CN106897657A - A kind of human face in-vivo detection method and device - Google Patents
A kind of human face in-vivo detection method and device Download PDFInfo
- Publication number
- CN106897657A CN106897657A CN201510960637.XA CN201510960637A CN106897657A CN 106897657 A CN106897657 A CN 106897657A CN 201510960637 A CN201510960637 A CN 201510960637A CN 106897657 A CN106897657 A CN 106897657A
- Authority
- CN
- China
- Prior art keywords
- sequence
- detection
- testing result
- detected user
- face
- 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
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
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/40—Spoof detection, e.g. liveness detection
- G06V40/45—Detection of the body part being alive
Abstract
A kind of human face in-vivo detection method and device are the embodiment of the invention provides, method therein includes：The detection sequence of random generation predetermined bit length, wherein, the predetermined bit length value is identical with default detection time numerical value；Detection sequence according to generation performs the detected user characteristics collection of preset times time, and the detected user characteristics is the detected user's face provincial characteristics；Testing result generation testing result sequence according to the detected user characteristics for collecting, wherein, the one digit number value in a testing result correspondence testing result sequence；Determine the fuzzy matching degree of detection sequence and testing result sequence；When the fuzzy matching degree of detection sequence and testing result sequence is more than the first predetermined threshold value, determine that the detected user is live body.The embodiment of the present invention can improve the safety and reliability of face In vivo detection system.
Description
Technical field
The present invention relates to technical field of face recognition, and in particular to a kind of human face in-vivo detection method and device.
Background technology
Recognition of face as living things feature recognition authentication important means, using more and more extensive.For example
Card is opened in the remote human face identification authentication of the social security recognition of face authentication system, financial field in portion of people society
System, recognition of face student status Verification System of education sector etc..
As face identification system application is more and more extensive, the counterfeit behavior to validated user face grows in intensity.
Because face is easily replicated with modes such as photo, videos, therefore it is people to the personation of validated user face
The important threat of face identification and authentication system safety.Therefore, face In vivo detection technology is arisen at the historic moment, face
In vivo detection briefly, exactly determines whether the biological characteristic for detecting is direct using human face detection tech
From the process of the live body corresponding to the biological characteristic.
In the last few years, face In vivo detection technology had made some progress, but existing face In vivo detection side
Method need to be improved in security and reliability.
The content of the invention
In order to solve the problems, such as face In vivo detection scheme in the prior art it cannot be guaranteed that safety and reliability,
The embodiment of the present invention is expected to provide a kind of human face in-vivo detection method and device.
A kind of human face in-vivo detection method is the embodiment of the invention provides, including：
The detection sequence of random generation predetermined bit length, wherein, the predetermined bit length value and default inspection
Survey time numerical value identical；
Detection sequence according to generation performs the detected user characteristics collection of preset times time, the detected use
Family is characterized as the detected user's face provincial characteristics；
Testing result generation testing result sequence according to the detected user characteristics for collecting, wherein, one
One digit number value in testing result correspondence testing result sequence；
Determine the fuzzy matching degree of detection sequence and testing result sequence；
When the fuzzy matching degree of detection sequence and testing result sequence is more than the first predetermined threshold value, it is determined that described
It is live body to be detected user.
Preferably, the detection sequence of the random generation predetermined bit length is comprised the following steps：
Random formation sequence；
It is determined that the sequence of random generation and the last detection sequence for carrying out face In vivo detection are matched
Degree；
Judge the sequence of random generation and the matching degree of the last detection sequence for carrying out face In vivo detection
Whether the second predetermined threshold value is less than；
When the sequence of random generation is small with the matching degree of the last detection sequence for carrying out face In vivo detection
When the second predetermined threshold value, the sequence of random generation is defined as currently carrying out the inspection of face In vivo detection
Sequencing row；Otherwise, the step of re-executing the random formation sequence.
Preferably, the detection sequence according to generation performs the detected user characteristics collection of preset times time,
Comprise the following steps：
Order reads the current bit value of the detection sequence；
It is determined that the facial physiological motion corresponding to the current numerical value for reading, wherein, a kind of numerical value correspondence is a kind of
Facial physiological motion；
Indicate to be detected user's execution facial physiological motion；
The detected user characteristics is gathered in Preset Time；
The detection sequence numerical value not read is judged whether, if it is, re-executing the order reads institute
The step of stating detection sequence current bit value；If not, terminating to be detected user characteristics collection.
Preferably, the testing result generation testing result sequence of the detected user characteristics that the basis is collected,
Including：
Determine to be detected user whether according to the instruction execution facial physiology according to the detected user characteristics
Property motion；
If it is, the numerical value corresponding with the facial physiological motion is defined as into working as testing result sequence
Preceding bit value；
If not, being defined as detection knot by the different numerical value of corresponding numerical value is moved from the facial physiological
The current bit value of infructescence row.
Preferably, the detection sequence and testing result sequence are 01 sequence；
It is relative with the facial physiological motion when it is 0 that the facial physiological moves corresponding numerical value
The different numerical value of the numerical value answered is 1；
It is relative with the facial physiological motion when it is 1 that the facial physiological moves corresponding numerical value
The different numerical value of the numerical value answered is 0.
Preferably, before the fuzzy matching degree for determining detection sequence and testing result sequence, methods described
Also include：
Determine that the detected user characteristics is not that video is attacked.
Preferably, it is described to determine that the detected user characteristics is not that video is attacked, including：
Extract the face detection region of the detected user characteristics；
Determine the aspect ratio value in the face detection region；
Judge the aspect ratio value whether in the 3rd threshold range；
If it is, determining that the detected user characteristics is not that video is attacked.
A kind of face living body detection device is the embodiment of the invention provides, including：
First generation module, acquisition module, the second generation module, the first determining module, judge module and
Two determining modules；Wherein,
First generation module, for the detection sequence of random generation predetermined bit length, wherein, it is described
Predetermined bit length value is identical with default detection time numerical value；
The acquisition module, the detected user characteristics of preset times time is performed for the detection sequence according to generation
Collection, the detected user characteristics is the detected user's face provincial characteristics；
Second generation module, inspection is generated for the testing result according to the detected user characteristics for collecting
Result sequence is surveyed, wherein, the one digit number value in a testing result correspondence testing result sequence；
First determining module, the fuzzy matching degree for determining detection sequence and testing result sequence；
The judge module, for judging whether detection sequence is more than with the fuzzy matching degree of testing result sequence
First predetermined threshold value；
Second determining module, for the fuzzy matching degree when detection sequence and testing result sequence more than the
During one predetermined threshold value, determine that the detected user is live body.
