CN106897700A - A kind of single sample face recognition method and system - Google Patents

A kind of single sample face recognition method and system Download PDF

Info

Publication number
CN106897700A
CN106897700A CN201710107890.XA CN201710107890A CN106897700A CN 106897700 A CN106897700 A CN 106897700A CN 201710107890 A CN201710107890 A CN 201710107890A CN 106897700 A CN106897700 A CN 106897700A
Authority
CN
China
Prior art keywords
subgraph
facial image
sample
training set
weight
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201710107890.XA
Other languages
Chinese (zh)
Other versions
CN106897700B (en
Inventor
张朦
张莉
王邦军
张召
李凡长
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou University
Original Assignee
Suzhou University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Suzhou University filed Critical Suzhou University
Priority to CN201710107890.XA priority Critical patent/CN106897700B/en
Publication of CN106897700A publication Critical patent/CN106897700A/en
Application granted granted Critical
Publication of CN106897700B publication Critical patent/CN106897700B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06V40/168—Feature extraction; Face representation

Landscapes

  • Engineering & Computer Science (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)
  • Collating Specific Patterns (AREA)
  • Image Processing (AREA)

Abstract

This application discloses a kind of single sample face recognition method and system, the method includes：Facial image to be identified is divided into t subgraph；Wherein, t is positive integer；Extract and the one-to-one t feature of t subgraph of facial image to be identified；T feature of facial image to be identified is input into the single sample human face recognition model for being in advance based on training set and being obtained with reference to the conjunction of subgraph weight sets, the corresponding classification of facial image to be identified is obtained；Wherein, subgraph weight sets is combined into the set that the between class distance of the face images sample in training set obtained after maximization treatment；R-th weight in subgraph weight set is the corresponding weight of r-th subgraph of any facial image sample in training set.The application effectively improves the discrimination of single sample face recognition technology.

