CN105117712A - Single-sample human face recognition method compatible for human face aging recognition - Google Patents
Single-sample human face recognition method compatible for human face aging recognition Download PDFInfo
- Publication number
- CN105117712A CN105117712A CN201510586207.6A CN201510586207A CN105117712A CN 105117712 A CN105117712 A CN 105117712A CN 201510586207 A CN201510586207 A CN 201510586207A CN 105117712 A CN105117712 A CN 105117712A
- Authority
- CN
- China
- Prior art keywords
- image
- group
- model
- face
- image model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—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)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a single-sample human face recognition method compatible for human face aging recognition, which comprises the steps of conducting the aging simulation on the pre-stored image model of a human face sample to re-construct the image model of the human face sample; conducting the global feature matching for a to-be-recognized human face image model with the image model of the human face sample, wherein if the matching fails, regarding the recognition result as mismatching; and conducting the local feature matching for the to-be-recognized human face image model with the image model of the human face sample, wherein if the matching fails, regarding the recognition result as mismatching. The above to-be-recognized human face image model is an active appearance model of a to-be-recognized human face image. The image model of the human face sample is an active appearance model of a reserved human face sample image. According to the technical scheme of the invention, the recognition effect compatible for human face aging influence is realized and improved based on the combination of the AAM technique with the IBSDT technique. Meanwhile, based on the combination of the AAM technique with the triangulation matching technique, the matching reliability of global features is greatly improved. Based on the combination of the LBP technique with the SURF technique, the matching reliability of local features and the illumination robustness are improved. Finally, the high recognition rate for the reserved human face image as a single sample is realized.
Description
Technical field
The present invention relates to technical field of face recognition, particularly relate to a kind of single sample face recognition method of compatible face aging identification.
Background technology
Face image processing is the important subject of computer vision field and image processing field with identification, receives the concern of many researchists always.In the fields such as criminal investigation, medical treatment, amusement, information, space, the process of portrait and conversion have a wide range of applications demand.
The age of people is a time dependent long process, and the external presentation of face can be made to produce obvious change.The problem such as noise, distortion in same person contemporaneity facial image not only will be solved, the identification problem of the facial image absorbed under also will solving same person different times (span time may be decades-long), varying environment in face identification system.This academicly be all challenge in the design of application system.
Different according to research emphasis and technological means, the development course of face recognition algorithms roughly experienced by three phases:
The first, before nineteen ninety: the recognition of face of this one-phase, based on geometric properties, mainly comprises: based on the recognition of face of face geometry feature, based on the recognition of face etc. of facial contours architectural feature.But because do not consider gray scale texture information, therefore promotion and application are bad.
, there are famous FERET face recognition algorithms test and some commercial face identification systems in second, 1991-1997: this one-phase; The two-dimension human face linear subspaces that mainly concentrate on of face recognition algorithms are analyzed and the aspect such as statistical-simulation spectrometry, such as: Eigenface method (PCA) and Fisherface method (LDA) remain the conventional main flow algorithm of field of face identification up to now, but are mainly used in multisample recognition of face; Elastic graph matching (EGM) describes face with attributed graph, can be used for single sample recognition of face, but due to a large amount of data and complicated deformation ratio pair, causes time, space complexity high, its application is deteriorated; Active shape model ASM and active apparent model AAM is Corpus--based Method describing method, can be used for recognition of face, ASM only carries out statistical modeling to sample shape, and AAM not only establishes shape Statistics model, set up overall texture variations model simultaneously, be widely used in the fields such as target detection, identification, attitude correction.
3rd, 1998 so far, increasing researchist and fund input are to the research of field of face identification, commercial face identification system obtains tremendous development, and the problems such as the illumination produced a very large impact practical application, attitude, expression, age and single specimen discerning become research emphasis, wherein, single specimen discerning refers to, in database, a people only has a reserved sample for training and can identify, such as, in public security database, identity card picture is reserved sample and only had one.
Current, commercial face identification system is in the application of reality, and many recognizers are under different restrictive conditions, and discrimination can be a greater impact, and is difficult to reach desirable recognition effect.Difficult point mainly concentrates on the key areas such as illumination, attitude, expression, age and single specimen discerning:
The first, face profile has instability, can produce various different expression, and, observe in different angles, easily produce different visual imaging effects, challenge is brought for the stability of recognition of face effect and accuracy.
The second, along with the increase at age, can there is aging change in facial characteristics.
