; Cordless telephones, i.e. devices for establishing wireless links to base stations without route selection
- H04M1/724—User interfaces specially adapted for cordless or mobile telephones
- H04M1/72448—User interfaces specially adapted for cordless or mobile telephones with means for adapting the functionality of the device according to specific conditions
Abstract
The present invention provides the device that a kind of utilization ear recognition changes state, it is characterised in that including：Module is projected, for emitting structural light image to the region comprising human ear；Imaging modules, for gathering the structure light image for including human ear；Process circuit, ear detection and identification are carried out using the structure light image, and according to the ear detection and the state of recognition result change described device.Device provided by the present invention carries out the functions such as auto-pickup using the detection to human ear and identification, greatly improves the facility of communication device.
Description
Technical field
The invention belongs to field of computer technology, more specifically, it is related to and a kind of is conversed using ear recognition
Method.
Background technology
Human body has many unique features, such as face, fingerprint, iris, human ear etc., and these features are collectively referred to as biological spy
Levy.Living things feature recognition is widely used in the various fields such as security protection, household, Intelligent hardware, at present more ripe biological characteristic
Identification such as fingerprint recognition, iris recognition etc. have been widely used in the terminals such as mobile phone, computer.And for features such as faces, to the greatest extent
The related research of pipe is very deep, and the identification for features such as faces is not popularized then yet, and this is primarily due to existing
Recognition methods causes the stability of discrimination and identification relatively low in the presence of limitation.These limitations are main include by ambient light light intensity and
Direction of illumination influences, human face expression influences discrimination and is easily cheated by artificial feature.
The identification of the features such as existing face, is based primarily upon face Two-dimensional Color Image, when environmental light intensity is weaker, can be tight
Ghost image rings recognition effect.In addition, when the direction of illumination is different, can have shade on facial image, equally can also influence identification
Effect.Gathered in the case of referenced facial image is not being expressed one's feelings, and be currently at the lower collection of expression of smiling
Facial image, the effect of recognition of face can also decline.In addition, if identified object is not real human face, but the face of two dimension
During picture, it often can also pass through identification.
The problem of for the above, the living things feature recognition based on near-infrared or thermal infrared images is generally used at present, it is near red
Outer image will not can be improved identification stability by the interference of ambient light, but but be difficult to solve asking for artificial feature deception
Topic；Thermal infrared images is only imaged to real human face, therefore can solve the problem of artificial feature is cheated, but thermal infrared images divides
Resolution is low, has a strong impact on recognition effect.
Can be used for unlocking based on living things feature recognition, pay etc. be equipped with these identification functions on task, mobile device will
It is favorably improved the convenience used.The present invention will provide a kind of automatic call scheme based on ear recognition, will greatly carry
Rise call convenience.
The content of the invention
The present invention provides a kind of method conversed using ear recognition, the following institute of the technical solution adopted by the present invention
State.
The present invention provides the device that a kind of utilization ear recognition changes state, it is characterised in that including：Module is projected, is used
In emitting structural light image to the region comprising human ear；Imaging modules, for gathering the structure light image for including human ear；Place
Circuit is managed, ear detection and identification is carried out using the structure light image, and change according to the ear detection and recognition result
The state of described device.
In some embodiments, the structure light image includes infrared speckle image.The infrared speckle image dissipate
Spot grain density is arranged to not block the main textural characteristics of the human ear.
In some embodiments, the ear detection is to be calculated using the detection based on machine learning with identification with identification
Sample Storehouse in method, the algorithm for model learning is made up of coloured image and/or gray level image.
In some embodiments, the ear detection with recognize include to the position of the human ear with apart from knowing
Not.
In some embodiments, the ear detection and identification include judging current human ear whether with target human ear phase
Together.
In some embodiments, described device includes communication device, and the state of the change described device includes device
Origin electricity condition changes into the state of answering.
In some embodiments, described device includes audio devices, and the state of the change described device includes device
Sound status is put outside loudspeaker and changes into earpiece sound state.
The present invention also provides a kind of method that utilization ear recognition changes state, it is characterised in that including：Utilize projective module
The group emitting structural light image extremely region comprising human ear；The structure light image for including human ear is gathered using imaging modules；It is logical
Cross process circuit and the structure light image is subjected to ear detection and identification, and changed according to the ear detection and recognition result
The state of described device.
In some embodiments, the structure light image include infrared speckle image, the infrared speckle image dissipate
Spot grain density is arranged to not block the main textural characteristics of the human ear.
Beneficial effects of the present invention are：Device provided by the present invention is carried out automatic using the detection to human ear and identification
The function such as answer, greatly improve the facility of communication device.