Preferably, first generation module, including：
Sequence generating unit, for random formation sequence；
First determining unit, the sequence for determining random generation carries out face In vivo detection with the last time
Detection sequence matching degree；
First judging unit, the sequence for judging random generation carries out face In vivo detection with the last time
Detection sequence matching degree whether be less than the second predetermined threshold value；
Second determining unit, the sequence and last time for that ought generate at random carry out face In vivo detection
When the matching degree of detection sequence is less than the second predetermined threshold value, the sequence of random generation is defined as when advance
The detection sequence of pedestrian's face In vivo detection simultaneously terminates random formation sequence；It is additionally operable to the sequence that ought be generated at random
When row are less than or equal to the second predetermined threshold value with the matching degree of the last detection sequence for carrying out face In vivo detection,
Go to sequence generating unit.
Preferably, the acquisition module includes：
Reading unit, for sequentially reading the current bit value of the detection sequence；
3rd determining unit, for determining the facial physiological motion corresponding to the current numerical value for reading, wherein,
A kind of a kind of facial physiological motion of numerical value correspondence；
Indicating member, for indicating to be detected user's execution facial physiological motion；
Collecting unit, user characteristics is detected for being gathered in Preset Time；
Second judging unit, at the end of Preset Time, judging whether the detection sequence not read
Numerical value, if it is, reading unit is gone to, if not, terminating to be detected user characteristics collection.
Preferably, second generation module, including：
3rd judging unit, for judging to be detected user whether according to finger according to the detected user characteristics
Show the execution facial physiological motion；
4th determining unit, for the 3rd judging unit judged result for be when, will be with the facial physiology
Property the corresponding numerical value of motion be defined as the current bit value of testing result sequence；It is additionally operable to judge single the 3rd
When first judged result is no, it is defined as the different numerical value of corresponding numerical value is moved from the facial physiological
The current bit value of testing result sequence.
Preferably, the detection sequence and testing result sequence are 01 sequence；
4th determining unit, for when it is 0 that the facial physiological moves corresponding numerical value, inciting somebody to action
The different numerical value of corresponding numerical value is moved from the facial physiological be defined as 1；It is additionally operable to when the face
When the corresponding numerical value of physiological motion is 1, corresponding numerical value will be moved from the facial physiological different
Numerical value be defined as 0.
Preferably, described device also includes：3rd determining module, it is described for determining in the second determining module
Before the fuzzy matching degree of detection sequence and testing result sequence, determine that the detected user characteristics is not regarded
Frequency is attacked.
Preferably, the 3rd determining module, including：
Extraction unit, the face detection region for extracting the detected user characteristics；
5th determining unit, the aspect ratio value for determining the face detection region；
4th judging unit, for judging the aspect ratio value whether in the 3rd preset threshold range；
6th determining unit, for when the 4th judging unit judged result is to be, determining the detected use
Family feature is not that video is attacked.
Compared with prior art, the embodiment of the present invention includes advantages below：
A kind of human face in-vivo detection method and device that the embodiment of the present invention is provided, generate predetermined bit at random
The detection sequence of length, wherein, the predetermined bit length value is identical with default detection time numerical value；According to life
Into detection sequence perform time detected user characteristics collection of preset times, the detected user characteristics is institute
State detected user's face provincial characteristics；Testing result generation inspection according to the detected user characteristics for collecting
Result sequence is surveyed, wherein, the one digit number value in a testing result correspondence testing result sequence；It is determined that detection
The fuzzy matching degree of sequence and testing result sequence；When detection sequence and the fuzzy matching degree of testing result sequence
During more than the first predetermined threshold value, determine that the detected user is live body.On the one hand, in this way,
When face In vivo detection is performed, the detection sequence according to random generation performs the detection process of face live body,
It is randomly generated due to detection sequence, detection sequence when therefore, it can reduce face In vivo detection each time
Repeatability, also avoid the repeatability of face In vivo detection process performed each time.So, may be used
To increase forgery face In vivo detection result during other users illegal invasion face biopsy system, for example, adopt
The difficulty of face detection system is escaped with video attack pattern, so as to improve the safe and reliable of human body In vivo detection
Property.On the other hand determine to be detected whether user is live body by fuzzy match mode, i.e. it is determined that detection
The fuzzy matching degree of sequence and testing result sequence, when detection sequence and the fuzzy matching degree of testing result sequence
During more than the first predetermined threshold value, it is determined that it is live body to be detected user, so, the method relative to matching completely,
The stability of face In vivo detection system can be strengthened, because, face In vivo detection system is generally first given
Go out detected user to indicate, afterwards, be detected user and just make a response, the reaction of detected user relative to
The acquisition time of system can time delay, if requiring that detection sequence and testing result sequence are matched completely, have
Certain difficulty, and fuzzy matching is at this moment used, the stability of a system can be strengthened on the contrary.
Brief description of the drawings
Fig. 1 is human face in-vivo detection method flow chart one provided in an embodiment of the present invention；
Fig. 2 is human face in-vivo detection method flowchart 2 provided in an embodiment of the present invention；
Fig. 3 is that the detection sequence in human face in-vivo detection method provided in an embodiment of the present invention according to generation is performed
The method flow diagram of the detected user characteristics collection of preset times time；
Fig. 4 is that pitch rotates schematic diagram in the prior art；
Fig. 5 is the basic block diagram one of face living body detection device provided in an embodiment of the present invention；
Fig. 6 is the basic block diagram two of face In vivo detection system provided in an embodiment of the present invention.
Specific embodiment
Embodiment one
Reference picture 1, flow chart the step of show a kind of human face in-vivo detection method embodiment one of the invention,
Can specifically include：
The detection sequence of step 101, random generation predetermined bit length, wherein, the predetermined bit length
Value is identical with default detection time numerical value；
The step in, by face In vivo detection system generate predetermined bit length detection sequence, preset ratio
Length value is determined by default detection time numerical value, it is, default detection time numerical value is how many, predetermined bit
Length value is just how many.For example, when default detection number of times is seven times, i.e., face In vivo detection system is performed
Seven face In vivo detections, then need the detection sequence of generation seven, afterwards, face In vivo detection system root
Face In vivo detection process is performed according to the detection sequence.