3rd, man face image acquiring process is by the impact of the many factors such as illumination, covering (decoration, glasses, hair, cosmetic etc.) and low-quality image.
4th, for same algorithm, when number of training reduces, recognition correct rate also may decline to a great extent, and therefore single specimen discerning problem hard is especially outstanding.Such as, it is 8 ~ 10 that the best of Eigenface and the Fisherface method of main flow reserves picture number, and its recognition correct rate declines gradually along with the minimizing of training sample number.
It is worth noting, for commercial face identification system, in a lot of situation, multiple above-mentioned negative effect factor may occur simultaneously, affects the recognition correct rate of whole system.
Summary of the invention
Provide hereinafter about brief overview of the present invention, to provide about the basic comprehension in some of the present invention.Should be appreciated that this general introduction is not summarize about exhaustive of the present invention.It is not that intention determines key of the present invention or pith, and nor is it intended to limit the scope of the present invention.Its object is only provide some concept in simplified form, in this, as the preorder in greater detail discussed after a while.
The invention provides a kind of by compatible face aging identification to improve single sample face recognition method of the compatible face aging identification of recognition accuracy, and the reliability of recognition result is improved further by AAM models coupling Delaunay triangle division coupling, improved the illumination robustness of recognition result by LBP and SURF determination methods, and finally realize the high discrimination of single sample.
The invention provides a kind of single sample face recognition method of compatible face aging identification, comprising:
Aging simulation is carried out to the face sample image model prestored, reconstructs described face sample image model;
Carry out global characteristics coupling to facial image model to be identified and described face sample image model, if it fails to match, then recognition result is not for mate;
Carry out local feature coupling to described facial image model to be identified and described face sample image model, if it fails to match, then recognition result is not for mate;
Described facial image model to be identified is the active apparent model of facial image to be identified, and described face sample image model is the active apparent model of reserved face sample image.
Single sample face recognition method of the compatible face aging identification that the many embodiments of the present invention provide, by carrying out Aging simulation to the face sample image model prestored, achieves the identification of compatible face aging impact; And then by improve the accuracy rate of face aging identification in conjunction with active apparent model (AAM) and face grain details mapping algorithm (IBSDT);
Single sample face recognition method of the compatible face aging identification that some embodiments of the invention provide is by mating in conjunction with active apparent model (AAM) and Delaunay triangle division, using leg-of-mutton for Delaunay matching number as face global characteristics coupling mark, significantly improve the reliability of global characteristics matching result;
Single sample face recognition method of the compatible face aging identification that some embodiments of the invention provide is mated face local critical area by LBP and SURF two kinds of determination methods, because LBP and SURF possesses stronger illumination robustness separately, the recognition result of therefore recognition methods provided by the present invention also has very strong robustness to illumination;
Single sample face recognition method of the compatible face aging identification that some embodiments of the invention provide achieves when only there being 1 reserved image by the combination of above-mentioned every technology, still can obtain good recognition effect, recognition result is significantly better than Eigenface and the Fisherface method of current main-stream.
Accompanying drawing explanation
Below with reference to the accompanying drawings illustrate embodiments of the invention, above and other objects, features and advantages of the present invention can be understood more easily.Parts in accompanying drawing are just in order to illustrate principle of the present invention.In the accompanying drawings, same or similar technical characteristic or parts will adopt same or similar Reference numeral to represent.
The process flow diagram of single sample face recognition method of the compatible face aging identification that Fig. 1 provides for an embodiment of the present invention.
Fig. 2 be embodiment illustrated in fig. 1 on a facial image, set up active apparent model extract the schematic diagram of feature.
Fig. 3 is the process flow diagram of step S50 in sample face recognition method single shown in Fig. 1.
Fig. 4 is the process flow diagram of step S70 in sample face recognition method single shown in Fig. 1.
Fig. 5 carries out based on the unique point of active apparent model the schematic diagram that triangle division obtains the triangulation network for using Delaunay triangle division method in step S71 shown in Fig. 4.
The process flow diagram of step S90 in a kind of preferred embodiment that Fig. 6 is sample face recognition method single shown in Fig. 1.
Fig. 7 is for dividing the schematic diagram of critical area according to unique point in step S91 shown in Fig. 6.
Fig. 8 is the process flow diagram of step S93 in sample face recognition method single shown in Fig. 6.
Fig. 9 is the process flow diagram of step S95 in sample face recognition method single shown in Fig. 6.
Figure 10 is the process flow diagram of the preferred embodiment of sample face recognition method single shown in Fig. 1.