Brief description of the drawings
The recognition of face schematic diagram of a scenario according to an embodiment of the present invention shown in Fig. 1.
Fig. 2 is the structural representation of face identification device according to an embodiment of the invention.
Fig. 3 be it is according to an embodiment of the invention utilize structure light image carry out recognition of face block diagram.
Fig. 4 is according to an embodiment of the invention to utilize structure light image and visible images to carry out recognition of face step
Figure.
Fig. 5 is according to an embodiment of the invention to utilize structure light image and depth image to carry out recognition of face step
Figure.
Fig. 6 is according to an embodiment of the invention to utilize structure light image and thermal infrared images to carry out recognition of face step
Figure.
Fig. 7 is the task executing method block diagram according to embodiments of the present invention based on recognition of face.
Embodiment
In order that technical problem to be solved of the embodiment of the present invention, technical scheme and beneficial effect are more clearly understood,
Below in conjunction with drawings and Examples, the present invention will be described in further detail.It should be appreciated that specific implementation described herein
Example is not intended to limit the present invention only to explain the present invention.
It should be noted that when element is referred to as " being fixed on " or " being arranged at " another element, it can be directly another
On one element or it is connected on another element.When an element is known as " being connected to " another element, it can
To be directly to another element or be indirectly connected on another element.In addition, connection can be to be used to fix
Effect can also be used for circuit communication effect.
It is to be appreciated that term " length ", " width ", " on ", " under ", "front", "rear", "left", "right", " vertical ",
The orientation or position relationship of the instruction such as " level ", " top ", " bottom " " interior ", " outer " are to be closed based on orientation shown in the drawings or position
System, is for only for ease of the description embodiment of the present invention and simplifies description, rather than indicate or imply that the device or element of meaning must
There must be specific orientation, with specific azimuth configuration and operation, therefore be not considered as limiting the invention.
In addition, term " first ", " second " are only used for describing purpose, and it is not intended that indicating or implying relative importance
Or the implicit quantity for indicating indicated technical characteristic.Thus, define " first ", the feature of " second " can express or
Implicitly include one or more this feature.In the description of the embodiment of the present invention, " multiple " are meant that two or two
More than, unless otherwise specifically defined.
The invention provides a kind of devices and methods therefor that living things feature recognition is carried out using structure light image.Following
Will be so that face characteristic be recognized as an example in elaboration.
Face recognition technology can be used for safety check, monitoring, now with the popularization of intelligent terminal such as mobile phone, flat board,
Face recognition technology can also be applied to unlock, pay, or even many aspects such as amusement game.Intelligent terminal, such as hand
Machine, flat board, computer, TV etc. are provided with color camera greatly, are gathered using color camera after the image comprising face, utilize this
Image carries out Face datection and identification, so as to further perform other related applications using the result of identification.However, for picture
For the mobile terminal devices such as mobile phone, flat board, its application environment usually changes, and environmental change can influence the imaging of color camera,
Then face can not be well imaged when such as light is weaker.On the other hand, color camera None- identified is identified object
Whether it is real human face.
It is that can distinguish truth from false face also not by the face identification method and dress of ambient light interference that the present invention, which will be provided a kind of,
Put.
The recognition of face schematic diagram of a scenario according to an embodiment of the present invention shown in Fig. 1.The hand-held recognition of face of user 10
Device 11（Mobile terminal, such as mobile phone, flat board）, the inside of mobile terminal 11 preposition a projection module 111 and imaging mould
Group 112, when the user oriented head of mobile terminal 11 and after have activated recognition of face task, projection module 111 is thrown to user face
Penetrate structure light image（Such as speckle image 12）, imaging modules 112, which are used to gather in the image for including face, image, also includes scattered
Spot image 12.Process circuit (not marked in figure) is also configured with inside terminal 11, for realizing to containing speckle image 12
The processing of facial image.For face identification system, process circuit generally requires the following task of execution：Image preprocessing, face
Detection, face segmentation, feature extraction, recognition of face and task of correlation is performed according to recognition result, such as unlock, pay
Deng.The process circuit can be single special processor or multiple processor groups into, required execution task with
The form of software algorithm is written in process circuit and performed.Process circuit can also perform corresponding appoint according to current application
Business, such as the application of depth image is needed, then can perform the task of depth calculation.
In certain embodiments, face identification device 11 can also be fixed terminal device, such as computer, TV, machine
Top box, game machine, safety check gate etc..
