CN110113319A - Identity identifying method, device, computer equipment and storage medium - Google Patents
Identity identifying method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110113319A CN110113319A CN201910304302.0A CN201910304302A CN110113319A CN 110113319 A CN110113319 A CN 110113319A CN 201910304302 A CN201910304302 A CN 201910304302A CN 110113319 A CN110113319 A CN 110113319A
- Authority
- CN
- China
- Prior art keywords
- lip
- recognition
- user
- network model
- neural network
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/32—User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—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
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
- H04L63/0861—Network architectures or network communication protocols for network security for authentication of entities using biometrical features, e.g. fingerprint, retina-scan
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/08—Network architectures or network communication protocols for network security for authentication of entities
- H04L63/0876—Network architectures or network communication protocols for network security for authentication of entities based on the identity of the terminal or configuration, e.g. MAC address, hardware or software configuration or device fingerprint
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Computer Security & Cryptography (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Computing Systems (AREA)
- Signal Processing (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Computer Networks & Wireless Communication (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Mathematical Physics (AREA)
- Geometry (AREA)
- Power Engineering (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Collating Specific Patterns (AREA)
Abstract
This application involves field of artificial intelligence, are applied to financial industry, more particularly to a kind of identity identifying method, device, computer equipment and storage medium.Method in one embodiment includes: to obtain the certification request of user to be certified, certification request carries the user identifier of user to be certified and the certification video of user to be certified, Face datection and face feature point detection are carried out to certification video, obtain lip image sequence, feature extraction is carried out to lip image sequence, obtain the feature vector of lip image sequence, using feature vector as lid speech characteristic authentication information, pre-set user registration information is searched according to user identifier, obtain lid speech characteristic registration information corresponding with user identifier, when lid speech characteristic authentication information is matched with lid speech characteristic registration information, it obtains user to be certified and passes through the authentication result of certification.It may be implemented to carry out authentication by user identifier+lip reading identification in this way, can effectively improve the safety of authentication.
Description
Technical field
This application involves field of artificial intelligence, more particularly to a kind of identity identifying method, device, computer equipment
And storage medium.
Background technique
With the development of science and technology, earth-shaking variation has occurred in people's lives, and mobile internet device becomes
The necessity of daily life, personally identifiable information are play an important role in Internet technology.
Traditional identity authentication method, for example identity is confirmed by password, password, certificate, there are verification informations to be easy quilt
The problem of stealing, being assumed another's name by people, i.e., traditional identity identifying method have that safety is low.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of authentication side that can be improved authentication security
Method, device, computer equipment and storage medium.
A kind of identity identifying method, which comprises
The certification request of user to be certified is obtained, the certification request carries the user identifier of user to be certified and wait recognize
Demonstrate,prove the certification video of user；
Face datection and face feature point detection are carried out to the certification video, obtain lip image sequence；
Feature extraction is carried out to the lip image sequence, the feature vector of the lip image sequence is obtained, with described
Feature vector is lid speech characteristic authentication information；
Pre-set user registration information is searched according to the user identifier, obtains lid speech characteristic corresponding with the user identifier
Registration information；
When the lid speech characteristic authentication information is matched with the lid speech characteristic registration information, obtains user to be certified and pass through
The authentication result of certification.
In one embodiment, described that feature extraction is carried out to the lip image sequence, obtain the lip image sequence
The feature vector of column, comprising:
The corresponding video block of the lip image sequence is divided into multiple sub-video blocks；
Each sub-video block is passed through into the first independence subspace analysis respectively and carries out characteristic vector pickup, obtains fisrt feature
Vector set；
Dimension-reduction treatment is carried out to the feature vector in the first eigenvector set, by the feature vector after dimension-reduction treatment
The second independence subspace analysis is inputted, second feature vector set is obtained；
According to the first eigenvector set and the second feature vector set, the lip image sequence is obtained
Feature vector.
In one embodiment, before the progress feature extraction to the lip image sequence, further includes:
Obtain video sample and Recognition with Recurrent Neural Network model, the video sample include lip reading image and with the lip
Sonagram is as corresponding lip reading recognition result；
Using the lip reading image as the input of the Recognition with Recurrent Neural Network model, using the lip reading recognition result as institute
The output for stating Recognition with Recurrent Neural Network model is trained the Recognition with Recurrent Neural Network model, the circulation nerve trained
Network model；
It is described that feature extraction is carried out to the lip image sequence, comprising:
Feature extraction is carried out to the lip image sequence by the Recognition with Recurrent Neural Network model trained.
In one embodiment, described using the lip reading image as the input of the Recognition with Recurrent Neural Network model, by institute
Output of the lip reading recognition result as the Recognition with Recurrent Neural Network model is stated, the Recognition with Recurrent Neural Network model is trained,
Before the Recognition with Recurrent Neural Network model trained, further includes:
Image procossing is carried out to the video sample, obtains syllable adhesion data；
It is described using the lip reading image as the input of the Recognition with Recurrent Neural Network model, the lip reading recognition result is made
For the output of the Recognition with Recurrent Neural Network model, the Recognition with Recurrent Neural Network model is trained, the circulation trained
Neural network model, comprising:
Using the lip reading image and the syllable adhesion data as the input of the Recognition with Recurrent Neural Network model, by institute
Output of the lip reading recognition result as the Recognition with Recurrent Neural Network model is stated, the Recognition with Recurrent Neural Network model is trained,
The Recognition with Recurrent Neural Network model trained.
In one embodiment, described that image procossing is carried out to the video sample, obtain syllable adhesion data, comprising:
Mouth sequence picture is converted by the video sample；
Mouth-Shape Recognition is carried out to the mouth sequence picture, obtains mouth sequence；
Syllable sequence is converted by the mouth sequence, based on default semantic base and the syllable sequence, obtains syllable
Adhesion data.