Figure 11 is the process flow diagram of the preferred embodiment of sample face recognition method single shown in Figure 10.
Figure 12 is the process flow diagram of face sample image model acquisition method in step S30 shown in Figure 11.
Figure 13 is the process flow diagram of the preferred embodiment of sample face recognition method single shown in Figure 11.
Figure 14 is the process flow diagram of step S20 in sample face recognition method single shown in Figure 13.
Embodiment
With reference to the accompanying drawings embodiments of the invention are described.The element described in an accompanying drawing of the present invention or a kind of embodiment and feature can combine with the element shown in one or more other accompanying drawing or embodiment and feature.It should be noted that for purposes of clarity, accompanying drawing and eliminate expression and the description of unrelated to the invention, parts known to persons of ordinary skill in the art and process in illustrating.
The process flow diagram of single sample face recognition method of the compatible face aging identification that Fig. 1 provides for an embodiment of the present invention.
As shown in Figure 1, in the present embodiment, single sample face recognition method of the compatible face aging identification of the present invention comprises:
S50: carry out Aging simulation to the face sample image model prestored, reconstructs described face sample image model.
S70: carry out global characteristics coupling to facial image model to be identified and described face sample image model, if it fails to match, then recognition result is not for mate.
S90: carry out local feature coupling to described facial image model to be identified and described face sample image model, if it fails to match, then recognition result is not for mate.
Wherein, described facial image model to be identified is the active apparent model of facial image to be identified, and described face sample image model is the active apparent model of reserved face sample image.
Fig. 2 be embodiment illustrated in fig. 1 on a facial image, set up active apparent model extract the schematic diagram of feature.As shown in Figure 2, active apparent model by marking the feature of feature point extraction face on facial image.Described facial image model to be identified and described face sample image model include the unique point of facial image to be identified and the unique point of face sample image respectively.
Single sample face recognition method of the compatible face aging identification that the above embodiment of the present invention provides, by carrying out Aging simulation to the face sample image model prestored, achieves the identification of compatible face aging impact.
Fig. 3 is the process flow diagram of step S50 in sample face recognition method single shown in Fig. 1.
As shown in Figure 3, in a preferred embodiment, in the single sample face recognition method shown in Fig. 1, step S50 comprises:
S51: extract the unique point in described face sample image model.
S53: according to described unique point, utilizes face grain details mapping algorithm (IBSDT) to reconstruct described face sample image model.
Particularly, face grain details mapping algorithm (IBSDT) is not changing on the basis of human face characteristic point, by conversion grain details, reduces the difference that face aging that the age brings causes.
The above-mentioned embodiment provided of the present invention is further by the accuracy rate that improve face aging identification in conjunction with active apparent model (AAM) and face grain details mapping algorithm (IBSDT).
Fig. 4 is the process flow diagram of step S70 in sample face recognition method single shown in Fig. 1.
As shown in Figure 4, in a preferred embodiment, in the single sample face recognition method shown in Fig. 1, step S70 comprises:
S71: carry out triangle division by the unique point of Delaunay triangle division method to described facial image model to be identified and described face sample image model respectively, obtains corresponding first triangulation network of described facial image model to be identified and second triangulation network of correspondence described face sample image model.
S73: whether each triangle in first triangulation network described in comparison and described second triangulation network mates, statistical match quantity.
S75: judge whether described number of matches is greater than predetermined threshold value, if described number of matches is greater than described predetermined threshold value, the result of described global characteristics coupling is that the match is successful, otherwise then result is that it fails to match.
Fig. 5 carries out based on the unique point of active apparent model the schematic diagram that triangle division obtains the triangulation network for using Delaunay triangle division method in step S71 shown in Fig. 4.As shown in Figure 5, Delaunay triangle division method distinguished point based defines the triangulation network of triangle composition.
In a preferred embodiment, step S73 specifically comprises:
Extract each triangle Δ in described first triangulation network respectively
lthree limit features, form each triangle Δ
lproper vector
wherein L=1,2 ..., m, m are the number of triangles of described first triangulation network,
be respectively Δ
lthe three limit length of sides;
Extract each triangle Δ in described second triangulation network respectively
dthree limit features, form each triangle Δ
dproper vector
wherein D=1,2 ..., n, n are the number of triangles of described second triangulation network,
be respectively Δ
dthe three limit length of sides;
The triangle Δ of first triangulation network described in comparison one by one
lwith the triangle Δ of described second triangulation network
dwhether meet:
Wherein, i=0,1,2, T
Δfor the matching threshold obtained according to statistics;
If meet, then Δ
lwith Δ
dcoupling; If do not meet, then Δ
lwith Δ
ddo not mate.