In certain embodiments, face identification device 11 can also be separated multiple devices composition, such as by camera
（Include projection module 111 and imaging modules 112）And computing device composition, connect to pass between camera and computing device
Transmission of data, connected mode includes wired and wireless connection.Usually, camera is used for obtaining the structure light image of face, image
The corresponding task that further performed by the process circuit in computing device is transferred to after computing device by network connection.Can be with
Understand, on camera some process circuits can also be set to carry out executable portion task.
Fig. 2 is the structural representation of face identification device according to an embodiment of the invention.Project module 111 and include light
Source, lens and structure light maker（Such as diffraction optical element DOE）, it is collimated by or focuses on after source emissioning light beam, then
Outwards launch speckle image 12 after DOE beam splitting.Usually, light source is near-infrared laser, such as edge-emitting laser or
VCSEL lasers, sightless speckle image 12 can be outwards launched using near-infrared laser, thus will not to it is artificial into regarding
Feel interference, on the other hand, near-infrared laser is easily gathered by infrared imaging module 112.It is understood that light source can be used
Any suitable wavelength, it is without limitations herein.
Projection module 111 is connected with imaging modules 112 with the mainboard 115 in face identification device 11, is additionally useful for performing
The processor 113 of calculating task is connected also by mainboard with projection module and imaging modules 112.
In one embodiment, the Visible Light Camera for obtaining texture image is also provided in face identification device 11
114, such as RGB camera, gray scale camera etc..Visible Light Camera 114 can also unite two into one with imaging modules 112, i.e., in imaging
Imaging sensor (such as CMOS, CCD) each Pixel surface inside module 112 is respectively provided for the optical filtering that different wave length passes through
Piece, to gather structure light image and visible images respectively.
In one embodiment, the thermal infrared for obtaining target thermal infrared images is also provided in face identification device 11
Camera 114.
Fig. 3 be it is according to an embodiment of the invention utilize structure light image carry out recognition of face block diagram.Including following
Step.
In step 301, by projecting the projective structure light into the area of space comprising face of module 111, such as speckle 12.
In step 302, the speckle image for including face is gathered using imaging modules 112.
In step 303, according to the speckle image collected, the human face region image in speckle image is detected.
In step 304, recognition of face is carried out based on detected human face region image.
In step 303, the step of Face datection is based on directly on speckle image, and this is due to speckle image and other
Structure light image（Such as phase fringes, binary-coding）Compare, most information of face are all retained, change an angle
Say, speckle image is equal to visible ray gray level image plus some noises, therefore when carrying out Face datection, in one embodiment
First speckle image can be pre-processed, such as carry out noise remove etc. using morphological images processing method.For striped,
The structure light images such as binary-coding, when being projected to face, the face information more than half will be blocked by structure light image, and
It is blocked partially due to area is larger and continuously lead to not recover by image algorithm, and speckle image is although be covered in face
On, but because speckle particle is smaller, and discontinuously, larger distortion will not be caused to face texture.
Recognition of face task typically has face authentication to be identified with face, and face authentication refers to known current face and is present in face
In database, the task of face authentication is to identify that whom the face is；Face identification refer to do not know current face whether there is in
In face database, the task of face identification is to judge, and output is present and non-existent result.But either any side
Formula, recognition of face is inherently comprised the steps of：Feature extraction and characteristic matching.By based on detecting in step 304
Human face region image carries out recognition of face, main to include carrying out feature extraction to human face region speckle image, further utilizes
Feature carries out recognition of face.
Fig. 4 is according to an embodiment of the invention to utilize structure light image and visible images to carry out recognition of face step
Figure.Comprise the following steps.
In step 401, by projecting the projective structure light into the area of space comprising face of module 111, such as speckle 12.
In step 402, the speckle image for including face is gathered using imaging modules 112.
In step 403, the visible images for including face are gathered using Visible Light Camera 114.
In step 404, according to the speckle image and visible images collected, speckle image and visible images are detected
In human face region image.
In step 405, recognition of face is carried out based on detected human face region image.
Wherein, step 402 and step 403 can synchronously carry out, such as imaging modules 112 and visible ray are controlled by controller
Camera 114 synchronizes collection.Visible images can be coloured image, such as RGB image or gray level image, this
In visible images refer to reflection face textural characteristics and do not include the image of structure optical information.When projection module projection
When being also visible ray, influence is produced during in order to prevent that structure light from gathering on visible images, step 402 should stagger with step 403
Carry out, its sequencing can arbitrarily be set.