Here, due to performing each time during face In vivo detection, what detection sequence was randomly generated, therefore,
The repeatability of detection sequence when can reduce face In vivo detection each time, also avoids performed each time
Face In vivo detection process repeatability.So, the inspection of other users illegal invasion face live body can be increased
Face In vivo detection result is forged when looking into system, face detection system is escaped for example with video attack pattern
Difficulty, so as to improve the security reliability of human body In vivo detection.
Step 102, the detected user characteristics collection of detection sequence execution preset times time according to generation, institute
Detected user characteristics is stated for the detected user's face provincial characteristics；
Face In vivo detection system requirements is detected user and can cooperate with face's life that the system completes regulation on one's own initiative
Rationality is moved.
The step in, it is tested that face In vivo detection system performs preset times time according to the detection sequence of generation
Survey user characteristics collection.Specifically, face In vivo detection system needs to have been distinguished according to the detection sequence
Gathered into user characteristics is detected each time；More specifically, face In vivo detection system is needed successively according to inspection
Sequencing row in each sequential value come determine when time need detection be detected user any facial physiological fortune
It is dynamic.
Step 103, the testing result generation testing result sequence according to the detected user characteristics for collecting,
Wherein, the one digit number value in a testing result correspondence testing result sequence；
Specifically, it is necessary to judge the detected user for gathering each time after collecting detected user characteristics
The testing result of feature, it is, judging the correct still mistake of the testing result for being detected user characteristics.Cause
This, testing result has two kinds, i.e. correct and mistake, therefore, detection knot is generated according to testing result
Infructescence is arranged, i.e. the testing result for detecting each time is represented with each sequential value.
Step 104, the fuzzy matching degree for determining detection sequence and testing result sequence；
The step in, the fuzzy matching degree for determining detection sequence and testing result sequence is to determine detection
The accuracy of result.Specifically, face In vivo detection system is according to quilt in the detected user characteristics for collecting
Detect that the facial physiological motion characteristics value of user is compared with default facial physiological motion characteristics value
Compared with calculating similarity, when similarity reaches default similarity threshold, it is determined that being detected user when time detection
Testing result be correct；Otherwise, it determines time testing result of working as being detected user is mistake.
As an example it is assumed that face In vivo detection system performs seven detections, this seven times detections detect quilt respectively
Detection user's dehisces and reaction of remaining silent, and detection sequence is 0100110, wherein, 0 represents dehisce detection, 1
Representative is remained silent detection, then face In vivo detection system is successively read the detection sequence, and comes according to sequential value
Indicate to be detected the corresponding face physiological motion of user's execution.Indicate to be detected when the numerical value for reading is 1
User remains silent, and when the numerical value for reading is 0, indicates detected user to dehisce, and accordingly, is detected user
Facial physiological motion according to indicated by indicating to complete；Face In vivo detection system judges to be detected user institute
Whether the facial physiological motion of completion is correct, draws testing result, when testing result is correct, input
The corresponding value of the face degree physiological motion, i.e., when the detected user of instruction completes to dehisce detection, detection
Result correctly then exports 0, when the detected user of instruction completes to remain silent detection, when testing result is correct then defeated
Go out 1；Accordingly, when testing result is incorrect, opposite value is exported, i.e. when instruction is detected user
When completion dehisces to detect, testing result mistake then exports 1, when the detected user of instruction completes to remain silent detection,
When testing result mistake then exports 0.So, directly detection sequence and testing result sequence are matched,
Determine its fuzzy matching degree, you can determine the accuracy of testing result.
Step 105, when detection sequence and testing result sequence fuzzy matching degree be more than the first predetermined threshold value when,
Determine that the detected user is live body.
The fuzzy matching degree of detection sequence and testing result sequence is pre-set, because, because face is lived
Physical examination examining system generally first provides detected user and indicates, and afterwards, is detected user and just makes a response, and is detected
Survey user reaction relative to system acquisition time can time delay, if require detection sequence and detection tie
Infructescence row are matched completely, there is certain difficulty, and at this moment use fuzzy matching, and the system can be strengthened on the contrary
Stability, by presetting suitable fuzzy matching degree threshold value (that is, the first predetermined threshold value), every detection sequence
The fuzzy matching degree of row and testing result sequence is more than this threshold value, it is determined that it is live body to be detected user.
Accordingly, when the fuzzy matching degree of detection sequence and testing result sequence is not more than the first predetermined threshold value,
Determine detected user's non-living body.
To sum up, the human face in-vivo detection method that the embodiment of the present invention one is provided, random generation predetermined bit length
Detection sequence, wherein, the predetermined bit length value is identical with default detection time numerical value；According to generation
Detection sequence performs the detected user characteristics collection of preset times time, and the detected user characteristics is the quilt
Detection user's face provincial characteristics；Testing result generation detection knot according to the detected user characteristics for collecting
Infructescence is arranged, wherein, the one digit number value in a testing result correspondence testing result sequence；Determine detection sequence
With the fuzzy matching degree of testing result sequence；When detection sequence is more than with the fuzzy matching degree of testing result sequence
During the first predetermined threshold value, determine that the detected user is live body.On the one hand, in this way, holding
During pedestrian's face In vivo detection, the detection sequence according to random generation performs the detection process of face live body, due to
What detection sequence was randomly generated, the weight of detection sequence when therefore, it can reduce face In vivo detection each time
Renaturation, also avoids the repeatability of face In vivo detection process performed each time.So, Ke Yizeng
Plus face In vivo detection result is forged during other users illegal invasion face biopsy system, for example with regarding
Frequency attack pattern escapes the difficulty of face detection system, so as to improve the security reliability of human body In vivo detection.
On the other hand determine to be detected whether user is live body by fuzzy match mode, i.e. determine detection sequence
With the fuzzy matching degree of testing result sequence, when detection sequence is more than with the fuzzy matching degree of testing result sequence
During the first predetermined threshold value, it is determined that it is live body to be detected user, so, the method relative to matching completely can
To strengthen the stability of face In vivo detection system, because, face In vivo detection system is generally first given
It is detected user to indicate, afterwards, is detected user and just makes a response, the reaction of detected user is relative to being
The acquisition time of system can time delay, if requiring that detection sequence and testing result sequence are matched completely, in reality
Border realizes there is certain difficulty, and at this moment uses fuzzy matching, and the stability of a system can be strengthened on the contrary.