Single sample face recognition method of the compatible face aging identification that the above embodiment of the present invention provides is by mating in conjunction with active apparent model (AAM) and Delaunay triangle division, using leg-of-mutton for Delaunay matching number as face global characteristics coupling mark, significantly improve the reliability of global characteristics matching result.
In a preferred embodiment, step S90 comprises:
S91: be first group of critical area image according to respective unique point by described facial image model partition to be identified respectively, be second group of critical area image by described face sample image model partition.
S93: Corresponding matching is carried out to described first group of critical area image and described second group of critical area image according to LBP algorithm; And/or,
S95: Corresponding matching is carried out to described first group of critical area image and described second group of critical area image according to SURF algorithm.
Namely step S90 may comprise step S91+S93, step S91+S95, step S91+S93+S95 tri-kinds of schemes.
The process flow diagram of step S90 in a kind of preferred embodiment that Fig. 6 is sample face recognition method single shown in Fig. 1.In preferred embodiment shown in Fig. 6, step S90 comprises above-mentioned steps S91+S93+S95.
Particularly, the LBP algorithm that step S93 uses and the SURF algorithm that step S95 uses possess stronger illumination robustness separately, and in conjunction with two kinds of algorithms carry out successively local feature coupling achieve the better illumination robustness of recognition effect.
Fig. 7 is for dividing the schematic diagram of critical area according to unique point in step S91 shown in Fig. 6.
As shown in Figure 7, in a preferred embodiment, described first group of critical area image and described second group of critical area image comprise left eye region image, right eye region image and mouth region image respectively.
Fig. 8 is the process flow diagram of step S93 in sample face recognition method single shown in Fig. 6.
As shown in Figure 8, in a preferred embodiment, step S93 comprises:
S931: calculate corresponding first group of LBP image of described first group of critical area image and second group of LBP image of the described second group of critical area image of correspondence respectively according to LBP algorithm;
S933: the histogram calculating described first group of LBP image and described second group of LBP image respectively;
S935: whether the histogram of first group of LBP image described in comparison and the histogram of described second group of LBP image mate one by one.
Particularly, in the present embodiment, step S931 calculates the first left eye region LBP image of corresponding described facial image model to be identified, the first right eye region LBP image and the first mouth region LBP image respectively according to LBP algorithm, and the second left eye region LBP image of corresponding described face sample image model, the second right eye region LBP image and the second mouth region LBP image;
Step S933 calculates described first left eye region LBP image, the first right eye region LBP image and the first mouth region LBP image histogram separately respectively, and described second left eye region LBP image, the second right eye region LBP image and the second mouth region LBP image histogram separately;
Step S935 compares with regard to the first left eye region LBP image histogram and the second left eye region LBP image histogram, the first right eye region LBP image histogram and the second right eye region LBP image histogram, the first mouth region LBP image histogram and the second mouth region LBP image histogram respectively, obtains three groups of comparison results.If wherein arbitrary group of comparison result be not for mate, then local feature matching result is not for mate; If three groups of comparison results are all coupling, then local feature matching result is coupling.
In a preferred embodiment, step S935 comprises:
The coupling mark of each critical area image in described first group of LBP image is gone out respectively according to histogram calculation
with the coupling mark of each critical area image in described second group of LBP image
wherein i=0,1,2;
The each region of comparison
with
whether all satisfied:
Wherein, i=0,1,2,
for the LBP matching threshold of each critical area;
If all satisfied, then described first group of critical area image and described second group of critical area image are based on LBP algorithmic match; Otherwise then do not mate.
Particularly, the LBP matching threshold of each critical area
obtain according to statistics respectively.
Fig. 9 is the process flow diagram of step S95 in sample face recognition method single shown in Fig. 6.
As shown in Figure 9, in a preferred embodiment, step S95 comprises:
S951: the second group of SURF feature extracting the corresponding described second group of critical area image of first group of SURF characteristic sum of corresponding described first group of critical area image according to SURF algorithm respectively.
S953: described in comparison, described in first group of SURF characteristic sum, whether second group of SURF feature mates one by one.