When carrying out human face region detection in step 404, pedestrian can be entered to speckle image and visible images respectively
Face is detected, only wherein piece image can also be detected, with reference to two cameras relative position relation so as to directly obtain
Human face region on another piece image, relative position relation is needed by being demarcated in advance.Usually, it is seen that the people of light image
Face detection tech more mature and reliable, therefore in one embodiment, by carrying out human face region detection to visible images, its
The secondary relative position relation according to testing result and two cameras obtains the human face region on speckle image.
In step 405, recognition of face can be carried out merely with the human face region image in speckle image, can also combined
Human face region image in visible images, to improve the accuracy of identification.
Fig. 5 is according to an embodiment of the invention to utilize structure light image and depth image to carry out recognition of face step
Figure.Comprise the following steps.
In step 501, by projecting the projective structure light into the area of space comprising face of module 111, such as speckle 12.
In step 502, the speckle image for including face is gathered using imaging modules 112.
In step 503, corresponding depth image is calculated using speckle image.
In step 504, according to obtained speckle image and depth image, speckle image and the people in depth image are detected
Face area image.
In step 505, recognition of face is carried out based on detected human face region image.
In step 503, the corresponding depth image of speckle image can be calculated based on structure light trigonometry, specifically,
By speckle image with carrying out the deviation value that matching primitives obtain each pixel with reference to speckle image, because deviation value is direct with depth
Correlation, therefore depth value can be calculated according to deviation value.
In step 504 carry out human face region detection when, due to speckle image and depth image be it is one-to-one, therefore
Only need to detect wherein piece image, usually, image segmentation is carried out in depth image to extract human face region more
For convenience, therefore in one embodiment, it is secondly direct according to testing result by carrying out human face region detection to depth image
Obtain the human face region on speckle image.
In step 505, recognition of face can be carried out merely with the human face region image in speckle image, can also combined
Human face region image in depth image, to improve the accuracy of identification.Using depth image another advantage is that, can sentence
Whether disconnected face is stereoscopic face, to prevent the possibility that can also be identified using planar picture, improves the peace of recognition of face
Quan Xing.If it should be noted that when carrying out three-dimensional detection using depth image, can not then be utilized to depth image in step 504
Image split realize human face region extract.
Fig. 6 is according to an embodiment of the invention to utilize structure light image and thermal infrared images to carry out recognition of face step
Figure.Comprise the following steps.
In step 601, by projecting the projective structure light into the area of space comprising face of module 111, such as speckle 12.
In step 602, the speckle image for including face is gathered using imaging modules 112.
In step 603, the thermal infrared images for including face is gathered using thermal infrared camera 114.
In step 604, according to obtained speckle image and thermal infrared images, detect in speckle image and depth image
Human face region image.
In step 605, recognition of face is carried out based on detected human face region image.
When carrying out human face region detection in step 604, first thermal infrared images can be detected, due to thermal infrared figure
As unique imaging characteristic, it can easily recognise that there is face according to thermal infrared images, or whether be real human face.
When detecting real human face, then into step 605, if not detecting face, or detect when being false face, then need not enter
Enter the identification step to face.Therefore, thermal infrared images assume responsibility for the task of In vivo detection herein.
In step 605, recognition of face can be carried out merely with the human face region image in speckle image, can also combined
Human face region image in thermal infrared images, to improve the accuracy of identification.
Above in several recognition of face embodiments, process substantially is only described, it is to be understood that by above mistake
The equivalent substitution of one or more of journey step, adjustment will be also dropped into protection scope of the present invention.Next to wherein
The Face datection that is related to, recognition of face are introduced.
Face datection.The main purpose of Face datection is detected in image with the presence or absence of the position where face and face
Put, Face datection algorithm mainly includes knowledge based rule, invariant features, template matches and statistical model totally four class method.
In the step of various embodiments above, it is related to speckle image, visible images, thermal infrared images and depth image, for
Different images carry out that during Face datection institute suitable algorithm should be selected.Such as visible images, it can use based on constant
Feature (such as colour of skin feature）Face datection algorithm；And for depth image, due to reflection be face three-dimensional information, because
This, is more applicable using the method matched based on three-dimensional template；For thermal infrared images, threshold can be easily passed through in general pattern
Value distinguishes face, therefore can carry out Face datection according to knowledge rule (threshold value)；, can essentially for speckle image
Regard as and some noises are added on ordinary gamma image, when carrying out Face datection, a kind of method is will by image procossing
Speckle carries out Face datection after being eliminated as much as, another scheme is directly to carry out Face datection using speckle image.
Because facial image often is influenceed to cause picture quality relatively low by different degrees of extraneous factor, now it is based on
The Face datection algorithm of statistical model can provide more accurate testing result.Say from the statistical significance, Face datection problem
It is that a grader problem, i.e. pixel on image are only possible to be two kinds of situations, one kind is face, and one kind is not face.Than
As Adaboost algorithm is had been demonstrated in terms of Face datection with very high verification and measurement ratio.