Description

A kind of single sample face recognition method and system
Technical field
The present invention relates to technical field of face recognition, more particularly to a kind of single sample face recognition method and system.
Background technology
At present, with the fast development of artificial intelligence and computer technology, and people's demand is growing and lasting Change, face recognition technology has obtained more and more extensive concern.Traditional face identification method is had been generally acknowledged that in training process In multiple facial image samples can be used carry out the extraction of characteristic value, but, personally for each, the people in real life A facial image, such as enforcement, E-Passport and ID card verification are generally only stored in face system.Thus in these feelings , it is necessary to the performance of conventional face's recognizer of multisample is often not fully up to expectations under condition.
Therefore, there has been proposed single sample face recognition technology, so-called single sample face recognition technology refers to according to every People only stores the face database of a width known identities to identify the identity corresponding to new facial image.However, existing Single sample face recognition technology still there is a problem of that discrimination is poor, how to lift the identification of single sample face recognition technology Rate is the problem for needing further solution at present.
The content of the invention
In view of this, it is an object of the invention to provide a kind of single sample face recognition method and system, can be effectively The discrimination of the single sample face recognition technology of lifting.Its concrete scheme is as follows：
A kind of single sample face recognition method, including：
Facial image to be identified is divided into t subgraph；Wherein, t is positive integer；
Extract and the one-to-one t feature of t subgraph of the facial image to be identified；
The t feature input of the facial image to be identified is in advance based on training set and is closed with reference to subgraph weight sets to obtain Single sample human face recognition model, obtain the corresponding classification of the facial image to be identified；
Wherein, the subgraph weight sets is combined into the between class distance of the face images sample in the training set and carries out The set obtained after maximization treatment；R-th weight in the subgraph weight set is any face figure in the training set The corresponding weight of r-th subgraph of decent, r=1,2 ..., t.
Optionally, the establishment process of single sample human face recognition model, including：
Each facial image sample standard deviation in the training set is divided into t subgraph；
Extract respectively and the one-to-one t feature of t subgraph of each facial image sample in the training set；
Using default apart from computing formula, r-th son of face images sample in the training set is calculated respectively The average between class distance of figure, r=1,2 ..., t；Wherein, it is described to be apart from computing formula：R=1, 2 ..., t, in formula, drRepresent the average between class distance of r-th subgraph of face images sample in the training set, c tables Show the sum of facial image sample in the training set, hirRepresent i-th r-th of facial image sample in the training set The feature of subgraph, Represent hirWithBetween manhatton distance；
Using object function, the subgraph weight set is determined；Wherein, the object function is： Also, s.t.0≤ωr≤ 1, r=1,2 ..., t, in formula, ω represents the subgraph weight set, ωrRepresent r-th in ω Weight, d=[d1,d2,...,dt]T；
Feature and the subgraph weight sets corresponding to subgraph based on each facial image sample in the training set Close, construct single sample human face recognition model；Wherein, single sample human face recognition model is specially：
In formula, y represents the corresponding classification of the facial image to be identified, d (hir,hr) represent hirAnd hrBetween Manhattan Distance.
Optionally, the one-to-one t of t subgraph extracted respectively with each facial image sample in the training set The process of individual feature, including：
Extract respectively and the one-to-one t LBP features of t subgraph of each facial image sample in the training set.
Optionally, the process extracted with the one-to-one t feature of t subgraph of the facial image to be identified, Including：
Extract and the one-to-one t LBP features of t subgraph of the facial image to be identified.
The present invention further correspondingly discloses a kind of single sample face identification system, including：
Model creation module, closes for being in advance based on training set and combining subgraph weight sets, obtains single sample recognition of face Model；
Image division module, for facial image to be identified to be divided into t subgraph；Wherein, t is positive integer；
Characteristic extracting module, for extracting and the one-to-one t feature of t subgraph of the facial image to be identified；
Category determination module, for t feature of the facial image to be identified to be input into single sample recognition of face Model, obtains the corresponding classification of the facial image to be identified；
Wherein, the subgraph weight sets is combined into the between class distance of the face images sample in the training set and carries out The set obtained after maximization treatment；R-th weight in the subgraph weight set is any face figure in the training set The corresponding weight of r-th subgraph of decent, r=1,2 ..., t.