In one embodiment, described when the lid speech characteristic authentication information is matched with the lid speech characteristic registration information
When, it obtains user to be certified and passes through the authentication result of certification, comprising:
Obtain the similarity between the lid speech characteristic authentication information and the lid speech characteristic registration information；
When the similarity between the lid speech characteristic authentication information and the lid speech characteristic registration information is greater than preset threshold
When, it obtains user to be certified and passes through the authentication result of certification.
A kind of identification authentication system, described device include:
Certification request obtains module, and for obtaining the certification request of user to be certified, the certification request carries to be certified
The certification video of the user identifier of user and user to be certified；
Image sequence extraction module obtains mouth for carrying out Face datection and face feature point detection to certification video
Lip image sequence；
Feature vector obtains module, for carrying out feature extraction to lip image sequence, obtains the spy of lip image sequence
Vector is levied, using feature vector as lid speech characteristic authentication information；Registration information obtains module, for being searched according to the user identifier
Pre-set user registration information obtains lid speech characteristic registration information corresponding with the user identifier；
Authentication module, for obtaining when the lid speech characteristic authentication information is matched with the lid speech characteristic registration information
The authentication result that user to be certified passes through certification.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
Device performs the steps of when executing the computer program
The certification request of user to be certified is obtained, the certification request carries the user identifier of user to be certified and wait recognize
Demonstrate,prove the certification video of user；
Face datection and face feature point detection are carried out to the certification video, obtain lip image sequence；
Feature extraction is carried out to the lip image sequence, the feature vector of the lip image sequence is obtained, with described
Feature vector is lid speech characteristic authentication information；
Pre-set user registration information is searched according to the user identifier, obtains lid speech characteristic corresponding with the user identifier
Registration information；
When the lid speech characteristic authentication information is matched with the lid speech characteristic registration information, obtains user to be certified and pass through
The authentication result of certification.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
It is performed the steps of when row
The certification request of user to be certified is obtained, the certification request carries the user identifier of user to be certified and wait recognize
Demonstrate,prove the certification video of user；
Face datection and face feature point detection are carried out to the certification video, obtain lip image sequence；
Feature extraction is carried out to the lip image sequence, the feature vector of the lip image sequence is obtained, with described
Feature vector is lid speech characteristic authentication information；
Pre-set user registration information is searched according to the user identifier, obtains lid speech characteristic corresponding with the user identifier
Registration information；
When the lid speech characteristic authentication information is matched with the lid speech characteristic registration information, obtains user to be certified and pass through
The authentication result of certification.
Above-mentioned identity identifying method, device, computer equipment and storage medium, the certification by obtaining user to be certified are asked
It asks, certification request carries the user identifier of user to be certified and the certification video of user to be certified, carries out people to certification video
Face detection and face feature point detection, obtain lip image sequence, carry out feature extraction to lip image sequence, obtain lip
The feature vector of image sequence searches pre-set user registration according to user identifier using feature vector as lid speech characteristic authentication information
Information obtains lid speech characteristic registration information corresponding with user identifier, when lid speech characteristic authentication information and lid speech characteristic registration are believed
When breath matching, obtain user to be certified by the authentication result of certification, may be implemented to identify by user identifier+lip reading in this way into
Row authentication can effectively improve the safety of authentication；And by obtaining lip image sequence to certification video, then from
Feature vector is extracted in lip image sequence, obtains lid speech characteristic authentication information, the essence of the lid speech characteristic got can be improved
Parasexuality, to further increase the safety of authentication.
Detailed description of the invention
Fig. 1 is the flow diagram of identity identifying method in one embodiment；
Fig. 2 is the flow diagram of the feature vector generation step of lip image sequence in one embodiment；
Fig. 3 is the flow diagram of identity identifying method in another embodiment；
Fig. 4 is the flow diagram of one embodiment syllable adhesion data generation step；
Fig. 5 is the flow diagram of user authentication process in one embodiment；
Fig. 6 is the structural block diagram of identification authentication system in one embodiment；
Fig. 7 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
In one embodiment, as shown in Figure 1, providing a kind of identity identifying method, comprising the following steps:
Step 102, obtain the certification request of user to be certified, certification request carry user to be certified user identifier and
The certification video of user to be certified.
Certification request refers to be proposed by user to be certified, and the request of authentication is carried out to certification authority.User identifier is
Refer to the mark for distinguishing different user, each user to be certified is corresponding to be identified uniquely, for example can pass through number A, B, C etc.
Distinguish different users to be certified.Certification video is when referring to that user to be certified authenticates, voice that field real-time acquisition arrives and
Image information, such as user to be certified read aloud the video of one section of password.
Step 104, Face datection is carried out to certification video and face feature point detects, obtain lip image sequence.
Face datection refers to for given image, scans for image to determine whether containing face, if it is
Then return to position, size and the posture of face.Face datection can by the method based on feature, the method based on template and
Method based on statistical theory is realized.Specifically, rule derived from the priori knowledge using face is passed through based on the method for feature
Carry out Face datection, including edge and shape feature, textural characteristics, color characteristic etc.；Method based on template is by calculating people
Correlation between face template and image to be detected carries out Face datection, for example, calculate separately image to be detected eyes, nose,
The correlation between features and face template such as mouth, by the size of correlation to determine whether there are faces；It is managed based on statistics
The method of opinion refers to finds face and non-face sample characteristics with machine learning using statistical analysis respectively, passes through the feature found
Classifier is constructed, to realize Face datection.It can be using the method returned by cascade to face in the facial image detected
Portion's characteristic point is detected, and carries out lip region extraction according to the position of characteristic point.Based on cascade return method by using
A series of weak non-linear relations for returning device fitting complexity, study regression function map directly to testing result.