Embodiment two
Reference picture 2, flow chart the step of show a kind of human face in-vivo detection method embodiment of the invention,
The method can specifically include：
Step 201, random generation predetermined bit length sequences, wherein, the predetermined bit length value with it is pre-
If detection time numerical value is identical；
Specifically, when current face's vivo identification is carried out, face In vivo detection system can utilize stochastic ordering
Column-generation device generates predetermined bit length sequences at random, and the length of the predetermined bit length sequences is by default inspection
Time numerical value is surveyed to determine；Specifically, default detection time numerical value is how many, it is how many to be generated as bit long angle value
Sequence.
Step 202, the sequence for determining random generation and the last detection sequence for carrying out face In vivo detection
The matching degree of row；
After random formation sequence, determine the sequence with the last detection sequence for carrying out face In vivo detection
Matching degree.
Step 203, the sequence for judging random generation and the last detection sequence for carrying out face In vivo detection
Whether the matching degree of row is less than the second predetermined threshold value；
Step 204, the sequence and the last detection sequence for carrying out face In vivo detection when random generation
Matching degree when being less than the second predetermined threshold value, the sequence of random generation is defined as currently carrying out face work
The detection sequence that physical examination is surveyed；
The step in, when sequence and the last detection sequence for carrying out face In vivo detection of random generation
When matching degree is less than the second predetermined threshold value, the sequence is defined as currently carrying out the detection of face In vivo detection
Sequence, thus it is ensured that be used to be differed greatly between the detection sequence of face In vivo detection each time, so that plus
Face In vivo detection result is forged during big other users illegal invasion face biopsy system, for example with
Video attack pattern escapes the difficulty of face detection system, greatly improves the security reliability of human body In vivo detection.
It refers to that the picture that face is carried out specific physiological motion by lawless person is autotelic that video mentioned here is attacked
Video is organized into, so as to hide the means of face In vivo detection system.
Step 205, the detected user characteristics collection of detection sequence execution preset times time according to generation, institute
Detected user characteristics is stated for the detected user's face provincial characteristics；
Specifically, as shown in figure 3, the detection sequence according to generation performs preset times time is detected use
Family collection apparatus, comprise the following steps：
S301, sequentially read the current bit value of the detection sequence；
Facial physiological motion corresponding to S302, the current numerical value for reading of determination, wherein, a kind of numerical value pair
A kind of facial physiological is answered to move；
S303, instruction are detected user and perform the facial physiological motion；
S304, the detected user characteristics is gathered in Preset Time；
S305, the detection sequence numerical value that does not read is judged whether, if it is, step S301 is gone to, such as
It is really no, terminate the current flow for performing and being detected user characteristics collection；
Specifically, at the end of default collection is detected the time of user characteristics, judging whether not reading
Detection sequence numerical value.
Step 206, the testing result generation testing result sequence according to the detected user characteristics for collecting,
Wherein, the one digit number value in a testing result correspondence testing result sequence；
Specifically, the testing result generation testing result sequence of the detected user characteristics that the basis is collected,
Including：
Determine to be detected user whether according to the instruction execution facial physiology according to the detected user characteristics
Property motion；
If it is, the numerical value corresponding with the facial physiological motion is defined as into working as testing result sequence
Preceding bit value；
If not, being defined as detection knot by the different numerical value of corresponding numerical value is moved from the facial physiological
The current bit value of infructescence row.
As an example it is assumed that detection sequence is 23233323, wherein, the facial physiological motion corresponding to 2
For eyes open, 3 corresponding facial physiological motions are eyes closed；When to be detected user carry out face
During In vivo detection, it is assumed that third time and the 5th physiological motion error in detection performed by user, then may be used
So that the numerical value corresponding to third time testing result in testing result sequence to be defined as any numerical value beyond 2,
Such as 5, similar is defined as beyond 3 the numerical value corresponding to the 5th testing result in testing result sequence
Any numerical value, such as 2；The testing result sequence for then exporting is 23532323.
Specifically, the detection sequence and testing result sequence can be binary sequence, e.g., typical two-value
Sequence：01 sequence；
So, when it is 0 that the facial physiological moves corresponding numerical value, transported with the facial physiological
The different numerical value of dynamic corresponding numerical value is 1；When it is 1 that the facial physiological moves corresponding numerical value,
It is 0 to move the different numerical value of corresponding numerical value from the facial physiological.
In a kind of preferred embodiment, using 01 sequence as detection sequence, and the corresponding face of 0 and 1 difference
Physiological motion is the opening and closing of face, and because detection sequence only has two kinds of values, implementation complexity is relative
It is relatively low；And, face is that comparatively face compares a flexible position, and it opens or closes at sentences
When other comparatively, accuracy rate is high；In addition, the opening and closing of face the two action either to youth
All it is very easily, in actual realization, to be easy to for people or the more sluggish the elderly of reaction and children
Operation.
Step 207, the fuzzy matching degree for determining detection sequence and testing result sequence；
Specifically, in practical implementations, determine the fuzzy matching degree of detection sequence and testing result at least with
Lower two methods：
First method：Detection sequence and testing result sequence are directly matched in order, is obscured
Matching degree；
By taking the specific example in step 206 as an example, when detection sequence is that 23233323, testing result sequence is
When 23532323, can directly determine that its fuzzy matching degree is 75%, at this point, it is assumed that fuzzy matching degree threshold value
It is 87%, then the face In vivo detection does not pass through, determines that the detected user is not live body, it is possible to
It is that video is attacked；
Second method：The maximum length sequence fragment matched with detection sequence is searched in testing result sequence,
That is, the maximum length sequence fragment continuously matched with detection sequence in testing result sequence is extracted；In matching process
In, detection sequence can be kept motionless, line displacement is entered into testing result sequence step-by-step, but due to testing result
The hysteresis quality of time of occurrence, usually forward migration；
As an example it is assumed that testing result sequence is 00111100, detection sequence is 00011000, then in inspection
It is 1100 to survey the maximum length sequence matched with detection sequence extracted in result sequence, that is, have continuous 4bit
Sequence is matched completely with detection sequence, then be 50% according to this calculating fuzzy matching degree, it is assumed that fuzzy matching degree
Threshold value is 75%, then the face In vivo detection does not pass through, and it is not live body to determine the detected user, is had
It is probably that video is attacked.