Particularly, in the present embodiment, step S951 extracts the first left eye region SURF feature, the first right eye region SURF characteristic sum first mouth region SURF feature of corresponding described facial image model to be identified respectively according to SURF algorithm, and the second left eye region SURF feature of corresponding described face sample image model, the second right eye region SURF characteristic sum second mouth region SURF feature;
Step S953 compares with regard to the first left eye region SURF characteristic sum second left eye region SURF feature, the first right eye region SURF characteristic sum second right eye region SURF feature, the first mouth region SURF characteristic sum second mouth region SURF feature respectively, obtains three groups of comparison results.If wherein arbitrary group of comparison result be not for mate, then local feature matching result is not for mate; If three groups of comparison results are all coupling, then local feature matching result is coupling.
In a preferred embodiment, step S953 specifically comprises:
Calculate the coupling mark of each critical area image in described first group of SURF feature respectively
with the coupling mark of each critical area image in described first group of SURF feature
wherein i=0,1,2;
The each region of comparison
with
whether all satisfied:
Wherein, i=0,1,2,
for the SURF matching threshold of each critical area;
If all satisfied, then described first group of critical area image and described second group of critical area image are based on SURF algorithmic match; Otherwise then do not mate.
Figure 10 is the process flow diagram of the preferred embodiment of sample face recognition method single shown in Fig. 1.
As shown in Figure 10, in a preferred embodiment, also comprise before step S50:
S40: the age gap distance according to facial image model to be identified and face sample image model judges whether to carry out Aging simulation to described face sample image model.
Particularly, in the present embodiment, when the age gap of facial image model to be identified and face sample image model is apart from when being less than predetermined threshold value, no longer performs step S50, and directly jump to step S70.
In a preferred embodiment, when the age gap of facial image model to be identified and face sample image model is apart from when being not less than predetermined threshold value, described age gap apart from as parameter, can be realized more accurate face aging effect by the IBSDT algorithm in step S53.
Figure 11 is the process flow diagram of the preferred embodiment of sample face recognition method single shown in Figure 10.
As shown in figure 11, in a preferred embodiment, also comprise before step S40:
S30: extract the face sample image model stored in a database.
Figure 12 is the process flow diagram of face sample image model acquisition method in step S30 shown in Figure 11.
As shown in figure 12, in a preferred embodiment, the face sample image model that step S30 extracts from database gathers in advance to be formed and stored in database, and its generation method comprises:
S11: gather reserved facial image;
S13: mark unique point on described facial image;
S15: the active apparent model setting up described facial image according to described unique point, obtains face sample image model;
S17: by described face sample image model storage in database.
Figure 13 is the process flow diagram of the preferred embodiment of sample face recognition method single shown in Figure 11.
As shown in figure 13, in a preferred embodiment, also comprise before step S30:
S20: gather facial image to be identified and generate facial image model to be identified.
Figure 14 is the process flow diagram of step S20 in sample face recognition method single shown in Figure 13.
As shown in figure 14, in a preferred embodiment, step S20 specifically comprises:
S21: gather facial image to be identified;
S23: mark unique point on described facial image to be identified;
S25: the active apparent model setting up described facial image to be identified according to described unique point, obtains facial image model to be identified.
Single sample face recognition method of the compatible face aging identification that the more above-mentioned embodiments of the present invention provide is mated face local critical area by LBP and SURF two kinds of determination methods, because LBP and SURF possesses stronger illumination robustness separately, the recognition result of therefore recognition methods provided by the present invention also has very strong robustness to illumination.
Single sample face recognition method of the compatible face aging identification that the more above-mentioned embodiments of the present invention provide achieves when only there being 1 reserved image by the combination of above-mentioned every technology, still can obtain good recognition effect, recognition result is significantly better than Eigenface and the Fisherface method of current main-stream.
Last it is noted that above embodiment is only in order to illustrate technical scheme of the present invention, be not intended to limit; Although with reference to previous embodiment to invention has been detailed description, those of ordinary skill in the art is to be understood that: it still can be modified to the technical scheme described in foregoing embodiments, or carries out equivalent replacement to wherein portion of techniques feature; And these amendments or replacement, do not make the essence of appropriate technical solution depart from the spirit and scope of various embodiments of the present invention technical scheme.
Claims (15)
1. a single sample face recognition method for compatible face aging identification, is characterized in that, comprising:
Aging simulation is carried out to the face sample image model prestored, reconstructs described face sample image model;
Carry out global characteristics coupling to facial image model to be identified and described face sample image model, if it fails to match, then recognition result is not for mate;
Carry out local feature coupling to described facial image model to be identified and described face sample image model, if it fails to match, then recognition result is not for mate;
Described facial image model to be identified is the active apparent model of facial image to be identified, and described face sample image model is the active apparent model of reserved face sample image.