By human face detection tech and face feature point detection technique, lip region is extracted from video image, is realized
Positioning to face lip.For example, can calculate using two key points of the corners of the mouth, the translation and rotation relative to standard mouth are obtained
Transposon.Use eye spacing as benchmark, the lip of different people is transformed into identical scale by scale parameter.According to translation
With twiddle factor and scale parameter, the lip image sequence being aligned.Lip alignment can eliminate mouth in different frame image
The position of lip, angle and scale it is inconsistent, lip reading identification is carried out by the lip image sequence of alignment, not only may be implemented
The opposite change information for talking about mouth, can also be normalized different speaker's difference mouths, position of the reduction because of lip, angle
With the inconsistent influence for causing to identify lip reading of scale.
Step 106, feature extraction is carried out to lip image sequence, the feature vector of lip image sequence is obtained, with feature
Vector is lid speech characteristic authentication information.
Lip image sequence refers to speaker's continuous shape of the mouth as one speaks variation characteristic in speech.Lip reading identification be can directly from
The technology that speech content is identified in the image of someone's speech continuously identifies face from certification video by lip reading identification, sentences
It is disconnected to extract the continuous shape of the mouth as one speaks variation characteristic of this person wherein just in talker, the feature of consecutive variations is input to lip reading identification
In model, the corresponding pronunciation of teller's shape of the mouth as one speaks is identified, further according to the pronunciation identified, obtain lip reading recognition result.Lip reading is special
Sign authentication information, which refers to, when carrying out authentication, is parsing certification video, obtained lip reading recognition result to user to be certified.
Step 108, pre-set user registration information is searched according to user identifier, obtains lid speech characteristic corresponding with user identifier
Registration information.
User acquires user identifier and registration video when registering, by user identifier with register that video is corresponding deposits
Storage.Registration video refers to that user reads aloud customized or setting password video, parses, obtains corresponding to registration video
User identifier and corresponding lid speech characteristic registration information associated storage are obtained user's registration information by lid speech characteristic registration information.
Lid speech characteristic registration information, which refers to, when registering, is parsing registration video, obtained lip reading recognition result to user.
Step 110, it when lid speech characteristic authentication information is matched with lid speech characteristic registration information, obtains user to be certified and passes through
The authentication result of certification.
For example, user's first, in registration, typing registration video, collected registration video are parsed, obtained lip reading
Recognition result is 201810, i.e., lid speech characteristic registration information is 201810.When carrying out authentication to user's first, acquisition certification
Video parses collected certification video, and obtained lip reading recognition result is 201810, i.e. lid speech characteristic authentication information is
201810.The lid speech characteristic authentication information of first is matched with lid speech characteristic registration information, thus obtains user's first to be certified by recognizing
The authentication result of card.
Above-mentioned identity identifying method, by obtaining the certification request of user to be certified, certification request carries user to be certified
User identifier and user to be certified certification video, Face datection and face feature point are carried out to certification video and detected,
Lip image sequence is obtained, feature extraction is carried out to lip image sequence, the feature vector of lip image sequence is obtained, with feature
Vector is lid speech characteristic authentication information, searches pre-set user registration information according to user identifier, obtains corresponding with user identifier
It is logical to obtain user to be certified when lid speech characteristic authentication information is matched with lid speech characteristic registration information for lid speech characteristic registration information
The authentication result for crossing certification may be implemented to carry out authentication by user identifier+lip reading identification in this way, can effectively improve body
The safety of part certification；And by certification video obtain lip image sequence, then from lip image sequence extract feature to
Amount, obtains lid speech characteristic authentication information, the accuracy of lip reading identification can be improved, to further increase the safety of authentication
Property.
In one embodiment, as shown in Fig. 2, carrying out feature extraction to lip image sequence, lip image sequence is obtained
Feature vector, comprising: step 202, the corresponding video block of lip image sequence is divided into multiple sub-video blocks；Step 204,
Each sub-video block is passed through into the first independence subspace analysis respectively and carries out characteristic vector pickup, obtains first eigenvector collection
It closes；Step 206, dimension-reduction treatment is carried out to the feature vector in first eigenvector set, by the feature vector after dimension-reduction treatment
The second independence subspace analysis is inputted, second feature vector set is obtained；Step 208, according to first eigenvector set and
Two feature vector set, obtain the feature vector of lip image sequence.According to lip region image sequence, the feature of image is obtained
Vector, as lid speech characteristic information.Since lip reading video includes image and motion information, a feature vector connects from multiframe
It is extracted in the subimage sequence that continuous image is formed.Independence subspace analysis is a kind of unsupervised feature learning method, can be from image
The feature of phase invariant is provided in middle school's acquistion.That the first independence subspace analysis extracts is the spy on mobile side in video sequence
Sign, the data that the first independence subspace analysis is extracted carry out dimensionality reduction by principal component analysis, can accelerate calculating speed,
What the second independence subspace analysis was handled is the input data compared with low dimensional, and what is extracted is more abstract feature.Specifically may be used
To realize feature extraction by stacking convolution independence subspace analysis network, convolution independence subspace analysis network is stacked by only
Vertical subspace analysis and principal component analysis successively stack composition.When calculating video features first by the picture in lesser video block
Element is input to first layer independence subspace analysis network after pulling into a vector, and then video block adjacent in bigger region is only
Vertical subspace analysis output is joined together, and after the pretreatment of principal component analysis dimensionality reduction, is input to second layer Independent subspace
Network is analyzed, and so on, the output of last every layer of independence subspace analysis network is connected into a vector as the view
The feature vector of frequency block.Feature vector in order to obtain is divided into multiple small by the video block of multiple continuous lip reading image constructions
Video block, each small video block, which is input into stack in convolution independence subspace analysis network, extracts feature vector, last each
The feature vector of small video block and treated that multiple feature vectors are connected is unified into final feature by principal component analysis
Vector.