Optionally, the model creation module, including：
Image division unit, for each the facial image sample standard deviation in the training set to be divided into t subgraph；
Feature extraction unit, for extracting respectively with t subgraph of each facial image sample in the training set one by one Corresponding t feature；
Metrics calculation unit, for using default apart from computing formula, all faces in the training set being calculated respectively The average between class distance of r-th subgraph of image pattern, r=1,2 ..., t；Wherein, it is described to be apart from computing formula：R=1,2 ..., t, in formula, drRepresent r-th son of face images sample in the training set The average between class distance of figure, c represents the sum of facial image sample in the training set, hirRepresent i-th in the training set The feature of r-th subgraph of individual facial image sample, Represent hirWithBetween manhatton distance；
Weight calculation unit, for utilizing object function, determines the subgraph weight set；Wherein, the object function For：Also, s.t.0≤ωr≤ 1, r=1,2 ..., t, in formula, ω represents the subgraph weight sets Close, ωrRepresent r-th weight in ω, d=[d1,d2,...,dt]T；
Construction of A Model unit, for the feature corresponding to the subgraph based on each facial image sample in the training set with And the subgraph weight set, construct single sample human face recognition model；Wherein, single sample human face recognition model is specific For：
In formula, y represents the corresponding classification of the facial image to be identified, d (hir,hr) represent hirAnd hrBetween Manhattan Distance.
Optionally, the feature extraction unit is：
LBP feature extraction units, for extracting the t subgraph one with each facial image sample in the training set respectively One corresponding t LBP features.
Optionally, the characteristic extracting module is：
LBP characteristic extracting modules, for extracting t LBP one-to-one with t subgraph of the facial image to be identified Feature.
In the present invention, single sample face recognition method, including：Facial image to be identified is divided into t subgraph；Wherein, t It is positive integer；Extract and the one-to-one t feature of t subgraph of facial image to be identified；By the t of facial image to be identified Feature input is in advance based on training set and the single sample human face recognition model obtained with reference to the conjunction of subgraph weight sets, obtains people to be identified The corresponding classification of face image；Wherein, subgraph weight sets is combined into the between class distance of the face images sample in training set and enters The set obtained after row maximization treatment；R-th weight in subgraph weight set is any facial image sample in training set The corresponding weight of r-th subgraph, r=1,2 ..., t.
It can be seen that, the present invention is to utilize the single sample recognition of face for being in advance based on training set and being obtained with reference to the conjunction of subgraph weight sets Model is identified to facial image to be identified, also, the subgraph weight that is related in above-mentioned single sample human face recognition model Set is that the between class distance of the face images sample in training set obtained after maximization treatment, so can The effectively difference degree in increase training set between inhomogeneity, so that based on the single sample obtained by the subgraph weight set Human face recognition model can more accurately be recognized to facial image to be identified, namely effectively improve list in the present invention The discrimination of sample face recognition technology.
Brief description of the drawings
In order to illustrate more clearly about the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing The accompanying drawing to be used needed for having technology description is briefly described, it should be apparent that, drawings in the following description are only this Inventive embodiment, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis The accompanying drawing of offer obtains other accompanying drawings.
Fig. 1 is a kind of single sample face recognition method flow chart disclosed in the embodiment of the present invention；
Fig. 2 is a kind of visioning procedure figure of single sample human face recognition model disclosed in the embodiment of the present invention；
Fig. 3 is a kind of single sample face identification system structural representation disclosed in the embodiment of the present invention.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than whole embodiments.It is based on Embodiment in the present invention, it is every other that those of ordinary skill in the art are obtained under the premise of creative work is not made Embodiment, belongs to the scope of protection of the invention.
The embodiment of the invention discloses a kind of single sample face recognition method, shown in Figure 1, the method includes：
Step S11：Facial image to be identified is divided into t subgraph；Wherein, t is positive integer；
Step S12：Extract and the one-to-one t feature of t subgraph of facial image to be identified；
Step S13：It is in advance based on the t feature input of facial image to be identified training set and combines subgraph weight sets to close The single sample human face recognition model for obtaining, obtains the corresponding classification of facial image to be identified；