Accordingly, because above two method stresses to have nothing in common with each other, therefore, fuzzy matching degree threshold value is being set
When, can be configured according to actual conditions, choose an optimal value, it is ensured that detection is safe and reliable.
In another alternative embodiment of the invention, the mould for determining detection sequence and testing result sequence
Before paste matching degree, methods described also includes：
Determine that the detected user characteristics is not that video is attacked.
Specifically, described determine that the detected user characteristics is not that video is attacked, including：
Extract the face detection region of the detected user characteristics；
Determine the aspect ratio value in the face detection region；
Judge the aspect ratio value whether in the 3rd threshold range；
If it is, determining that the detected user characteristics is not that video is attacked.
When it is determined that the detected user characteristics is video attack, it is not necessary to perform follow-up step again, can
Directly to determine detected user's non-living body.
Have already been mentioned, video attack refers to the figure that face is carried out lawless person specific physiological motion
Piece is autotelic to be organized into video, so as to hide the means of face In vivo detection system.Generally, regarded
When frequency is attacked, for certain some face's physiological feature, for example, dehisce, remain silent, or, eyes open,
Eyes closed, disabled user can be rotated by carrying out pitching as shown in Figure 4 to picture to (pitch),
The shape of the mouth as one speaks that can cause to dehisce is mistaken as the shape of the mouth as one speaks of remaining silent, as shown in figure 4, pitch rotations refer to x by picture
Rotated along pitch directions centered on axle.But, for by the postrotational pictures of pitch, in picture
The aspect ratio value in Face datection region will have greatly changed, therefore, in this step, can be by setting
Suitable 3rd threshold value is put, quickly to find that video is attacked.
Step 208, when detection sequence and testing result sequence fuzzy matching degree be more than the first predetermined threshold value when,
Determine that the detected user is live body.
The fuzzy matching degree of detection sequence and testing result sequence is pre-set, because, because face is lived
Physical examination examining system generally first provides detected user and indicates, and afterwards, is detected user and just makes a response, and is detected
Survey user reaction relative to system acquisition time can time delay, if require detection sequence and detection tie
Infructescence row are matched completely, there is certain difficulty, and at this moment use fuzzy matching, and the system can be strengthened on the contrary
Stability, by presetting suitable fuzzy matching degree threshold value (that is, the first predetermined threshold value), every detection sequence
The fuzzy matching degree of row and testing result sequence is more than this threshold value, it is determined that it is live body to be detected user.
Accordingly, when the fuzzy matching degree of detection sequence and testing result sequence is more than the first predetermined threshold value,
Determine detected user's non-living body.
It can be seen that, the human face in-vivo detection method provided using the embodiment of the present invention two generates predetermined bit at random
Length sequences, wherein, the predetermined bit length value is identical with default detection time numerical value, it is determined that random generation
The sequence and the matching degree of the last detection sequence for carrying out face In vivo detection, afterwards, judge random
Whether the sequence of generation is less than second with the matching degree of the last detection sequence for carrying out face In vivo detection
Predetermined threshold value, when random generation the sequence and the last detection sequence for carrying out face In vivo detection
During with degree less than the second predetermined threshold value, the sequence of random generation is defined as currently carrying out face live body inspection
The detection sequence of survey, compared to the scheme that embodiment one is provided, this mode can further ensure that each time
For being differed greatly between the detection sequence of face In vivo detection, so as to increase other users illegal invasion face
Face In vivo detection result is forged during biopsy system, Face datection is escaped for example with video attack pattern
The difficulty of system, greatly improves the security reliability of human body In vivo detection；In addition, passing through fuzzy match mode
It is determined that being detected whether user is live body, i.e. determine the fuzzy matching of detection sequence and testing result sequence
Degree, when the fuzzy matching degree of detection sequence and testing result sequence is more than the first predetermined threshold value, it is determined that tested
Survey user is live body, and so, the method relative to matching completely in the prior art can strengthen face live body
The stability of detecting system, because, face In vivo detection system generally first provides detected user and indicates,
Afterwards, it is detected user just to make a response, the reaction for being detected user has relative to the acquisition time of system
Institute's time delay, if requiring that detection sequence and testing result sequence are matched completely, there is certain difficulty, and at this moment
Using fuzzy matching, the stability of a system can be strengthened on the contrary.
Device embodiment one
Reference picture 5, shows a kind of face living body detection device of the invention, positioned at face In vivo detection system
On；Wherein, described device includes：First generation module 51, acquisition module 52, the second generation module 53,
First determining module 54, the determining module 56 of judge module 55 and second；Wherein,
First generation module 51, for the detection sequence of random generation predetermined bit length, wherein, institute
State predetermined bit length value identical with default detection time numerical value；
The acquisition module 52, performs the detected user of preset times time special for the detection sequence according to generation
Collection is levied, the detected user characteristics is the detected user's face provincial characteristics；
Second generation module 53, for the testing result generation according to the detected user characteristics for collecting
Testing result sequence, wherein, the one digit number value in a testing result correspondence testing result sequence；
First determining module 54, the fuzzy matching degree for determining detection sequence and testing result sequence；
Whether the judge module 55 is big with the fuzzy matching degree of testing result sequence for judging detection sequence
In the first predetermined threshold value；
Second determining module 56, for being more than with the fuzzy matching degree of testing result sequence when detection sequence
During the first predetermined threshold value, determine that the detected user is live body.
Specifically, first generation module 51, including：
Sequence generating unit, for random formation sequence；
First determining unit, the sequence for determining random generation carries out face In vivo detection with the last time
Detection sequence matching degree；
First judging unit, the sequence for judging random generation carries out face In vivo detection with the last time
Detection sequence matching degree whether be less than the second predetermined threshold value；
Second determining unit, the sequence and last time for that ought generate at random carry out face In vivo detection
When the matching degree of detection sequence is less than the second predetermined threshold value, the sequence of random generation is defined as when advance
The detection sequence of pedestrian's face In vivo detection simultaneously terminates random formation sequence flow；It is additionally operable to the institute that ought be generated at random
State sequence and be less than or equal to the second predetermined threshold value with the matching degree of the last detection sequence for carrying out face In vivo detection
When, go to sequence generating unit.