In one embodiment, as shown in figure 3, before to the progress feature extraction of lip image sequence, further includes: step
302, obtain video sample and Recognition with Recurrent Neural Network model, video sample includes lip reading image and corresponding with lip reading image
Lip reading recognition result；Step 304, using lip reading image as the input of Recognition with Recurrent Neural Network model, using lip reading recognition result as
The output of Recognition with Recurrent Neural Network model is trained Recognition with Recurrent Neural Network model, the Recognition with Recurrent Neural Network mould trained
Type；Feature extraction is carried out to lip image sequence, comprising: step 306, by the Recognition with Recurrent Neural Network model trained to lip
Image sequence carries out feature extraction.Certification video is made of continuous multiple image, and authentication authorization and accounting video is sequence data, and sequence
There is relevance, the picture frame that front occurs has significant impact or even subsequent figure to subsequent picture frame between before and after data
As data frame of the frame to front is also to have a major impact.Recognition with Recurrent Neural Network needs to consider front when handling current information
The information of appearance, Recognition with Recurrent Neural Network may include all Given informations before current information, recycle the pre-treatment referred to
Information can be utilized always and help subsequent information.Identification for lip reading passes through the Recognition with Recurrent Neural Network model trained
It realizes, obtains using this model lip shape and change with time feature, to carry out predicted letters, for example, acquisition speaker is bright
The image for reading Arabic numerals 0, in this, as lip reading image, corresponding lip reading recognition result is Arabic numerals 0.By speaker
The image for reading aloud Arabic numerals 0 played the input of model as circulation nerve, was that Arabic numerals 0 are made by lip reading recognition result
For the standard output of Recognition with Recurrent Neural Network, Recognition with Recurrent Neural Network model is trained, according to the reality of Recognition with Recurrent Neural Network model
Border output and standard output, adjust the training weight of Recognition with Recurrent Neural Network model, until the reality of Recognition with Recurrent Neural Network model
Output meets preset condition compared with standard output, the Recognition with Recurrent Neural Network model trained.
In one embodiment, using lip reading image as the input of Recognition with Recurrent Neural Network model, lip reading recognition result is made
For the output of Recognition with Recurrent Neural Network model, Recognition with Recurrent Neural Network model is trained, the Recognition with Recurrent Neural Network trained
Before model, further includes: carry out image procossing to video sample, obtain syllable adhesion data；Using lip reading image as circulation mind
Input through network model, using lip reading recognition result as the output of Recognition with Recurrent Neural Network model, to Recognition with Recurrent Neural Network model
It is trained, the Recognition with Recurrent Neural Network model trained, comprising: using lip reading image and syllable adhesion data as circulation
The input of neural network model, using lip reading recognition result as the output of Recognition with Recurrent Neural Network model, to Recognition with Recurrent Neural Network mould
Type is trained, the Recognition with Recurrent Neural Network model trained.It joined when carrying out Recognition with Recurrent Neural Network model training pair
The training of syllable adhesion when fast word speed keeps lip reading identification more acurrate.For example the video training data of fast word speed can be increased, so that
Recognition with Recurrent Neural Network model can learn the lip shape variation in the case of fast word speed；It can also reinforce video frame rate, reduce circulation nerve
The difficulty of network model training.
In one embodiment, as shown in figure 4, carrying out image procossing to video sample, syllable adhesion data are obtained, are wrapped
It includes: step 402, converting mouth sequence picture for video sample；Step 404, lip-sync sequence of pictures carries out Mouth-Shape Recognition, obtains
To mouth sequence；Step 406, syllable sequence is converted by mouth sequence, based on default semantic base and syllable sequence, obtains sound
Save adhesion data.Mouth sequence picture is converted by video sample by image procossing, mouth sequence picture is subjected to gray processing
And filtering processing, then binary conversion treatment is carried out, so as to the extraction of feature vector.Shape of the mouth as one speaks template is lip-sync according to daily life
Empirically determined, lip-sync sequence of pictures carries out characteristic vector pickup, and is matched with shape of the mouth as one speaks template, obtains mouth sequence.Language
Yi Ku can be using Chinese phonetic alphabet as basic unit the shape of the mouth as one speaks template library established, and include all phonetic words in semantic base
The shape of the mouth as one speaks picture and multi-C vector parameter of mother's pronunciation.Syllable sequence is converted into word, language by the efficient combination of syllable sequence
The semantic constraint in adopted library refers to grammer based on natural language and syntactic relation and Chinese character during being combined into syllables by syllable
The dependence of inter-syllable.It is influenced by the association of study front and back syllable, reduces the case where misrecognition is at single syllable.
In one embodiment, as shown in figure 5, when lid speech characteristic authentication information is matched with lid speech characteristic registration information,
It obtains user to be certified and passes through the authentication result of certification, comprising: step 502, obtain lid speech characteristic authentication information and lid speech characteristic
Similarity between registration information；Step 504, when the similarity between lid speech characteristic authentication information and lid speech characteristic registration information
When greater than preset threshold, obtains user to be certified and pass through the authentication result of certification.By lid speech characteristic authentication information and lid speech characteristic
Registration information is respectively converted into the form of feature vector, and the phase of the two can be calculated by Euclidean distance or cosine similarity
Like degree, wherein Euclidean distance refers to the natural length in actual distance or vector in m-dimensional space between two points, two
Euclidean distance in peacekeeping three-dimensional space is exactly the actual range between two o'clock；Cosine similarity refers to by calculating two vectors
Included angle cosine value come the similarity both assessed.
It should be understood that although each step in the flow chart of Fig. 1-5 is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, these steps can execute in other order.Moreover, at least one in Fig. 1-5
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, the execution sequence in these sub-steps or stage is also not necessarily successively
It carries out, but can be at least part of the sub-step or stage of other steps or other steps in turn or alternately
It executes.