Wherein, subgraph weight sets is combined into the between class distance of the face images sample in training set and carries out at maximization The set obtained after reason；R-th weight in subgraph weight set is r-th subgraph of any facial image sample in training set Corresponding weight, r=1,2 ..., t.
Specifically, shown in Figure 2, the establishment process of the single sample human face recognition model in above-mentioned steps S13, can wrap Include below step S131 to S135：
Step S131：Each facial image sample standard deviation in training set is divided into t subgraph；
Step S132：T spy one-to-one with t subgraph of each facial image sample in training set is extracted respectively Levy；
Step S133：Using default apart from computing formula, the r of face images sample in training set is calculated respectively The average between class distance of individual subgraph, r=1,2 ..., t；Wherein, it is apart from computing formula：R=1, 2 ..., t, in formula, drThe average between class distance of r-th subgraph of face images sample in training set is represented, c represents instruction Practice the sum for concentrating facial image sample, hirThe feature of r-th subgraph of i-th facial image sample in expression training set, Represent hirWithBetween manhatton distance；
Step S134：Using object function, subgraph weight set is determined；Wherein, object function is：Also, s.t.0≤ωr≤ 1, r=1,2 ..., t, in formula, ω represents subgraph weight set, ωrTable Show r-th weight in ω, d=[d1,d2,...,dt]T；
Step S135：Based on the feature corresponding to the subgraph of each facial image sample in training set and subgraph weight sets Close, the single sample human face recognition model of construction；Wherein, single sample human face recognition model is specially：
In formula, y represents the corresponding classification of facial image to be identified, d (hir,hr) represent hirAnd hrBetween Manhattan away from From.
In the present embodiment, the feature of each subgraph preferentially using LBP features (LBP, i.e. Local Binary Pattern, Local binary patterns).That is, in above-mentioned steps S132, the t subgraph with each facial image sample in training set is extracted respectively The one-to-one t process of feature, specifically includes：The t subgraph with each facial image sample in training set is extracted respectively One-to-one t LBP feature.
Accordingly, in above-mentioned steps S12, extract and the one-to-one t feature of t subgraph of facial image to be identified Process, specifically includes：Extract and the one-to-one t LBP features of t subgraph of facial image to be identified.
It can be seen that, the embodiment of the present invention is to utilize the single sample people for being in advance based on training set and being obtained with reference to the conjunction of subgraph weight sets Face identification model is identified to facial image to be identified, also, the son that is related in above-mentioned single sample human face recognition model Figure weight set is that the between class distance of the face images sample in training set obtained after maximization treatment, this Sample can effectively increase the difference degree between inhomogeneity in training set, so that obtained by based on the subgraph weight set Single sample human face recognition model can more accurately be recognized to facial image to be identified, namely effectively improve this hair The discrimination of bright middle single sample face recognition technology.
The embodiment of the present invention describes single sample recognition of face mould in a upper embodiment in detail by way of concrete example Process and corresponding face recognition process that type is created.
Assuming that training set includes 100 facial image samples of people, namely including 100 facial image samples, then this The building process of single sample human face recognition model is specific as follows in embodiment：
Each facial image sample standard deviation in above-mentioned training set is divided into 36 subgraphs, is then extracted respectively and everyone The one-to-one 36 LBP features of 36 subgraphs of face image sample, so, i-th LBP eigenmatrix of facial image sample Can be expressed as：
Hi=[hi1,hi2,...,hi36]；
That is, in the present embodiment, hir∈R256Represent r-th subgraph of i-th facial image sample in above-mentioned training set LBP features, r=1,2 ..., 36.
Using following apart from computing formula：All faces in above-mentioned training set are calculated respectively The average between class distance of r-th subgraph of image pattern, r=1,2 ..., 36；Wherein,Represent face images sample R-th average of the LBP features of subgraph, namely Represent hirWithBetween manhatton distance.
Then, the between class distance to face images sample in training set carries out maximization treatment, to obtain subgraph power Gather again, namely solve following object function：
Also, s.t.0≤ωr≤ 1, r=1,2 ..., 36；
In formula, ω represents subgraph weight set, ωrRepresent r-th weight in ω, d=[d1,d2,...,d36]T；
Feature and subgraph weight set corresponding to subgraph based on each facial image sample in above-mentioned training set, can Construct following single sample human face recognition model：
In formula, y represents the corresponding classification of facial image to be identified, d (hir,hr) represent hirAnd hrBetween Manhattan away from From.
Thus, after facial image to be identified is got, just the facial image to be identified can be divided into 36 subgraphs, Then extract and the one-to-one 36 LBP features of 36 subgraphs, then can be above-mentioned by 36 LBP features inputs Single sample human face recognition model that establishment is finished, so as to obtain corresponding recognition result.