2. single sample face recognition method according to claim 1, is characterized in that, the described face sample image model to prestoring carries out Aging simulation, reconstructs described face sample image model and comprises:
Extract the unique point in described face sample image model;
According to described unique point, face grain details mapping algorithm (IBSDT) is utilized to reconstruct described face sample image model.
3. single sample face recognition method according to claim 1, is characterized in that, described to facial image model to be identified and described face sample image model carry out global characteristics coupling comprise:
Carry out triangle division by the unique point of Delaunay triangle division method to described facial image model to be identified and described face sample image model respectively, obtain corresponding first triangulation network of described facial image model to be identified and second triangulation network of correspondence described face sample image model;
Whether each triangle in first triangulation network described in comparison and described second triangulation network mates, statistical match quantity;
Judge whether described number of matches is greater than predetermined threshold value, if described number of matches is greater than described predetermined threshold value, the result of described global characteristics coupling is that the match is successful, otherwise then result is that it fails to match.
4. single sample face recognition method according to claim 3, is characterized in that, whether each triangle in first triangulation network described in described comparison and described second triangulation network mates comprises:
Extract each triangle Δ in described first triangulation network respectively
lthree limit features, form each triangle Δ
lproper vector
wherein L=1,2 ..., m, m are the number of triangles of described first triangulation network,
be respectively Δ
lthe three limit length of sides;
Extract each triangle Δ in described second triangulation network respectively
dthree limit features, form each triangle Δ
dproper vector
wherein D=1,2 ..., n, n are the number of triangles of described second triangulation network,
be respectively Δ
dthe three limit length of sides;
The triangle Δ of first triangulation network described in comparison one by one
lwith the triangle Δ of described second triangulation network
dwhether meet:
Wherein, i=0,1,2, T
Δfor the matching threshold obtained according to statistics;
If meet, then Δ
lwith Δ
dcoupling; If do not meet, then Δ
lwith Δ
ddo not mate.
5. single sample face recognition method according to claim 1, is characterized in that, described to described facial image model to be identified and described face sample image model carry out local feature coupling comprise:
Be first group of critical area image according to respective unique point by described facial image model partition to be identified respectively, be second group of critical area image by described face sample image model partition;
According to LBP algorithm, Corresponding matching is carried out to described first group of critical area image and described second group of critical area image; And/or,
According to SURF algorithm, Corresponding matching is carried out to described first group of critical area image and described second group of critical area image.
6. single sample face recognition method according to claim 5, is characterized in that, described first group of critical area image and described second group of critical area image comprise left eye region image, right eye region image and mouth region image respectively.
7. single sample face recognition method according to claim 6, is characterized in that, described use LBP algorithm carries out Corresponding matching to described first group of critical area image and described second group of critical area image and comprises:
Corresponding first group of LBP image of described first group of critical area image and second group of LBP image of the described second group of critical area image of correspondence is calculated respectively according to LBP algorithm;
Calculate the histogram of described first group of LBP image and described second group of LBP image respectively;
Whether the histogram of first group of LBP image described in comparison and the histogram of described second group of LBP image mate one by one.
8. single sample face recognition method according to claim 7, is characterized in that, one by one whether the histogram of first group of LBP image described in described comparison and the histogram of described second group of LBP image coupling comprise:
The coupling mark of each critical area image in described first group of LBP image is gone out respectively according to histogram calculation
with the coupling mark of each critical area image in described second group of LBP image
wherein i=0,1,2;
The each region of comparison
with
whether all satisfied:
Wherein, i=0,1,2,
for the LBP matching threshold of each critical area;
If all satisfied, then described first group of critical area image and described second group of critical area image are based on LBP algorithmic match; Otherwise then do not mate.
9. single sample face recognition method according to claim 6, is characterized in that, describedly carries out Corresponding matching according to SURF algorithm to described first group of critical area image and described second group of critical area image and comprises:
Second group of SURF feature of the corresponding described second group of critical area image of first group of SURF characteristic sum of corresponding described first group of critical area image is extracted respectively according to SURF algorithm;
Described in comparison, described in first group of SURF characteristic sum, whether second group of SURF feature mates one by one.
10. single sample face recognition method according to claim 9, is characterized in that, whether second group of SURF feature described in first group of SURF characteristic sum described in described comparison mates one by one comprises:
Calculate the coupling mark of each critical area image in described first group of SURF feature respectively
with the coupling mark of each critical area image in described first group of SURF feature
wherein i=0,1,2;
The each region of comparison
with
whether all satisfied:
Wherein, i=0,1,2,
for the SURF matching threshold of each critical area;
If all satisfied, then described first group of critical area image and described second group of critical area image are based on SURF algorithmic match; Otherwise then do not mate.