In one embodiment, as shown in fig. 6, providing a kind of identification authentication system, comprising: certification request obtains module
602, image sequence extraction module 604, feature vector obtain module 606, registration information obtains module 608 and authentication module 610.
Certification request obtains module, and for obtaining the certification request of user to be certified, certification request carries user's mark of user to be certified
The certification video of knowledge and user to be certified；Image sequence extraction module, for carrying out Face datection and face to certification video
The detection of portion's characteristic point, obtains lip image sequence；Feature vector obtains module, mentions for carrying out feature to lip image sequence
It takes, obtains the feature vector of lip image sequence, using feature vector as lid speech characteristic authentication information；Registration information obtains module,
For searching pre-set user registration information according to user identifier, lid speech characteristic registration information corresponding with user identifier is obtained；Recognize
Module is demonstrate,proved, for user to be certified being obtained and passing through certification when lid speech characteristic authentication information is matched with lid speech characteristic registration information
Authentication result.
In one embodiment, it includes video block division unit that feature vector, which obtains module, is used for lip image sequence
Corresponding video block is divided into multiple sub-video blocks；Vector extraction unit, for each sub-video block to be passed through first solely respectively
Vertical subspace analysis carries out characteristic vector pickup, obtains first eigenvector set；Dimensionality reduction unit, for first eigenvector
Feature vector in set carries out dimension-reduction treatment, and the feature vector after dimension-reduction treatment is inputted the second independence subspace analysis, is obtained
To second feature vector set；Output unit, for obtaining mouth according to first eigenvector set and second feature vector set
The feature vector of lip image sequence.
It in one embodiment, further include that model obtains module before feature vector obtains module, for obtaining video sample
Originally and Recognition with Recurrent Neural Network model, video sample include lip reading image and lip reading recognition result corresponding with lip reading image；
Model training module, for using lip reading image as the input of Recognition with Recurrent Neural Network model, using lip reading recognition result as circulation
The output of neural network model is trained Recognition with Recurrent Neural Network model, the Recognition with Recurrent Neural Network model trained；It is special
Sign vector obtains module and is also used to the Recognition with Recurrent Neural Network model by having trained to the progress feature extraction of lip image sequence.
It in one embodiment, further include that syllable adhesion obtains module before model training module, for video sample
Image procossing is carried out, syllable adhesion data are obtained；Model training module be used for using lip reading image and syllable adhesion data as
The input of Recognition with Recurrent Neural Network model, using lip reading recognition result as the output of Recognition with Recurrent Neural Network model, to circulation nerve net
Network model is trained, the Recognition with Recurrent Neural Network model trained.
In one embodiment, it includes the first conversion unit that syllable adhesion, which obtains module, for converting video sample to
Mouth sequence picture；Mouth-Shape Recognition unit carries out Mouth-Shape Recognition for lip-sync sequence of pictures, obtains mouth sequence；Second turn
Change unit, for converting syllable sequence for mouth sequence, based on default semantic base and syllable sequence, obtains syllable adhesion number
According to.
In one embodiment, authentication module includes similarity acquiring unit, for obtain lid speech characteristic authentication information with
Similarity between lid speech characteristic registration information；Similarity-rough set unit, for working as lid speech characteristic authentication information and lid speech characteristic
When similarity between registration information is greater than preset threshold, obtains user to be certified and pass through the authentication result of certification.
Specific about identification authentication system limits the restriction that may refer to above for identity identifying method, herein not
It repeats again.Modules in above-mentioned identification authentication system can be realized fully or partially through software, hardware and combinations thereof.On
Stating each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also store in a software form
In memory in computer equipment, the corresponding operation of the above modules is executed in order to which processor calls.
In one embodiment, a kind of computer equipment is provided, which can be terminal, internal structure
Figure can be as shown in Figure 7.The computer equipment includes processor, the memory, network interface, display connected by system bus
Screen and input unit.Wherein, the processor of the computer equipment is for providing calculating and control ability.The computer equipment is deposited
Reservoir includes non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system and computer journey
Sequence.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The network interface of machine equipment is used to communicate with external terminal by network connection.When the computer program is executed by processor with
Realize a kind of identity identifying method.The display screen of the computer equipment can be liquid crystal display or electric ink display screen,
The input unit of the computer equipment can be the touch layer covered on display screen, be also possible to be arranged on computer equipment shell
Key, trace ball or Trackpad, can also be external keyboard, Trackpad or mouse etc..
It will be understood by those skilled in the art that structure shown in Fig. 7, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, which is stored with
Computer program, the processor perform the steps of the certification request for obtaining user to be certified when executing computer program, authenticate
Request carries the user identifier of user to be certified and the certification video of user to be certified；To certification video carry out Face datection with
And face feature point detection, obtain lip image sequence；Feature extraction is carried out to lip image sequence, obtains lip image sequence
Feature vector, using feature vector as lid speech characteristic authentication information；Pre-set user registration information is searched according to user identifier, is obtained
Lid speech characteristic registration information corresponding with user identifier；When lid speech characteristic authentication information is matched with lid speech characteristic registration information,
It obtains user to be certified and passes through the authentication result of certification.
In one embodiment, it also performs the steps of when processor executes computer program by lip image sequence pair
The video block answered is divided into multiple sub-video blocks；Each sub-video block is passed through into the first independence subspace analysis respectively and carries out feature
Vector extracts, and obtains first eigenvector set；Dimension-reduction treatment is carried out to the feature vector in first eigenvector set, will be dropped
Dimension treated feature vector inputs the second independence subspace analysis, obtains second feature vector set；According to fisrt feature to
Duration set and second feature vector set, obtain the feature vector of lip image sequence.