Accordingly, it is shown in Figure 3 the embodiment of the invention also discloses a kind of single sample face identification system, including：
Model creation module 11, closes for being in advance based on training set and combining subgraph weight sets, obtains single sample face and knows Other model；
Image division module 12, for facial image to be identified to be divided into t subgraph；Wherein, t is positive integer；
Characteristic extracting module 13, for extracting and the one-to-one t feature of t subgraph of facial image to be identified；
Category determination module 14, for by the single sample human face recognition model of t feature input of facial image to be identified, obtaining To the corresponding classification of facial image to be identified；
Wherein, above-mentioned subgraph weight sets is combined into the between class distance of the face images sample in training set and carries out maximum The set obtained after change treatment；R-th weight in above-mentioned subgraph weight set is any facial image sample in training set The corresponding weight of r-th subgraph, r=1,2 ..., t.
Specifically, above-mentioned model creation module 11 can include that image division unit, feature extraction unit, distance calculate single Unit, weight calculation unit and Construction of A Model unit；Wherein,
Image division unit, for each the facial image sample standard deviation in training set to be divided into t subgraph；
Feature extraction unit, corresponds for extracting respectively with t subgraph of each facial image sample in training set T feature；
Metrics calculation unit, apart from computing formula, face images in training set are calculated for using default respectively The average between class distance of r-th subgraph of sample, r=1,2 ..., t；Wherein, it is apart from computing formula：R=1,2 ..., t, in formula, drRepresent r-th subgraph of face images sample in training set Average between class distance, c represents the sum of facial image sample in training set, hirRepresent i-th facial image sample in training set The feature of this r-th subgraph, Represent hirWithBetween manhatton distance；
Weight calculation unit, for utilizing object function, determines subgraph weight set；Wherein, object function is：Also, s.t.0≤ωr≤ 1, r=1,2 ..., t, in formula, ω represents subgraph weight set, ωrTable Show r-th weight in ω, d=[d1,d2,...,dt]T；
Construction of A Model unit, for based on the feature and son corresponding to the subgraph of each facial image sample in training set Figure weight set, the single sample human face recognition model of construction；Wherein, single sample human face recognition model is specially：
In formula, y represents the corresponding classification of facial image to be identified, d (hir,hr) represent hirAnd hrBetween Manhattan away from From.
Wherein, features described above extraction unit is specifically as follows：
LBP feature extraction units are right one by one with t subgraph of each facial image sample in training set for extracting respectively The t LBP feature answered.
Accordingly, features described above extraction module 13 is specifically as follows：
LBP characteristic extracting modules, it is special for extracting t LBP one-to-one with t subgraph of facial image to be identified Levy.
It can be seen that, the embodiment of the present invention is to utilize the single sample people for being in advance based on training set and being obtained with reference to the conjunction of subgraph weight sets Face identification model is identified to facial image to be identified, also, the son that is related in above-mentioned single sample human face recognition model Figure weight set is that the between class distance of the face images sample in training set obtained after maximization treatment, this Sample can effectively increase the difference degree between inhomogeneity in training set, so that obtained by based on the subgraph weight set Single sample human face recognition model can more accurately be recognized to facial image to be identified, namely effectively improve this hair The discrimination of bright middle single sample face recognition technology.
Finally, in addition it is also necessary to explanation, herein, such as first and second or the like relational terms be used merely to by One entity or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or operation Between there is any this actual relation or order.And, term " including ", "comprising" or its any other variant meaning Covering including for nonexcludability, so that process, method, article or equipment including a series of key elements not only include that A little key elements, but also other key elements including being not expressly set out, or also include for this process, method, article or The intrinsic key element of equipment.In the absence of more restrictions, the key element limited by sentence "including a ...", does not arrange Except also there is other identical element in the process including the key element, method, article or equipment.
A kind of single sample face recognition method provided by the present invention and system are described in detail above, herein Apply specific case to be set forth principle of the invention and implementation method, the explanation of above example is only intended to help Understand the method for the present invention and its core concept；Simultaneously for those of ordinary skill in the art, according to thought of the invention, Will change in specific embodiments and applications, in sum, this specification content should not be construed as to this The limitation of invention.