11. single sample face recognition method according to claim 1, is characterized in that, the described face sample image model to prestoring carries out Aging simulation, also comprises before reconstructing described face sample image model:
Age gap distance according to facial image model to be identified and face sample image model judges whether to carry out Aging simulation to described face sample image model.
12. single sample face recognition method according to claim 1, is characterized in that, the described face sample image model to prestoring carries out Aging simulation, also comprises before reconstructing described face sample image model:
Extract the face sample image model stored in a database.
13. single sample face recognition method according to claim 12, is characterized in that, the generation method of described face sample image model comprises:
Gather reserved facial image;
Described facial image marks unique point;
Set up the active apparent model of described facial image according to described unique point, obtain face sample image model;
By described face sample image model storage in database.
14. single sample face recognition method according to claim 1, is characterized in that, the described face sample image model to prestoring carries out Aging simulation, also comprises before reconstructing described face sample image model:
Gather facial image to be identified and generate facial image model to be identified.
15. single sample face recognition method according to claim 14, is characterized in that, described collection facial image to be identified also generates facial image model to be identified and comprises:
Gather facial image to be identified;
Described facial image to be identified marks unique point;
Set up the active apparent model of described facial image to be identified according to described unique point, obtain facial image model to be identified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510586207.6A CN105117712A (en) | 2015-09-15 | 2015-09-15 | Single-sample human face recognition method compatible for human face aging recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510586207.6A CN105117712A (en) | 2015-09-15 | 2015-09-15 | Single-sample human face recognition method compatible for human face aging recognition |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105117712A true CN105117712A (en) | 2015-12-02 |
Family
ID=54665695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510586207.6A Pending CN105117712A (en) | 2015-09-15 | 2015-09-15 | Single-sample human face recognition method compatible for human face aging recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105117712A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570822A (en) * | 2016-10-25 | 2017-04-19 | 宇龙计算机通信科技(深圳)有限公司 | Human face mapping method and device |
| CN106599872A (en) * | 2016-12-23 | 2017-04-26 | 北京旷视科技有限公司 | Method and equipment for verifying living face images |
| CN107145829A (en) * | 2017-04-07 | 2017-09-08 | 电子科技大学 | A kind of vena metacarpea recognition methods for merging textural characteristics and scale invariant feature |
| CN108664925A (en) * | 2018-05-10 | 2018-10-16 | 王逸人 | The method and user terminal of recognition of face |
| CN108681752A (en) * | 2018-05-28 | 2018-10-19 | 电子科技大学 | A kind of image scene mask method based on deep learning |
| CN109584327A (en) * | 2018-11-08 | 2019-04-05 | 深圳市威富视界有限公司 | Face aging analogy method, device and equipment |
| CN109685029A (en) * | 2018-12-28 | 2019-04-26 | 东莞中国科学院云计算产业技术创新与育成中心 | Flexible article recognition methods, device, equipment and the medium of Complicated Spatial Structure |
| CN110119746A (en) * | 2019-05-08 | 2019-08-13 | 北京市商汤科技开发有限公司 | A kind of characteristic recognition method and device, computer readable storage medium |
| CN114821720A (en) * | 2022-04-25 | 2022-07-29 | 广州瀚信通信科技股份有限公司 | Face detection method, device, system, equipment and storage medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20050039765A (en) * | 2005-02-21 | 2005-04-29 | 주식회사 아이티펌 | Method for extracting face outline and face feature points |
| CN103927511A (en) * | 2014-02-25 | 2014-07-16 | 华北电力大学（保定） | Image identification method based on difference feature description |
-
2015
- 2015-09-15 CN CN201510586207.6A patent/CN105117712A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20050039765A (en) * | 2005-02-21 | 2005-04-29 | 주식회사 아이티펌 | Method for extracting face outline and face feature points |
| CN103927511A (en) * | 2014-02-25 | 2014-07-16 | 华北电力大学（保定） | Image identification method based on difference feature description |