In one embodiment, processor execute computer program when also perform the steps of obtain video sample and
Recognition with Recurrent Neural Network model, video sample include lip reading image and lip reading recognition result corresponding with lip reading image；By lip reading
Input of the image as Recognition with Recurrent Neural Network model, using lip reading recognition result as the output of Recognition with Recurrent Neural Network model, to following
Ring neural network model is trained, the Recognition with Recurrent Neural Network model trained；Pass through the Recognition with Recurrent Neural Network trained
Model carries out feature extraction to lip image sequence.
In one embodiment, it is also performed the steps of when processor executes computer program and figure is carried out to video sample
As processing, syllable adhesion data are obtained；Using lip reading image and syllable adhesion data as the input of Recognition with Recurrent Neural Network model,
Using lip reading recognition result as the output of Recognition with Recurrent Neural Network model, Recognition with Recurrent Neural Network model is trained, has been instructed
Experienced Recognition with Recurrent Neural Network model.
In one embodiment, it is also performed the steps of when processor executes computer program and converts video sample to
Mouth sequence picture；Lip-sync sequence of pictures carries out Mouth-Shape Recognition, obtains mouth sequence；Syllable sequence is converted by mouth sequence
Column obtain syllable adhesion data based on default semantic base and syllable sequence.
In one embodiment, it is also performed the steps of when processor executes computer program and obtains lid speech characteristic certification
Similarity between information and lid speech characteristic registration information；When between lid speech characteristic authentication information and lid speech characteristic registration information
When similarity is greater than preset threshold, obtains user to be certified and pass through the authentication result of certification.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program performs the steps of the certification request for obtaining user to be certified when being executed by processor, certification request carries to be certified
The certification video of the user identifier of user and user to be certified；Face datection and face feature point inspection are carried out to certification video
It surveys, obtains lip image sequence；Feature extraction is carried out to lip image sequence, obtains the feature vector of lip image sequence, with
Feature vector is lid speech characteristic authentication information；Pre-set user registration information is searched according to user identifier, is obtained and user identifier pair
The lid speech characteristic registration information answered；When lid speech characteristic authentication information is matched with lid speech characteristic registration information, use to be certified is obtained
The authentication result that family passes through certification.
In one embodiment, it is also performed the steps of when computer program is executed by processor by lip image sequence
Corresponding video block is divided into multiple sub-video blocks；Each sub-video block is passed through into the first independence subspace analysis respectively and carries out spy
It levies vector to extract, obtains first eigenvector set；Dimension-reduction treatment is carried out to the feature vector in first eigenvector set, it will
Feature vector after dimension-reduction treatment inputs the second independence subspace analysis, obtains second feature vector set；According to fisrt feature
Vector set and second feature vector set, obtain the feature vector of lip image sequence.
In one embodiment, when computer program is executed by processor also perform the steps of obtain video sample with
And Recognition with Recurrent Neural Network model, video sample include lip reading image and lip reading recognition result corresponding with lip reading image；By lip
Input of the sonagram picture as Recognition with Recurrent Neural Network model, it is right using lip reading recognition result as the output of Recognition with Recurrent Neural Network model
Recognition with Recurrent Neural Network model is trained, the Recognition with Recurrent Neural Network model trained；Pass through the circulation nerve net trained
Network model carries out feature extraction to lip image sequence.
In one embodiment, it is also performed the steps of when computer program is executed by processor and video sample is carried out
Image procossing obtains syllable adhesion data；Using lip reading image and syllable adhesion data as the defeated of Recognition with Recurrent Neural Network model
Enter, using lip reading recognition result as the output of Recognition with Recurrent Neural Network model, Recognition with Recurrent Neural Network model is trained, is obtained
Trained Recognition with Recurrent Neural Network model.
In one embodiment, it is also performed the steps of when computer program is executed by processor and converts video sample
For mouth sequence picture；Lip-sync sequence of pictures carries out Mouth-Shape Recognition, obtains mouth sequence；Syllable sequence is converted by mouth sequence
Column obtain syllable adhesion data based on default semantic base and syllable sequence.
In one embodiment, acquisition lid speech characteristic is also performed the steps of when computer program is executed by processor to recognize
Demonstrate,prove the similarity between information and lid speech characteristic registration information；When between lid speech characteristic authentication information and lid speech characteristic registration information
Similarity be greater than preset threshold when, obtain user to be certified by certification authentication result.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of identity identifying method, which comprises
The certification request of user to be certified is obtained, the certification request carries the user identifier and use to be certified of user to be certified
The certification video at family；
Face datection and face feature point detection are carried out to the certification video, obtain lip image sequence；
Feature extraction is carried out to the lip image sequence, the feature vector of the lip image sequence is obtained, with the feature
Vector is lid speech characteristic authentication information；
Pre-set user registration information is searched according to the user identifier, obtains lid speech characteristic registration corresponding with the user identifier
Information；
When the lid speech characteristic authentication information is matched with the lid speech characteristic registration information, obtains user to be certified and pass through certification
Authentication result.
2. the method according to claim 1, wherein it is described to the lip image sequence carry out feature extraction,
Obtain the feature vector of the lip image sequence, comprising:
The corresponding video block of the lip image sequence is divided into multiple sub-video blocks；Each sub-video block is passed through respectively
One independence subspace analysis carries out characteristic vector pickup, obtains first eigenvector set；
Dimension-reduction treatment is carried out to the feature vector in the first eigenvector set, the feature vector after dimension-reduction treatment is inputted
Second independence subspace analysis obtains second feature vector set；
According to the first eigenvector set and the second feature vector set, the feature of the lip image sequence is obtained
Vector.