Claims (8)

1. a kind of single sample face recognition method, it is characterised in that including：
Facial image to be identified is divided into t subgraph；Wherein, t is positive integer；
Extract and the one-to-one t feature of t subgraph of the facial image to be identified；
T feature of the facial image to be identified is input into the list for being in advance based on training set and being obtained with reference to the conjunction of subgraph weight sets Sample human face recognition model, obtains the corresponding classification of the facial image to be identified；
Wherein, the subgraph weight sets is combined into the between class distance of the face images sample in the training set and carries out maximum The set obtained after change treatment；R-th weight in the subgraph weight set is any facial image sample in the training set This corresponding weight of r-th subgraph, r=1,2 ..., t.
2. single sample face recognition method according to claim 1, it is characterised in that single sample human face recognition model Establishment process, including：
Each facial image sample standard deviation in the training set is divided into t subgraph；
Extract respectively and the one-to-one t feature of t subgraph of each facial image sample in the training set；
Using default apart from computing formula, r-th subgraph of face images sample in the training set is calculated respectively Average between class distance, r=1,2 ..., t；Wherein, it is described to be apart from computing formula：R=1,2 ..., T, in formula, drThe average between class distance of r-th subgraph of face images sample in the training set is represented, c represents described The sum of facial image sample, h in training setirRepresent r-th subgraph of i-th facial image sample in the training set Feature, Represent hirWithBetween manhatton distance；
Using object function, the subgraph weight set is determined；Wherein, the object function is：And And, s.t.0≤ωr≤ 1, r=1,2 ..., t, in formula, ω represents the subgraph weight set, ωrRepresent r-th power in ω Weight, d=[d1,d2,...,dt]T；
Feature and the subgraph weight set corresponding to subgraph based on each facial image sample in the training set, structure Make single sample human face recognition model；Wherein, single sample human face recognition model is specially：
y = argmin i = 1 , 2 , ... , c Σ r = 1 t ω r d ( h i r , h r ) ;
In formula, y represents the corresponding classification of the facial image to be identified, d (hir,hr) represent hirAnd hrBetween Manhattan away from From.
3. single sample face recognition method according to claim 2, it is characterised in that described to extract respectively and the training The process of the one-to-one t feature of t subgraph of each facial image sample is concentrated, including：
Extract respectively and the one-to-one t LBP features of t subgraph of each facial image sample in the training set.
4. single sample face recognition method according to claim 3, it is characterised in that the extraction and the people to be identified The process of the one-to-one t feature of t subgraph of face image, including：
Extract and the one-to-one t LBP features of t subgraph of the facial image to be identified.
5. a kind of single sample face identification system, it is characterised in that including：
Model creation module, closes for being in advance based on training set and combining subgraph weight sets, obtains single sample human face recognition model；
Image division module, for facial image to be identified to be divided into t subgraph；Wherein, t is positive integer；
Characteristic extracting module, for extracting and the one-to-one t feature of t subgraph of the facial image to be identified；
Category determination module, for t feature of the facial image to be identified to be input into single sample human face recognition model, Obtain the corresponding classification of the facial image to be identified；
Wherein, the subgraph weight sets is combined into the between class distance of the face images sample in the training set and carries out maximum The set obtained after change treatment；R-th weight in the subgraph weight set is any facial image sample in the training set This corresponding weight of r-th subgraph, r=1,2 ..., t.
6. single sample face identification system according to claim 5, it is characterised in that the model creation module, including：
Image division unit, for each the facial image sample standard deviation in the training set to be divided into t subgraph；
Feature extraction unit, corresponds for extracting respectively with t subgraph of each facial image sample in the training set T feature；
Metrics calculation unit, for using default apart from computing formula, face images in the training set being calculated respectively The average between class distance of r-th subgraph of sample, r=1,2 ..., t；Wherein, it is described to be apart from computing formula：R=1,2 ..., t, in formula, drRepresent r-th son of face images sample in the training set The average between class distance of figure, c represents the sum of facial image sample in the training set, hirRepresent i-th in the training set The feature of r-th subgraph of individual facial image sample, Represent hirWithBetween manhatton distance；
Weight calculation unit, for utilizing object function, determines the subgraph weight set；Wherein, the object function is：Also, s.t.0≤ωr≤ 1, r=1,2 ..., t, in formula, ω represents the subgraph weight set, ωrRepresent r-th weight in ω, d=[d1,d2,...,dt]T；
Construction of A Model unit, for the feature corresponding to the subgraph based on each facial image sample in the training set and institute Subgraph weight set is stated, single sample human face recognition model is constructed；Wherein, single sample human face recognition model is specially：
y = argmin i = 1 , 2 , ... , c Σ r = 1 t ω r d ( h i r , h r ) ;
In formula, y represents the corresponding classification of the facial image to be identified, d (hir,hr) represent hirAnd hrBetween Manhattan away from From.
7. single sample face identification system according to claim 6, it is characterised in that the feature extraction unit is：
LBP feature extraction units are right one by one with t subgraph of each facial image sample in the training set for extracting respectively The t LBP feature answered.
8. single sample face identification system according to claim 7, it is characterised in that the characteristic extracting module is：
LBP characteristic extracting modules, it is special for extracting t LBP one-to-one with t subgraph of the facial image to be identified Levy.
CN201710107890.XA 2017-02-27 2017-02-27 Single-sample face recognition method and system Active CN106897700B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710107890.XA CN106897700B (en) 2017-02-27 2017-02-27 Single-sample face recognition method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710107890.XA CN106897700B (en) 2017-02-27 2017-02-27 Single-sample face recognition method and system