Face datection algorithm is varied, the method for the above by way of example only, any suitable algorithm can by with
To carry out Face datection.
Recognition of face.Detect after facial image, it is necessary to the facial image be identified, face recognition algorithms mainly have
Using the method based on outward appearance of overall textural characteristics, subspace method, neural network and based on shape and texture based on
Method of model etc..Different methods can be selected for different images, such as depth image preferably by based on
The face identification method of model.For visible images, earliest face identification method is the algorithm based on geometric properties, the calculation
Then method calculates the similarity between two kinds of feature by extracting images to be recognized and feature in template image, than
Similarity is such as weighed using minimum distance metric to realize recognition of face.Any suitable algorithm can be used into pedestrian
Face is recognized.
Carrying out Face datection and recognition of face constantly, thermal infrared images is but warm although possessing the function of vivo identification
Infrared camera cost is higher, and thermal infrared images can influence with many factors such as identified person's moods in addition, cause merely with heat
The effect that infrared image carries out recognition of face is undesirable.And when carrying out human face detection and tracing using visible images, on the one hand
Influenceed seriously by illumination etc., another aspect visible images hold due to being only capable of reflecting the two-dimensional signal of identified person's face
It is easily caused the hidden danger that can be also identified when with the two dimensional image of identified person as identified object.Carried out using depth image
During human face detection and tracing, vivo identification can be easily carried out, the feature yet with depth image is less, carry out face
Difficulty is larger when identification feature is compared.
The 2 d texture information of most faces is contained on speckle image, the speckle on image is then directly and people in addition
The three-dimensional information of face is related, therefore not only can be very good to carry out human face detection and tracing using the speckle image of face, may be used also
To determine whether live body.It should be noted that live body mentioned here judges that not individually carrying out an In vivo detection appoints
Business, but saying to have using speckle image effectively reduces the identification phase when measurand is non-three-dimensional real human body
Like degree.As an example it is assumed that current recognition of face task be judge current identified face whether with the people that is stored in system
Whether face is same face, the standard speckle image of testee's real human face is saved in memory first, then to quilt
Tester's face gathers current speckle image, finally carries out similarity identification to current speckle image and standard speckle image, from
And determine whether same face.If it is apparent that when gathering current speckle image, collected object is real human face, similar
The result of degree identification will be displayed as same face；If collected object is false photograph two-dimentional comprising testee's face,
Although there is testee's face identical 2 d texture information in the speckle image collected, but the speckle on speckle image
Reflection is plane information rather than steric information, therefore causes final similarity identification result to be non-same face.
Coloured image, gray level image, depth can be used for based on any by carrying out human face detection and tracing using speckle image
The detection of image and recognizer, in some algorithms based on machine learning, the Sample Storehouse for study is preferably by multiple
The speckle image composition of face, can also be made up of coloured image, gray level image or depth image in certain embodiments.
It fact proved, when speckle image is identified the model that the Sample Storehouse being made up of coloured image learns, can be good at
Avoid the wrong identification as caused by false 2-dimentional photo.
The density of speckle particle influences whether the performance of recognition of face in speckle image, will be blocked if speckle image is overstocked
More face texture informations, if speckle image is excessively sparse and can cause three-dimensional feature information that it is reflected very little.Therefore
The density of speckle image should be controlled in rational scope, i.e., will not block the main texture information of face too much（Such as eye
Eyeball, nose, face etc.）The three-dimensional feature of face can relatively accurately be reflected again.In certain embodiments, projection module is set
The speckle image of a variety of density can be projected by being set to, and when carrying out human face detection and tracing, then project low-density speckle image, when
When needing to carry out 3-D scanning task, then high density speckle image is projected out.In certain embodiments, projection module can be projected
Go out a variety of density and can meet the speckle image for not blocking the main texture information of face, the relatively low speckle image of density is used for people
When face is recognized, recognition speed is fast, but accuracy of identification is low, and then recognition speed is slow for the higher speckle image of density, but accuracy of identification
It is high.
In the face identification method corresponding to Fig. 4 ~ Fig. 6 or the recognition of face of the speckle image progress using different densities
In method, it is all based on what two kinds or more different images were identified, such be advantageous in that goes for more
Scene and raising discrimination.When two kinds or more of image is identified, typically there are two kinds of identification integration methods, Yi Zhongshi
Fusion based on decision-making, i.e., various images are identified respectively, are then merged recognition result to obtain final knowledge
Other result；Another is data fusion, image that will be two kinds or more directly as face identification system input, in face
During recognizer, the feature on various images is all as the foundation of final result.