3. the method according to claim 1, wherein it is described to the lip image sequence carry out feature extraction it
Before, further includes:
Obtain video sample and Recognition with Recurrent Neural Network model, the video sample include lip reading image and with the lip reading figure
As corresponding lip reading recognition result；
Using the lip reading image as the input of the Recognition with Recurrent Neural Network model, the lip reading recognition result is followed as described in
The output of ring neural network model is trained the Recognition with Recurrent Neural Network model, the Recognition with Recurrent Neural Network trained
Model；
It is described that feature extraction is carried out to the lip image sequence, comprising:
Feature extraction is carried out to the lip image sequence by the Recognition with Recurrent Neural Network model trained.
4. according to the method described in claim 3, it is characterized in that, described using the lip reading image as the circulation nerve net
The input of network model, using the lip reading recognition result as the output of the Recognition with Recurrent Neural Network model, to the circulation nerve
Network model is trained, before the Recognition with Recurrent Neural Network model trained, further includes:
Image procossing is carried out to the video sample, obtains syllable adhesion data；
It is described using the lip reading image as the input of the Recognition with Recurrent Neural Network model, using the lip reading recognition result as institute
The output for stating Recognition with Recurrent Neural Network model is trained the Recognition with Recurrent Neural Network model, the circulation nerve trained
Network model, comprising:
Using the lip reading image and the syllable adhesion data as the input of the Recognition with Recurrent Neural Network model, by the lip
Output of the language recognition result as the Recognition with Recurrent Neural Network model, is trained the Recognition with Recurrent Neural Network model, obtains
The Recognition with Recurrent Neural Network model trained.
5. according to the method described in claim 4, it is characterized in that, it is described to the video sample carry out image procossing, obtain
Syllable adhesion data, comprising:
Mouth sequence picture is converted by the video sample；
Mouth-Shape Recognition is carried out to the mouth sequence picture, obtains mouth sequence；
Syllable sequence is converted by the mouth sequence, based on default semantic base and the syllable sequence, obtains syllable adhesion
Data.
6. the method according to claim 1, wherein described work as the lid speech characteristic authentication information and the lip reading
When feature registration information matches, obtains user to be certified and passes through the authentication result of certification, comprising:
Obtain the similarity between the lid speech characteristic authentication information and the lid speech characteristic registration information；
When the similarity between the lid speech characteristic authentication information and the lid speech characteristic registration information is greater than preset threshold, obtain
Pass through the authentication result of certification to user to be certified.
7. a kind of identification authentication system, which is characterized in that described device includes:
Certification request obtains module, and for obtaining the certification request of user to be certified, the certification request carries user to be certified
User identifier and user to be certified certification video；
Image sequence extraction module obtains lip figure for carrying out Face datection and face feature point detection to certification video
As sequence；
Feature vector obtains module, for carrying out feature extraction to lip image sequence, obtain the feature of lip image sequence to
Amount, using feature vector as lid speech characteristic authentication information；
Registration information obtains module, for searching pre-set user registration information according to the user identifier, obtains and the user
Identify corresponding lid speech characteristic registration information；
Authentication module, for obtaining wait recognize when the lid speech characteristic authentication information is matched with the lid speech characteristic registration information
It demonstrate,proves user and passes through the authentication result of certification.
8. device according to claim 7, which is characterized in that described eigenvector obtains module and includes:
Video block division unit, for the corresponding video block of lip image sequence to be divided into multiple sub-video blocks；
Vector extraction unit is mentioned for each sub-video block to be passed through the first independence subspace analysis progress feature vector respectively
It takes, obtains first eigenvector set；
Dimensionality reduction unit, for carrying out dimension-reduction treatment to the feature vector in first eigenvector set, by the spy after dimension-reduction treatment
It levies vector and inputs the second independence subspace analysis, obtain second feature vector set；
Output unit, for obtaining the spy of lip image sequence according to first eigenvector set and second feature vector set
Levy vector.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 6 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 6 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910304302.0A CN110113319A (en) | 2019-04-16 | 2019-04-16 | Identity identifying method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910304302.0A CN110113319A (en) | 2019-04-16 | 2019-04-16 | Identity identifying method, device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110113319A true CN110113319A (en) | 2019-08-09 |
Family
ID=67485500
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910304302.0A Pending CN110113319A (en) | 2019-04-16 | 2019-04-16 | Identity identifying method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110113319A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110955874A (en) * | 2019-10-12 | 2020-04-03 | 深圳壹账通智能科技有限公司 | Identity authentication method, identity authentication device, computer equipment and storage medium |
| CN111242029A (en) * | 2020-01-13 | 2020-06-05 | 湖南世优电气股份有限公司 | Device control method, device, computer device and storage medium |
| CN111507165A (en) * | 2020-01-21 | 2020-08-07 | 北京明略软件系统有限公司 | Face recognition method and device, electronic equipment and computer readable storage medium |
| CN111523405A (en) * | 2020-04-08 | 2020-08-11 | 绍兴埃瓦科技有限公司 | Face recognition method and system and electronic equipment |