Publications (2)

Publication Number Publication Date
CN106897700A true CN106897700A (en) 2017-06-27
CN106897700B CN106897700B (en) 2020-04-07

Family

ID=59184771

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710107890.XA Active CN106897700B (en) 2017-02-27 2017-02-27 Single-sample face recognition method and system

Country Status (1)

Country Link
CN (1) CN106897700B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107480628A (en) * 2017-08-10 2017-12-15 苏州大学 A kind of face identification method and device
CN107886090A (en) * 2017-12-15 2018-04-06 苏州大学 A kind of single sample face recognition method, system, equipment and readable storage medium storing program for executing
CN107958236A (en) * 2017-12-28 2018-04-24 深圳市金立通信设备有限公司 The generation method and terminal of recognition of face sample image
CN108520215A (en) * 2018-03-28 2018-09-11 电子科技大学 Single sample face recognition method based on multiple dimensioned union feature encoder
CN109635626A (en) * 2018-10-18 2019-04-16 西安理工大学 A kind of low resolution list class face identification method of list sample

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080166026A1 (en) * 2007-01-10 2008-07-10 Samsung Electronics Co., Ltd. Method and apparatus for generating face descriptor using extended local binary patterns, and method and apparatus for face recognition using extended local binary patterns
CN102855468A (en) * 2012-07-31 2013-01-02 东南大学 Single sample face recognition method in photo recognition
CN103714326A (en) * 2013-12-26 2014-04-09 江南大学 One-sample face identification method
US20140212044A1 (en) * 2012-09-26 2014-07-31 Carnegie Mellon University Image Matching Using Subspace-Based Discrete Transform Encoded Local Binary Patterns
CN105224957A (en) * 2015-10-23 2016-01-06 苏州大学 A kind of method and system of the image recognition based on single sample
CN106056074A (en) * 2016-05-27 2016-10-26 广东顺德中山大学卡内基梅隆大学国际联合研究院 Single training sample face identification method based on area sparse

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080166026A1 (en) * 2007-01-10 2008-07-10 Samsung Electronics Co., Ltd. Method and apparatus for generating face descriptor using extended local binary patterns, and method and apparatus for face recognition using extended local binary patterns
CN102855468A (en) * 2012-07-31 2013-01-02 东南大学 Single sample face recognition method in photo recognition
US20140212044A1 (en) * 2012-09-26 2014-07-31 Carnegie Mellon University Image Matching Using Subspace-Based Discrete Transform Encoded Local Binary Patterns
CN103714326A (en) * 2013-12-26 2014-04-09 江南大学 One-sample face identification method
CN105224957A (en) * 2015-10-23 2016-01-06 苏州大学 A kind of method and system of the image recognition based on single sample
CN106056074A (en) * 2016-05-27 2016-10-26 广东顺德中山大学卡内基梅隆大学国际联合研究院 Single training sample face identification method based on area sparse