Human face detection and tracing can be used in multiple-task, such as unblock, payment of smart machine etc..Usually,
The execution of task includes three steps：Mission-enabling, recognition of face and tasks carrying.
In certain embodiments, because the safe class of different tasks is different, if for the task of different safety class
It is clearly irrational using same face recognition scheme, for the relatively low task of safe class, such as unlock, can use
Relatively simple, quick face recognition scheme；And for the higher task of safe class, such as pay, then it is suitable using more
Complicated, accurate face recognition scheme.
Fig. 7 is the task executing method block diagram according to embodiments of the present invention based on recognition of face.Comprise the following steps.
In step 701, current task is activated.Activation can be carried out by various ways, such as button, inertia measurement equipment
(IMU) etc..In one embodiment, to be mobile device be locked to equipment to task by resting state solution opens, and activating the task can be with
Performed by some buttons, such as home keys, on ＆ off switch, volume key etc., internal IMU device, such as user can also be passed through
Picking up mobile device, the mobile corresponding acceleration of generation is obtained by IMU device rapidly（Such as user picks up movement from a certain place
Acceleration caused by equipment）, current task is activated when acceleration reaches a certain threshold value.In one embodiment, task is branch
The task of paying, activate the task directly can be performed by the virtual push button on related software, it is to be appreciated that activation task
Method can be by other any suitable modes.
In step 702, the safe class of current identification mission is judged.The peace to current task is needed after activation task
Congruent level is judged.In one embodiment, corresponding safe class is pre-set to various possible tasks, such as unblock is appointed
Business is that to open task be that safe class 2, payment task are safe class 3 for safe class 1, software, and safe class is higher to be meaned
The privacy of current task is higher, higher to the accuracy requirement of recognition of face.After activation, current task is pacified
The judgement of congruent level.
In step 703, face identification method corresponding with safe class is performed.Describe 4 kinds of faces altogether in Fig. 3 ~ Fig. 6
Recognition methods, the hardware and software algorithm that different recognition methods need is different, in one embodiment, if current face's identification dress
Above method can be performed by putting, and above method is classified according to the accuracy of algorithm, and by accuracy to difference
Method is classified and matched with safe class.The step in, phase is performed according to the safe class that is obtained in previous step
The face identification method answered, such as, for the minimum unblock task of level of security, perform face identification method as shown in Figure 3.
In certain embodiments, by projecting the speckle image of different densities come the task for different level of securitys, speckle image is close
The more high corresponding level of security of degree is also higher.It is understood that it is more than any two and different-effect face identification method
It can be used in the present invention with corresponding different safe class.
In step 704, corresponding task is performed according to face recognition result.Such as unblock task, work as recognition of face
When as a result showing identified object with the object of preservation in system for same people, perform corresponding instruction and unlocked.It can manage
Solution, recognition result generally comprises front and negative results, and different results should perform different tasks, or does not perform and appoint
What task.In certain embodiments, the result of recognition of face is except determining whether same people or belonging in standard personage storehouse
The situation such as a member outside, should also include the position of face and/or the distance recognized, only when face position and/or apart from up to
To can just perform corresponding task during preset value.
Different face identification devices does not cause face identification method that it can realize also not due to the difference of hardware configuration
Together, therefore in the above description, the face identification method performed by different safety class also can difference.Safe class
Quantity and the quantity of face identification method are also not necessarily the same, it is to be understood that the explanation scope of the claimed of the above will not
Limited to by this.Even if in addition, identical hardware configuration, can also set different recognizers with different safety of correspondence etc.
Level.
The safety applied in the present embodiment with system is classified, it is to be understood that other any classifications are all wrapped
Containing within the scope of the invention.
The device that face is identified above can also be used in the identification of other human body biological characteristics with scheme,
In some embodiments, it is possible to use structure light image is identified to human ear and further performs corresponding task.It will be situated between below
Continue a kind of call task that mobile terminal is performed using ear recognition.
Human ear is also that can distinguish the biological characteristic of identity, in some applications, especially for mobile communication terminal
It is finally to perform terminal in call task, existing method close to ear for call task, when terminal caller,
Call needs to answer by keys or buttons, but terminal is conversed close in one's ear.One will be provided in the present invention
More easily answering method is planted, i.e., when a call comes, is answered without carrying out keys or buttons, but directly by terminal close to people
Ear, and judge whether by the identification to human ear to answer.