| CN111666820A (en) * | 2020-05-11 | 2020-09-15 | 北京中广上洋科技股份有限公司 | Speaking state recognition method and device, storage medium and terminal |
| CN112055013A (en) * | 2020-09-01 | 2020-12-08 | 好活（昆山）网络科技有限公司 | Automatic authentication method, device, equipment and storage medium |
| CN112132095A (en) * | 2020-09-30 | 2020-12-25 | Oppo广东移动通信有限公司 | Dangerous state identification method and device, electronic equipment and storage medium |
| CN112417497A (en) * | 2020-11-11 | 2021-02-26 | 北京邮电大学 | Privacy protection method and device, electronic equipment and storage medium |
| CN113314125A (en) * | 2021-05-28 | 2021-08-27 | 深圳市展拓电子技术有限公司 | Voiceprint identification method, system and memory for monitoring room interphone |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016201679A1 (en) * | 2015-06-18 | 2016-12-22 | 华为技术有限公司 | Feature extraction method, lip-reading classification method, device and apparatus |
| CN106778496A (en) * | 2016-11-22 | 2017-05-31 | 重庆中科云丛科技有限公司 | Biopsy method and device |
| CN107404381A (en) * | 2016-05-19 | 2017-11-28 | 阿里巴巴集团控股有限公司 | A kind of identity identifying method and device |
| CN108491808A (en) * | 2018-03-28 | 2018-09-04 | 百度在线网络技术（北京）有限公司 | Method and device for obtaining information |
| CN109409195A (en) * | 2018-08-30 | 2019-03-01 | 华侨大学 | A kind of lip reading recognition methods neural network based and system |
-
2019
- 2019-04-16 CN CN201910304302.0A patent/CN110113319A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016201679A1 (en) * | 2015-06-18 | 2016-12-22 | 华为技术有限公司 | Feature extraction method, lip-reading classification method, device and apparatus |
| CN107404381A (en) * | 2016-05-19 | 2017-11-28 | 阿里巴巴集团控股有限公司 | A kind of identity identifying method and device |
| CN106778496A (en) * | 2016-11-22 | 2017-05-31 | 重庆中科云丛科技有限公司 | Biopsy method and device |
| CN108491808A (en) * | 2018-03-28 | 2018-09-04 | 百度在线网络技术（北京）有限公司 | Method and device for obtaining information |
| CN109409195A (en) * | 2018-08-30 | 2019-03-01 | 华侨大学 | A kind of lip reading recognition methods neural network based and system |
Non-Patent Citations (1)
| Title |
|---|
| 杨龙生等: "用于可靠身份认证的唇语识别", 《电视技术》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110955874A (en) * | 2019-10-12 | 2020-04-03 | 深圳壹账通智能科技有限公司 | Identity authentication method, identity authentication device, computer equipment and storage medium |
| CN111242029A (en) * | 2020-01-13 | 2020-06-05 | 湖南世优电气股份有限公司 | Device control method, device, computer device and storage medium |
| CN111507165A (en) * | 2020-01-21 | 2020-08-07 | 北京明略软件系统有限公司 | Face recognition method and device, electronic equipment and computer readable storage medium |
| CN111523405A (en) * | 2020-04-08 | 2020-08-11 | 绍兴埃瓦科技有限公司 | Face recognition method and system and electronic equipment |
| CN111666820A (en) * | 2020-05-11 | 2020-09-15 | 北京中广上洋科技股份有限公司 | Speaking state recognition method and device, storage medium and terminal |
| CN111666820B (en) * | 2020-05-11 | 2023-06-20 | 北京中广上洋科技股份有限公司 | Speech state recognition method and device, storage medium and terminal |
| CN112055013A (en) * | 2020-09-01 | 2020-12-08 | 好活（昆山）网络科技有限公司 | Automatic authentication method, device, equipment and storage medium |
| CN112132095A (en) * | 2020-09-30 | 2020-12-25 | Oppo广东移动通信有限公司 | Dangerous state identification method and device, electronic equipment and storage medium |
| CN112132095B (en) * | 2020-09-30 | 2024-02-09 | Oppo广东移动通信有限公司 | Dangerous state identification method and device, electronic equipment and storage medium |
| CN112417497A (en) * | 2020-11-11 | 2021-02-26 | 北京邮电大学 | Privacy protection method and device, electronic equipment and storage medium |
| CN112417497B (en) * | 2020-11-11 | 2023-04-25 | 北京邮电大学 | Privacy protection method, device, electronic equipment and storage medium |
| CN113314125A (en) * | 2021-05-28 | 2021-08-27 | 深圳市展拓电子技术有限公司 | Voiceprint identification method, system and memory for monitoring room interphone |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110113319A (en) | Identity identifying method, device, computer equipment and storage medium | |
| Yang et al. | Preventing deepfake attacks on speaker authentication by dynamic lip movement analysis | |
| EP3477519B1 (en) | Identity authentication method, terminal device, and computer-readable storage medium | |
| Gao et al. | Discriminative multiple canonical correlation analysis for information fusion | |
| CN104361276B (en) | A kind of multi-modal biological characteristic identity identifying method and system | |
| US6219639B1 (en) | Method and apparatus for recognizing identity of individuals employing synchronized biometrics | |
| US20210327431A1 (en) | 'liveness' detection system | |
| Tolosana et al. | Preprocessing and feature selection for improved sensor interoperability in online biometric signature verification | |
| Revett | Behavioral biometrics: a remote access approach | |
| WO2016172872A1 (en) | Method and device for verifying real human face, and computer program product | |
| CN106778450B (en) | Face recognition method and device | |
| CN111079791A (en) | Face recognition method, face recognition device and computer-readable storage medium | |
| JP7412496B2 (en) | Living body (liveness) detection verification method, living body detection verification system, recording medium, and training method for living body detection verification system | |
| Mandalapu et al. | Audio-visual biometric recognition and presentation attack detection: A comprehensive survey | |
| Saeed | New directions in behavioral biometrics | |
| Cheng et al. | Visual speaker authentication with random prompt texts by a dual-task CNN framework | |
| CN115511704B (en) | Virtual customer service generation method and device, electronic equipment and storage medium | |
| Fernández | New methodology for the integration of biometric features in speaker recognition systems applied to security environments | |
| CN116883900A (en) | Video authenticity identification method and system based on multidimensional biological characteristics | |
| Uzan et al. | I know that voice: Identifying the voice actor behind the voice | |
| Baró et al. | Biometric tools for learner identity in e-assessment | |
| Carneiro et al. | Whose emotion matters? Speaking activity localisation without prior knowledge | |
| Chen et al. | An Identity Authentication Method Based on Multi-modal Feature Fusion | |
| Czyz et al. | Scalability analysis of audio-visual person identity verification | |
| US12045333B1 (en) | Method and a device for user verification |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190809 |