Non-Patent Citations (3)
| Title |
|---|
| Y. SHAN等: "Image-Based Surface Detail Transfer", 《COMPUTER VISION AND PATTERN RECOGNITION》 * |
| 叶永青: "基于稀疏约束非负矩阵分解算法的人脸图像老化模拟及其在人脸识别中的应用", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 苏煜: "融合全局和局部特征的人脸识别", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570822A (en) * | 2016-10-25 | 2017-04-19 | 宇龙计算机通信科技(深圳)有限公司 | Human face mapping method and device |
| CN106570822B (en) * | 2016-10-25 | 2020-10-16 | 宇龙计算机通信科技(深圳)有限公司 | Face mapping method and device |
| CN106599872A (en) * | 2016-12-23 | 2017-04-26 | 北京旷视科技有限公司 | Method and equipment for verifying living face images |
| CN107145829B (en) * | 2017-04-07 | 2020-05-22 | 电子科技大学 | Palm vein identification method integrating textural features and scale invariant features |
| CN107145829A (en) * | 2017-04-07 | 2017-09-08 | 电子科技大学 | A kind of vena metacarpea recognition methods for merging textural characteristics and scale invariant feature |
| CN108664925A (en) * | 2018-05-10 | 2018-10-16 | 王逸人 | The method and user terminal of recognition of face |
| CN108681752A (en) * | 2018-05-28 | 2018-10-19 | 电子科技大学 | A kind of image scene mask method based on deep learning |
| CN108681752B (en) * | 2018-05-28 | 2023-08-15 | 电子科技大学 | Image scene labeling method based on deep learning |
| CN109584327A (en) * | 2018-11-08 | 2019-04-05 | 深圳市威富视界有限公司 | Face aging analogy method, device and equipment |
| CN109584327B (en) * | 2018-11-08 | 2023-07-25 | 深圳市威富视界有限公司 | Face aging simulation method, device and equipment |
| CN109685029A (en) * | 2018-12-28 | 2019-04-26 | 东莞中国科学院云计算产业技术创新与育成中心 | Flexible article recognition methods, device, equipment and the medium of Complicated Spatial Structure |
| CN110119746A (en) * | 2019-05-08 | 2019-08-13 | 北京市商汤科技开发有限公司 | A kind of characteristic recognition method and device, computer readable storage medium |
| CN110119746B (en) * | 2019-05-08 | 2021-11-30 | 北京市商汤科技开发有限公司 | Feature recognition method and device and computer readable storage medium |
| CN114821720A (en) * | 2022-04-25 | 2022-07-29 | 广州瀚信通信科技股份有限公司 | Face detection method, device, system, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105117712A (en) | Single-sample human face recognition method compatible for human face aging recognition | |
| CN109657631B (en) | Human body posture recognition method and device | |
| CN108334816B (en) | Multi-pose face recognition method based on contour symmetric constraint generation type countermeasure network | |
| CN109685013B (en) | Method and device for detecting head key points in human body posture recognition | |
| CN110490158B (en) | Robust face alignment method based on multistage model | |
| CN107145842A (en) | With reference to LBP characteristic patterns and the face identification method of convolutional neural networks | |
| Nalepa et al. | Wrist localization in color images for hand gesture recognition | |
| CN103514442A (en) | Video sequence face identification method based on AAM model | |
| CN110796101A (en) | Face recognition method and system of embedded platform | |
| Loconsole et al. | Real-time emotion recognition: an improved hybrid approach for classification performance | |
| Crabbe et al. | Skeleton-free body pose estimation from depth images for movement analysis | |
| KR101326691B1 (en) | Robust face recognition method through statistical learning of local features | |
| CN103902992A (en) | Human face recognition method | |
| CN102495999A (en) | Face recognition method | |
| Papatheodorou et al. | Evaluation of 3D face recognition using registration and PCA | |
| Jang et al. | Robust deep age estimation method using artificially generated image set | |
| Teijeiro-Mosquera et al. | Performance of active appearance model-based pose-robust face recognition | |
| Zafer et al. | Face recognition with expression variation via robust ncc | |
| Bourbakis et al. | Skin-based face detection-extraction and recognition of facial expressions | |
| Kuo et al. | Face recognition based on a two-view projective transformation using one sample per subject | |
| Li et al. | Face recognition by 3d registration for the visually impaired using a rgb-d sensor | |
| KR101501410B1 (en) | Method for detecting a position of eye | |
| Panchal et al. | 3d face recognition based on pose correction using Euler angle method | |
| Lei et al. | A novel shape-based interest point descriptor (SIP) for 3D ear recognition | |
| Mawale et al. | Facial wrinkles detection techniques and its application |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20151202 |
|
| RJ01 | Rejection of invention patent application after publication |