When mobile communication terminal has incoming call, that is, activation call task is performed, and pop instruction indicates whether to answer.
In one embodiment, the safe class of the task of call is set to higher, i.e., the only owner of terminal or the one or two people specified can
To answer；In some embodiments, it is also possible to which the safe class of the task of call to be set to low, i.e., owner can answer.
Next, performing human ear identification method corresponding with safe class.The safety that can be answered for only one or two people
Grade, the detection that terminal performs to current human ear during close to human ear, constantly has been protected with recognizing and judging whether to belong to
One in the people deposited the not ear of individual, if i.e. task is answered in execution, if otherwise performing rejection task.For owner all
The safe class that can be answered, terminal performs the detection and identification to human ear, and determine whether during close to human ear
Human ear, if i.e. task is answered in execution.
In one embodiment, the detection to face or human ear also includes face or human ear with identification relative to identification terminal
Position or apart from etc. identification, i.e., not only to identify face or human ear, in addition it is also necessary to judge whether face or human ear are in conjunction
Behind suitable position, then perform the task of next step.In one embodiment, just open to answer when human ear sufficiently closes to terminal and appoint
Business, such as, this distance could be arranged within 5cm.
Using the identification to human ear except call task can be performed, other changes to the SOT state of termination are can also carry out,
Such as current state is to carry out the state put outside sound using loudspeaker, when terminal close to human ear and it is identified after by terminal shape
State become by receiver send sound only close to when the state that can just hear.
Above content is to combine specific preferred embodiment further description made for the present invention, it is impossible to assert
The specific implementation of the present invention is confined to these explanations.For those skilled in the art, do not taking off
On the premise of from present inventive concept, some equivalent substitutes or obvious modification can also be made, and performance or purposes are identical, all should
When being considered as belonging to protection scope of the present invention.
Claims (10)
1. a kind of utilization ear recognition changes the device of state, it is characterised in that including：Module is projected, for emitting structural light
The image extremely region comprising human ear；
Imaging modules, for gathering the structure light image for including human ear；Process circuit, is entered using the structure light image
Row ear detection and identification, and according to the ear detection and the state of recognition result change described device.
2. according to the method described in claim 1, it is characterised in that the structure light image includes infrared speckle image.
3. method according to claim 2, it is characterised in that the speckle particle density of the infrared speckle image is set
Not block the main textural characteristics of the human ear.
4. according to the method described in claim 1, it is characterised in that the ear detection is to utilize to be based on machine learning with identification
Detection and recognizer, the Sample Storehouse that model learning is used in the algorithm is made up of coloured image and/or gray level image.
5. according to the method described in claim 1, it is characterised in that the ear detection includes the position to the human ear with identification
Put and be identified with distance.
6. according to the method described in claim 1, it is characterised in that the ear detection includes judging that current human ear is with identification
It is no identical with target human ear.
7. according to the method described in claim 1, it is characterised in that described device includes communication device, the change dress
The state put changes into the state of answering including device origin electricity condition.
8. according to the method described in claim 1, it is characterised in that described device includes audio devices, the change dress
The state put puts sound status including device outside loudspeaker and changes into earpiece sound state.
9. a kind of method that utilization ear recognition changes state, it is characterised in that including：Using projecting module emitting structural light
The image extremely region comprising human ear；
The structure light image for including human ear is gathered using imaging modules；The structure light image is entered by process circuit
Row ear detection and identification, and according to the ear detection and the state of recognition result change described device.