Specifically, the acquisition module 52 includes：
Reading unit, for sequentially reading the current bit value of the detection sequence；
3rd determining unit, for determining the facial physiological motion corresponding to the current numerical value for reading, wherein,
A kind of a kind of facial physiological motion of numerical value correspondence；
Indicating member, for indicating to be detected user's execution facial physiological motion；
Collecting unit, user characteristics is detected for being gathered in Preset Time；
Second judging unit, at the end of Preset Time, judging whether the detection sequence not read
Numerical value, if it is, reading unit is gone to, if not, terminating to be detected user characteristics collection.
Specifically, second generation module 53, including：
3rd judging unit, for judging to be detected user whether according to finger according to the detected user characteristics
Show the execution facial physiological motion；
4th determining unit, for the 3rd judging unit judged result for be when, will be with the facial physiology
Property the corresponding numerical value of motion be defined as the current bit value of testing result sequence；It is additionally operable to judge single the 3rd
When first judged result is no, it is defined as the different numerical value of corresponding numerical value is moved from the facial physiological
The current bit value of testing result sequence.
Specifically, the detection sequence and testing result sequence are 01 sequence；
4th determining unit, is additionally operable to when it is 0 that the facial physiological moves corresponding numerical value,
It is defined as 1 by the different numerical value of corresponding numerical value is moved from the facial physiological；It is additionally operable to when the face
When physiological motion corresponding numerical value in portion's is 1, by the numerical value corresponding with the facial physiological motion not
Same numerical value is defined as 0.
In another alternative embodiment of the invention, as shown in fig. 6, described device also includes：3rd is true
Cover half block 57, for determining that the detection sequence is fuzzy with testing result sequence in the second determining module 56
Before matching degree, determine that the detected user characteristics is not that video is attacked.
Specifically, the 3rd determining module 57, including：
Extraction unit, the face detection region for extracting the detected user characteristics；
5th determining unit, the aspect ratio value for determining the face detection region；
4th judging unit, for judging the aspect ratio value whether in the 3rd threshold range；
6th determining unit, for when the 4th judging unit judged result is to be, determining the detected use
Family feature is not that video is attacked.
In specific implementation process, above-mentioned first generation module 51, acquisition module 52, the second generation module
53rd, the first determining module 54, judge module 55, the second determining module 56 and the 3rd determining module 57
With by central processing unit (CPU, Central Processing Unit), the micro- place in face In vivo detection system
Reason device (MPU, Micro Processing Unit), digital signal processor (DSP, Digital Signal
) or programmable logic array (FPGA, Field-Programmable Gate Array) comes real Processor
It is existing.
For device embodiment, because it is substantially similar to embodiment of the method, so the comparing of description is simple
Single, the relevent part can refer to the partial explaination of embodiments of method.
Each embodiment in this specification is described by the way of progressive, what each embodiment was stressed
All be the difference with other embodiment, between each embodiment identical similar part mutually referring to.
It should be understood by those skilled in the art that, the embodiment of the embodiment of the present invention can be provided as method, device,
Or computer program product.Therefore, the embodiment of the present invention can use complete hardware embodiment, complete software reality
Apply example or the form with reference to the embodiment in terms of software and hardware.And, the embodiment of the present invention can be used
One or more wherein include the computer-usable storage medium of computer usable program code (including but not
Be limited to magnetic disk storage, CD-ROM, optical memory etc.) on implement computer program product form.
The embodiment of the present invention is with reference to method according to embodiments of the present invention, terminal device (system) and calculates
The flow chart and/or block diagram of machine program product is described.It should be understood that can be realized by computer program instructions
In each flow and/or square frame and flow chart and/or block diagram in flow chart and/or block diagram
The combination of flow and/or square frame.These computer program instructions to all-purpose computer, dedicated computing can be provided
The processor of machine, Embedded Processor or other programmable data processing terminal equipments to produce a machine,
So that produced by the instruction of computer or the computing device of other programmable data processing terminal equipments being used for
Realize what is specified in one flow of flow chart or multiple one square frame of flow and/or block diagram or multiple square frames
The device of function.
These computer program instructions may be alternatively stored in can guide computer or other programmable data processing terminals
In the computer-readable memory that equipment works in a specific way so that storage is in the computer-readable memory
In instruction produce include the manufacture of command device, the command device realization in one flow or many of flow chart
The function of being specified in one square frame of individual flow and/or block diagram or multiple square frames.
These computer program instructions can also be loaded into computer or other programmable data processing terminal equipments
On so that series of operation steps is performed on computer or other programmable terminal equipments to produce computer
The treatment of realization, so as to the instruction performed on computer or other programmable terminal equipments is provided for realizing
The function of being specified in one flow of flow chart or multiple one square frame of flow and/or block diagram or multiple square frames
The step of.
Although having been described for the preferred embodiment of the embodiment of the present invention, those skilled in the art once obtain
Cicada basic creative concept, then can make other change and modification to these embodiments.So, it is appended
Claim be intended to be construed to include preferred embodiment and fall into range of embodiment of the invention have altered and
Modification.
Finally, in addition it is also necessary to explanation, herein, such as first and second or the like relational terms are only
Only be used for by an entity or operation with another entity or operate make a distinction, and not necessarily require or
Imply between these entities or operation there is any this actual relation or order.And, term " bag
Include ", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion so that being including one
The process of row key element, method, article or terminal device not only include those key elements, but also including not having
Other key elements being expressly recited, or it is this process, method, article or terminal device institute also to include
Intrinsic key element.In the absence of more restrictions, the key element limited by sentence "including a ...",
It is not precluded from also the presence of other phase in the process including the key element, method, article or terminal device
Same key element.
Above to a kind of multi-channel video display methods provided by the present invention and device, it is described in detail,
Specific case used herein is set forth to principle of the invention and implementation method, above example
Illustrate that being only intended to help understands the method for the present invention and its core concept；Simultaneously for the general of this area
Technical staff, thought of the invention, will change in specific embodiments and applications,
In sum, this specification content should not be construed as limiting the invention.