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HAIBIN YAN 等: "Multi-feature multi-manifold learning for single-sample face recognition", 《NEUROCOMPUTING》 *
TINGWEI PEI 等: "Decision pyramid classifier for face recognition under complex variations using single sample per person", 《PATTERN RECOGNITION》 *
赵汝哲: "基于单样本训练集的人脸识别研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107480628A (en) * 2017-08-10 2017-12-15 苏州大学 A kind of face identification method and device
CN107480628B (en) * 2017-08-10 2020-08-25 苏州大学 Face recognition method and device
CN107886090A (en) * 2017-12-15 2018-04-06 苏州大学 A kind of single sample face recognition method, system, equipment and readable storage medium storing program for executing
CN107886090B (en) * 2017-12-15 2021-07-30 苏州大学 Single-sample face recognition method, system, equipment and readable storage medium
CN107958236A (en) * 2017-12-28 2018-04-24 深圳市金立通信设备有限公司 The generation method and terminal of recognition of face sample image
CN108520215A (en) * 2018-03-28 2018-09-11 电子科技大学 Single sample face recognition method based on multiple dimensioned union feature encoder
CN108520215B (en) * 2018-03-28 2022-10-11 电子科技大学 Single-sample face recognition method based on multi-scale joint feature encoder
CN109635626A (en) * 2018-10-18 2019-04-16 西安理工大学 A kind of low resolution list class face identification method of list sample
CN109635626B (en) * 2018-10-18 2022-11-25 北京和鸿盈科技术有限公司 Single-sample low-resolution single-class face recognition method

Also Published As

Publication number Publication date
CN106897700B (en) 2020-04-07

Similar Documents

Publication Publication Date Title
CN106897700A (en) A kind of single sample face recognition method and system
CN106503669B (en) Training and recognition method and system based on multitask deep learning network
CN107766850A (en) Based on the face identification method for combining face character information
CN104239858B (en) A kind of method and apparatus of face characteristic checking
CN109214360A (en) A kind of construction method of the human face recognition model based on ParaSoftMax loss function and application
CN107392082A (en) A kind of small area fingerprint comparison method based on deep learning
CN103824054A (en) Cascaded depth neural network-based face attribute recognition method
JP2022532177A (en) Forged face recognition methods, devices, and non-temporary computer-readable storage media
CN106022317A (en) Face identification method and apparatus
CN106250858A (en) A kind of recognition methods merging multiple face recognition algorithms and system
CN107871100A (en) The training method and device of faceform, face authentication method and device
CN106096535A (en) A kind of face verification method based on bilinearity associating CNN
CN106682602A (en) Driver behavior identification method and terminal
CN104866829A (en) Cross-age face verify method based on characteristic learning
CN108764041A (en) The face identification method of facial image is blocked for lower part
CN102902980B (en) A kind of biometric image analysis based on linear programming model and recognition methods
CN108573243A (en) A kind of comparison method of the low quality face based on depth convolutional neural networks
CN105678231A (en) Pedestrian image detection method based on sparse coding and neural network
CN107909104A (en) The face cluster method, apparatus and storage medium of a kind of picture
CN106384087A (en) Identity identification method based on multi-layer network human being features
CN111401292A (en) Face recognition network construction method fusing infrared image training
CN111339930A (en) Face recognition method combining mask attribute loss function
CN101493887A (en) Eyebrow image segmentation method based on semi-supervision learning and Hash index
CN114612941B (en) Palm vein feature-based multi-mode identity authentication method, device and system
WO2021204086A1 (en) Identity authentication method, and method and device for training identity authentication model

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