10. method according to claim 9, it is characterised in that the structure light image includes infrared speckle image, described
The speckle particle density of infrared speckle image is arranged to not block the main textural characteristics of the human ear.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710567421.6A CN107229925A (en) | 2017-07-12 | 2017-07-12 | Conversed using ear recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710567421.6A CN107229925A (en) | 2017-07-12 | 2017-07-12 | Conversed using ear recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107229925A true CN107229925A (en) | 2017-10-03 |
Family
ID=59957495
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710567421.6A Pending CN107229925A (en) | 2017-07-12 | 2017-07-12 | Conversed using ear recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107229925A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107800963A (en) * | 2017-10-27 | 2018-03-13 | 广东欧珀移动通信有限公司 | Image processing method, device and electronic installation |
| US11315268B2 (en) | 2017-10-27 | 2022-04-26 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Image processing methods, image processing apparatuses and electronic devices |
| CN114500795A (en) * | 2021-12-27 | 2022-05-13 | 奥比中光科技集团股份有限公司 | Laser safety control method and device, intelligent door lock and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102622591A (en) * | 2012-01-12 | 2012-08-01 | 北京理工大学 | 3D (three-dimensional) human posture capturing and simulating system |
| CN105513221A (en) * | 2015-12-30 | 2016-04-20 | 四川川大智胜软件股份有限公司 | ATM (Automatic Teller Machine) cheat-proof device and system based on three-dimensional human face identification |
| CN106599779A (en) * | 2016-10-28 | 2017-04-26 | 黑龙江省科学院自动化研究所 | Human ear recognition method |
-
2017
- 2017-07-12 CN CN201710567421.6A patent/CN107229925A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102622591A (en) * | 2012-01-12 | 2012-08-01 | 北京理工大学 | 3D (three-dimensional) human posture capturing and simulating system |
| CN105513221A (en) * | 2015-12-30 | 2016-04-20 | 四川川大智胜软件股份有限公司 | ATM (Automatic Teller Machine) cheat-proof device and system based on three-dimensional human face identification |
| CN106599779A (en) * | 2016-10-28 | 2017-04-26 | 黑龙江省科学院自动化研究所 | Human ear recognition method |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107800963A (en) * | 2017-10-27 | 2018-03-13 | 广东欧珀移动通信有限公司 | Image processing method, device and electronic installation |
| CN107800963B (en) * | 2017-10-27 | 2019-08-30 | Oppo广东移动通信有限公司 | Image processing method, device, electronic device and computer readable storage medium |
| US11315268B2 (en) | 2017-10-27 | 2022-04-26 | Guangdong Oppo Mobile Telecommunications Corp., Ltd. | Image processing methods, image processing apparatuses and electronic devices |
| CN114500795A (en) * | 2021-12-27 | 2022-05-13 | 奥比中光科技集团股份有限公司 | Laser safety control method and device, intelligent door lock and storage medium |
| CN114500795B (en) * | 2021-12-27 | 2024-03-15 | 奥比中光科技集团股份有限公司 | Laser safety control method and device, intelligent door lock and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107169483A (en) | Tasks carrying based on recognition of face | |
| CN107341481A (en) | It is identified using structure light image | |
| CN107292283A (en) | Mix face identification method | |
| US11238270B2 (en) | 3D face identity authentication method and apparatus | |
| CN107609383B (en) | 3D face identity authentication method and device | |
| CN107633165B (en) | 3D face identity authentication method and device | |
| CN106372601B (en) | Living body detection method and device based on infrared visible binocular images | |
| CN105874473A (en) | Apparatus and method for acquiring image for iris recognition using distance of facial feature | |
| CN111368811B (en) | Living body detection method, living body detection device, living body detection equipment and storage medium | |
| KR101919090B1 (en) | Apparatus and method of face recognition verifying liveness based on 3d depth information and ir information | |
| CN104598882A (en) | Method and system of spoofing detection for biometric authentication | |
| CN107563304A (en) | Unlocking terminal equipment method and device, terminal device | |
| CN112232155B (en) | Non-contact fingerprint identification method and device, terminal and storage medium | |
| KR20180134280A (en) | Apparatus and method of face recognition verifying liveness based on 3d depth information and ir information | |
| CN110383289A (en) | Device, method and the electronic equipment of recognition of face | |
| CN110008813A (en) | Face identification method and system based on In vivo detection technology | |
| CN110866454B (en) | Face living body detection method and system and computer readable storage medium | |
| KR101640014B1 (en) | Iris recognition apparatus for detecting false face image | |
| KR20210062381A (en) | Liveness test method and liveness test apparatus, biometrics authentication method and biometrics authentication apparatus | |
| CN107229925A (en) | Conversed using ear recognition | |
| CN109902604B (en) | High-safety face comparison system and method based on Feiteng platform | |
| CN111104833A (en) | Method and apparatus for in vivo examination, storage medium, and electronic device | |
| US20210256244A1 (en) | Method for authentication or identification of an individual | |
| CN112818722A (en) | Modular dynamically configurable living body face recognition system | |
| CN111144169A (en) | Face recognition method and device and electronic equipment |
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 | ||
| CB02 | Change of applicant information |
Address after: 518057 11-13 / F, joint headquarters building, high tech Zone, No.63 Xuefu Road, Yuehai street, Nanshan District, Shenzhen City, Guangdong Province Applicant after: Obi Zhongguang Technology Group Co., Ltd Address before: 518057 11-13 / F, joint headquarters building, high tech Zone, No.63 Xuefu Road, Yuehai street, Nanshan District, Shenzhen City, Guangdong Province Applicant before: SHENZHEN ORBBEC Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171003 |
|
| RJ01 | Rejection of invention patent application after publication |