Claims (14)
1. a kind of human face in-vivo detection method, it is characterised in that methods described includes：
The detection sequence of random generation predetermined bit length, wherein, the predetermined bit length value and default inspection
Survey time numerical value identical；
Detection sequence according to generation performs the detected user characteristics collection of preset times time, the detected use
Family is characterized as the detected user's face provincial characteristics；
Testing result generation testing result sequence according to the detected user characteristics for collecting, wherein, one
One digit number value in testing result correspondence testing result sequence；
Determine the fuzzy matching degree of detection sequence and testing result sequence；
When the fuzzy matching degree of detection sequence and testing result sequence is more than the first predetermined threshold value, it is determined that described
It is live body to be detected user.
2. method according to claim 1, it is characterised in that the random generation predetermined bit length
Detection sequence comprise the following steps：
Random formation sequence；
It is determined that the sequence of random generation and the last detection sequence for carrying out face In vivo detection are matched
Degree；
Judge the sequence of random generation and the matching degree of the last detection sequence for carrying out face In vivo detection
Whether the second predetermined threshold value is less than；
When the sequence of random generation is small with the matching degree of the last detection sequence for carrying out face In vivo detection
When the second predetermined threshold value, the sequence of random generation is defined as currently carrying out the inspection of face In vivo detection
Sequencing row；Otherwise, the step of re-executing the random formation sequence.
3. method according to claim 2, it is characterised in that the detection sequence according to generation is held
The detected user characteristics collection of row preset times time, comprises the following steps：
Order reads the current bit value of the detection sequence；
It is determined that the facial physiological motion corresponding to the current numerical value for reading, wherein, a kind of numerical value correspondence is a kind of
Facial physiological motion；
Indicate to be detected user's execution facial physiological motion；
The detected user characteristics is gathered in Preset Time；
The detection sequence numerical value not read is judged whether, if it is, re-executing the order reads institute
The step of stating detection sequence current bit value；If not, terminating to be detected user characteristics acquisition step.
4. method according to claim 3, it is characterised in that the detected use that the basis is collected
The testing result generation testing result sequence of family feature, including：
Determine to be detected user whether according to the instruction execution facial physiology according to the detected user characteristics
Property motion；
If it is, the numerical value corresponding with the facial physiological motion is defined as into working as testing result sequence
Preceding bit value；
If not, being defined as detection knot by the different numerical value of corresponding numerical value is moved from the facial physiological
The current bit value of infructescence row.
5. method according to claim 4, it is characterised in that the detection sequence and testing result sequence
It is classified as 01 sequence；
It is relative with the facial physiological motion when it is 0 that the facial physiological moves corresponding numerical value
The different numerical value of the numerical value answered is 1；
It is relative with the facial physiological motion when it is 1 that the facial physiological moves corresponding numerical value
The different numerical value of the numerical value answered is 0.
6. according to the method described in claim 1 to 5 any of which, it is characterised in that the determination inspection
Before the fuzzy matching degree of sequencing row and testing result sequence, methods described also includes：
Determine that the detected user characteristics is not that video is attacked.
7. method according to claim 6, it is characterised in that the determination detected user is special
It is not that video is attacked to levy, including：
Extract the face detection region of the detected user characteristics；
Determine the aspect ratio value in the face detection region；
Judge the aspect ratio value whether in the 3rd threshold range；
If it is, determining that the detected user characteristics is not that video is attacked.
8. a kind of face living body detection device, it is characterised in that described device includes：First generation module,
Acquisition module, the second generation module, the first determining module, judge module and the second determining module；Wherein,
First generation module, for the detection sequence of random generation predetermined bit length, wherein, it is described
Predetermined bit length value is identical with default detection time numerical value；
The acquisition module, the detected user characteristics of preset times time is performed for the detection sequence according to generation
Collection, the detected user characteristics is the detected user's face provincial characteristics；
Second generation module, inspection is generated for the testing result according to the detected user characteristics for collecting
Result sequence is surveyed, wherein, the one digit number value in a testing result correspondence testing result sequence；
First determining module, the fuzzy matching degree for determining detection sequence and testing result sequence；
The judge module, for judging whether detection sequence is more than with the fuzzy matching degree of testing result sequence
First predetermined threshold value；
Second determining module, for the fuzzy matching degree when detection sequence and testing result sequence more than the
During one predetermined threshold value, determine that the detected user is live body.
9. device according to claim 8, it is characterised in that first generation module, including：
Sequence generating unit, for random formation sequence；
First determining unit, the sequence for determining random generation carries out face In vivo detection with the last time
Detection sequence matching degree；
First judging unit, the sequence for judging random generation carries out face In vivo detection with the last time
Detection sequence matching degree whether be less than the second predetermined threshold value；
Second determining unit, the sequence and last time for that ought generate at random carry out face In vivo detection
When the matching degree of detection sequence is less than the second predetermined threshold value, the sequence of random generation is defined as when advance
The detection sequence of pedestrian's face In vivo detection simultaneously terminates random formation sequence；It is additionally operable to the sequence that ought be generated at random
When row are less than or equal to the second predetermined threshold value with the matching degree of the last detection sequence for carrying out face In vivo detection,
Go to sequence generating unit.
10. device according to claim 9, it is characterised in that the acquisition module includes：
Reading unit, for sequentially reading the current bit value of the detection sequence；
3rd determining unit, for determining the facial physiological motion corresponding to the current numerical value for reading, wherein,
A kind of a kind of facial physiological motion of numerical value correspondence；
Indicating member, for indicating to be detected user's execution facial physiological motion；
Collecting unit, user characteristics is detected for being gathered in Preset Time；
Second judging unit, at the end of Preset Time, judging whether the detection sequence not read
Numerical value, if it is, reading unit is gone to, if not, terminating to be detected user characteristics collection.
11. devices according to claim 10, it is characterised in that second generation module, including：
3rd judging unit, for judging to be detected user whether according to finger according to the detected user characteristics
Show the execution facial physiological motion；
4th determining unit, for the 3rd judging unit judged result for be when, will be with the facial physiology
Property the corresponding numerical value of motion be defined as the current bit value of testing result sequence；It is additionally operable to judge single the 3rd
When first judged result is no, it is defined as the different numerical value of corresponding numerical value is moved from the facial physiological
The current bit value of testing result sequence.
12. devices according to claim 11, it is characterised in that the detection sequence and testing result
Sequence is 01 sequence；
4th determining unit, for when it is 0 that the facial physiological moves corresponding numerical value, inciting somebody to action
The different numerical value of corresponding numerical value is moved from the facial physiological be defined as 1；It is additionally operable to when the face
When the corresponding numerical value of physiological motion is 1, corresponding numerical value will be moved from the facial physiological different
Numerical value be defined as 0.
13. device according to claim 8 to 12 any of which, it is characterised in that described device
Also include：3rd determining module, for determining the detection sequence and testing result sequence in the second determining module
Before the fuzzy matching degree of row, determine that the detected user characteristics is not that video is attacked.
14. devices according to claim 13, it is characterised in that the 3rd determining module, including：
Extraction unit, the face detection region for extracting the detected user characteristics；
5th determining unit, the aspect ratio value for determining the face detection region；
4th judging unit, for judging the aspect ratio value whether in the 3rd preset threshold range；
6th determining unit, for when the 4th judging unit judged result is to be, determining the detected use
Family feature is not that video is attacked.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510960637.XA CN106897657B (en) | 2015-12-18 | 2015-12-18 | Face living body detection method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510960637.XA CN106897657B (en) | 2015-12-18 | 2015-12-18 | Face living body detection method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106897657A true CN106897657A (en) | 2017-06-27 |
| CN106897657B CN106897657B (en) | 2020-10-27 |
Family
ID=59190168
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510960637.XA Active CN106897657B (en) | 2015-12-18 | 2015-12-18 | Face living body detection method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106897657B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107992845A (en) * | 2017-12-14 | 2018-05-04 | 广东工业大学 | A kind of face recognition the method for distinguishing and device, computer equipment |
| CN110633659A (en) * | 2019-08-30 | 2019-12-31 | 北京旷视科技有限公司 | Living body detection method, living body detection device, computer equipment and storage medium |
| CN111783644A (en) * | 2020-06-30 | 2020-10-16 | 百度在线网络技术（北京）有限公司 | Detection method, device, equipment and computer storage medium |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101984422A (en) * | 2010-10-18 | 2011-03-09 | 百度在线网络技术（北京）有限公司 | Fault-tolerant text query method and equipment |
| US20130254183A1 (en) * | 2003-09-10 | 2013-09-26 | International Business Machines Corporation | Semantic discovery and mapping between data sources |
| CN104683302A (en) * | 2013-11-29 | 2015-06-03 | 国际商业机器公司 | Authentication method, authentication device, terminal equipment, authentication server and system |
| CN105518708A (en) * | 2015-04-29 | 2016-04-20 | 北京旷视科技有限公司 | Method and equipment for verifying living human face, and computer program product |
| CN105760817A (en) * | 2016-01-28 | 2016-07-13 | 深圳泰首智能技术有限公司 | Method and device for recognizing, authenticating, unlocking and encrypting storage space by using human face |
-
2015
- 2015-12-18 CN CN201510960637.XA patent/CN106897657B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130254183A1 (en) * | 2003-09-10 | 2013-09-26 | International Business Machines Corporation | Semantic discovery and mapping between data sources |
| CN101984422A (en) * | 2010-10-18 | 2011-03-09 | 百度在线网络技术（北京）有限公司 | Fault-tolerant text query method and equipment |
| CN104683302A (en) * | 2013-11-29 | 2015-06-03 | 国际商业机器公司 | Authentication method, authentication device, terminal equipment, authentication server and system |
| CN105518708A (en) * | 2015-04-29 | 2016-04-20 | 北京旷视科技有限公司 | Method and equipment for verifying living human face, and computer program product |
| CN105760817A (en) * | 2016-01-28 | 2016-07-13 | 深圳泰首智能技术有限公司 | Method and device for recognizing, authenticating, unlocking and encrypting storage space by using human face |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107992845A (en) * | 2017-12-14 | 2018-05-04 | 广东工业大学 | A kind of face recognition the method for distinguishing and device, computer equipment |
| CN110633659A (en) * | 2019-08-30 | 2019-12-31 | 北京旷视科技有限公司 | Living body detection method, living body detection device, computer equipment and storage medium |
| CN110633659B (en) * | 2019-08-30 | 2022-11-04 | 北京旷视科技有限公司 | Living body detection method, living body detection device, computer equipment and storage medium |
| CN111783644A (en) * | 2020-06-30 | 2020-10-16 | 百度在线网络技术（北京）有限公司 | Detection method, device, equipment and computer storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106897657B (en) | 2020-10-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108182409B (en) | Living body detection method, living body detection device, living body detection equipment and storage medium | |
| CN106557726B (en) | Face identity authentication system with silent type living body detection and method thereof | |
| CN110321873A (en) | Sensitization picture recognition methods and system based on deep learning convolutional neural networks | |
| CN107066942A (en) | A kind of living body faces recognition methods and system | |
| CN106960177A (en) | Living body faces verification method and system, living body faces checking device | |
| CN110851835A (en) | Image model detection method and device, electronic equipment and storage medium | |
| CN111144293A (en) | Human face identity authentication system with interactive living body detection and method thereof | |
| CN106570489A (en) | Living body determination method and apparatus, and identity authentication method and device | |
| CN111126366B (en) | Method, device, equipment and storage medium for distinguishing living human face | |
| US11908240B2 (en) | Micro-expression recognition method based on multi-scale spatiotemporal feature neural network | |
| CN103761465A (en) | Method and device for identity authentication | |
| CN109858514A (en) | A kind of video behavior classification method neural network based | |
| CN107590473A (en) | A kind of human face in-vivo detection method, medium and relevant apparatus | |
| CN107316029A (en) | A kind of live body verification method and equipment | |
| CN107358209A (en) | Training method, device and method for detecting human face, the device of Face datection model | |
| CN109815797A (en) | Biopsy method and device | |
| CN107819790A (en) | The recognition methods of attack message and device | |
| CN106897657A (en) | A kind of human face in-vivo detection method and device | |
| CN112307984B (en) | Safety helmet detection method and device based on neural network | |
| CN112464690A (en) | Living body identification method, living body identification device, electronic equipment and readable storage medium | |
| Nguyen et al. | Deep visual saliency on stereoscopic images | |
| CN111291863A (en) | Training method of face changing identification model, face changing identification method, device and equipment | |
| CN110688878B (en) | Living body identification detection method, living body identification detection device, living body identification detection medium, and electronic device | |
| CN115953822B (en) | Human face video fake identification method and device based on rPPG physiological signals | |
| CN110393539A (en) | Psychological abnormality detection method, device, storage medium and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |