CN106250858B - Recognition method and system fusing multiple face recognition algorithms - Google Patents
Recognition method and system fusing multiple face recognition algorithms Download PDFInfo
- Publication number
- CN106250858B CN106250858B CN201610634586.6A CN201610634586A CN106250858B CN 106250858 B CN106250858 B CN 106250858B CN 201610634586 A CN201610634586 A CN 201610634586A CN 106250858 B CN106250858 B CN 106250858B
- Authority
- CN
- China
- Prior art keywords
- face
- face recognition
- face image
- recognition algorithm
- recognition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 20
- 239000013598 vector Substances 0.000 claims abstract description 118
- 238000000605 extraction Methods 0.000 claims abstract description 9
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000004927 fusion Effects 0.000 abstract description 42
- 238000007499 fusion processing Methods 0.000 abstract description 21
- 238000005516 engineering process Methods 0.000 description 4
- 241000282414 Homo sapiens Species 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000000513 principal component analysis Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Images
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a recognition method and a system fusing multiple face recognition algorithms, comprising the following steps: presetting at least two face recognition algorithms; performing feature extraction and recognition on the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain a feature vector of the corresponding face image to be recognized, a feature vector of the sample face image, a contrast similarity value and a recognition result; performing fusion processing on the extracted feature vector of the face image to be recognized and the extracted feature vector of the sample face image, and determining a final recognition result according to the fusion feature vector of the face image to be recognized and the fusion feature vector of the sample face image; or, carrying out fusion processing on the obtained contrast similarity values, and determining a final recognition result according to the fused contrast similarity values; or, carrying out decision processing on the recognition result, and determining a final recognition result according to the processing result. Therefore, the embodiment of the invention can improve the accuracy of face recognition.
Description
Technical Field
The invention relates to the technical field of face recognition, in particular to a recognition method and a recognition system fusing multiple face recognition algorithms.
Background
The face recognition technology is a technology for performing identity authentication by using physiological characteristics or behavior characteristics which are owned by human beings and can uniquely mark the identity of the human beings based on a biological characteristic recognition mode. With the development of artificial intelligence technology and computer mode recognition, face recognition technology has been developed from laboratories to practical applications.
At present, the existing face recognition system is realized by optimizing a face recognition algorithm, and the accuracy of face recognition are limited to a certain extent. Therefore, in order to solve the problem that the accuracy and accuracy of face recognition are limited by using a single face recognition algorithm, a method for performing face recognition by fusing multiple face recognition algorithms is urgently needed at present so as to improve the accuracy of face recognition.
Disclosure of Invention
In view of the above disadvantages of the prior art, an object of the present invention is to provide an identification method and system fusing multiple face identification algorithms, which can improve the accuracy of face identification by fusing multiple face identification algorithms for identification and authentication.
In order to achieve the above and other related objects, an embodiment of the present invention provides a recognition method fusing multiple face recognition algorithms, where at least two face recognition algorithms are preset, and the method further includes:
extracting the features of the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain the feature vector of the face image to be recognized and the feature vector of the sample face image corresponding to each face recognition algorithm;
respectively calculating the contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image;
respectively determining the recognition result of each face recognition algorithm according to the contrast similarity value of each face recognition algorithm;
performing fusion processing on the feature vectors of the face images to be recognized extracted by each face recognition algorithm, performing fusion processing on the feature vectors of the sample face images extracted by each face recognition algorithm, and determining a final recognition result according to the fusion feature vectors of the face images to be recognized and the fusion feature vectors of the sample face images; or,
fusing the contrast similarity values of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity values; or,
and performing decision processing on the recognition result of each face recognition algorithm, and determining a final recognition result according to the processing result.
Preferably, the fusing the feature vectors of the face images to be recognized extracted by each face recognition algorithm and the feature vectors of the sample face images extracted by each face recognition algorithm, and determining the final recognition result according to the fused feature vectors of the face images to be recognized and the fused feature vectors of the sample face images, includes:
the feature vectors of the face image to be recognized extracted by each face recognition algorithm are subjected to fusion processing to obtain fusion feature vectors f of the face image to be recognizedAAnd fusing the feature vectors f through a principal component analysis algorithm PCA pairAPerforming dimensionality reduction to obtain new fusion characteristicsVector FA：
The characteristic vectors of the sample face images extracted by each face recognition algorithm are fused to obtain the fusion characteristic vector f of the sample face imagesBAnd fusing the feature vectors f by PCA pairsBDimension reduction processing is carried out to obtain a new fusion characteristic vector FB：
Calculating a new fusion feature vector F according to a probability linear discriminant analysis algorithm (PLDA)AWith the new fused feature vector FBObtaining a fused contrast similarity value;
and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
Preferably, the fusing the contrast similarity values of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity values, includes:
presetting a weight coefficient of each face recognition algorithm;
obtaining a fused contrast similarity value according to the contrast similarity value and the weight coefficient of each face recognition algorithm:
and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
Preferably, the determining the recognition result of each face recognition algorithm according to the recognition result of each face recognition algorithm includes:
determining the number of successfully identified identification results and the number of all identification results;
judging whether the number of the successfully identified identification results is greater than one half of the number of all the identification results;
if the number of the successfully identified identification results is greater than one half of the number of all the identification results, determining that the final identification result is successfully identified;
and if the number of the successfully-identified identification results is less than or equal to one half of the number of all the identification results, determining that the final identification result is identification failure.
Preferably, the determining the recognition result of each face recognition algorithm according to the recognition result of each face recognition algorithm includes:
identifying each type of face image according to each face identification algorithm in advance to obtain the identification rate of each type of face image by each face identification algorithm;
determining the object types of the face image to be recognized and the sample face image;
selecting a face recognition algorithm with the highest recognition rate corresponding to the object type according to the determined object type and the recognition rate of each type of face recognition algorithm to each type of face image obtained in advance;
and determining the recognition result obtained by the selected face recognition algorithm as a final recognition result.
According to the above method, an embodiment of the present invention provides an identification system fusing multiple face recognition algorithms, including: the device comprises a presetting module, a feature extraction module, a calculation module, a first determination module and a second determination module; wherein,
the preset module is used for presetting at least two face recognition algorithms;
the feature extraction module is used for extracting features of the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain a feature vector of the face image to be recognized and a feature vector of the sample face image corresponding to each face recognition algorithm;
the computing module is used for respectively computing the contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image;
the first determining module is used for respectively determining the recognition result of each face recognition algorithm according to the contrast similarity value of each face recognition algorithm;
the second determining module is used for performing fusion processing on the feature vectors of the face images to be recognized extracted by each face recognition algorithm and the feature vectors of the sample face images extracted by each face recognition algorithm, and determining a final recognition result according to the fusion feature vectors of the face images to be recognized and the fusion feature vectors of the sample face images; or, fusing the contrast similarity values of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity values; or, the recognition result of each face recognition algorithm is subjected to decision processing, and the final recognition result is determined according to the processing result.
Preferably, the second determining module is specifically configured to:
the feature vectors of the face image to be recognized extracted by each face recognition algorithm are subjected to fusion processing to obtain fusion feature vectors f of the face image to be recognizedAAnd fusing the feature vectors f by PCA pairsADimension reduction processing is carried out to obtain a new fusion characteristic vector FA：
The characteristic vectors of the sample face images extracted by each face recognition algorithm are fused to obtain the fusion characteristic vector f of the sample face imagesBAnd fusing the feature vectors f by PCA pairsBDimension reduction processing is carried out to obtain a new fusion characteristic vector FB：
Calculating a new fusion eigenvector F according to the PLDAAWith the new fused feature vector FBObtaining a fused contrast similarity value;
and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
Preferably, the second determining module is specifically configured to:
presetting a weight coefficient of each face recognition algorithm;
obtaining a fused contrast similarity value according to the contrast similarity value and the weight coefficient of each face recognition algorithm:
and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
Preferably, the second determining module is specifically configured to:
determining the number of successfully identified identification results and the number of all identification results;
judging whether the number of the successfully identified identification results is greater than one half of the number of all the identification results;
if the number of the successfully identified identification results is greater than one half of the number of all the identification results, determining that the final identification result is successfully identified;
and if the number of the successfully-identified identification results is less than or equal to one half of the number of all the identification results, determining that the final identification result is identification failure.
Preferably, the second determining module is specifically configured to:
identifying each type of face image according to each face identification algorithm in advance to obtain the identification rate of each type of face image by each face identification algorithm;
determining the object types of the face image to be recognized and the sample face image;
selecting a face recognition algorithm with the highest recognition rate corresponding to the object type according to the determined object type and the recognition rate of each type of face recognition algorithm to each type of face image obtained in advance;
and determining the recognition result obtained by the selected face recognition algorithm as a final recognition result.
The invention provides an identification method and system fusing a plurality of face identification algorithms, comprising the following steps: presetting at least two face recognition algorithms; extracting the features of the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain the feature vector of the face image to be recognized and the feature vector of the sample face image corresponding to each face recognition algorithm; respectively calculating the contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image; respectively determining the recognition result of each face recognition algorithm according to the contrast similarity value of each face recognition algorithm; performing fusion processing on the feature vectors of the face images to be recognized extracted by each face recognition algorithm, performing fusion processing on the feature vectors of the sample face images extracted by each face recognition algorithm, and determining a final recognition result according to the fusion feature vectors of the face images to be recognized and the fusion feature vectors of the sample face images; or, carrying out decision processing on the contrast similarity value of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity value; or, the recognition result of each face recognition algorithm is subjected to decision processing, and the final recognition result is determined according to the processing result. Therefore, in the embodiment of the invention, a plurality of face recognition algorithms are fused for face recognition, decision fusion is carried out from a characteristic vector end, a comparison similarity value end or a recognition result end, and a final recognition result is determined according to the decision fusion result, so that the problem that the accuracy and the accuracy of face recognition are limited by using a single face recognition algorithm is solved, and the accuracy of face recognition is improved.
Drawings
FIG. 1 is a schematic flow chart of an identification method of the present invention incorporating multiple face recognition algorithms;
fig. 2 is a schematic diagram showing a composition structure of the recognition system of the present invention which integrates a plurality of face recognition algorithms.
Detailed Description
In the embodiment of the invention, in order to solve the problem that the accuracy and the accuracy of face recognition are limited by using a single face recognition algorithm, a plurality of face recognition algorithms are adopted for face recognition, decision fusion is carried out on the plurality of face recognition algorithms in different layers, the decision fusion can be carried out from a characteristic vector end, a contrast similarity value end or a recognition result end, and a final recognition result is determined according to the decision fusion result, so that the accuracy of face recognition is improved.
The invention is described in further detail below with reference to the figures and the embodiments.
The embodiment of the invention provides an identification method fusing multiple face identification algorithms, as shown in figure 1, the method comprises the following steps:
step S100: at least two face recognition algorithms are preset.
In this step, at least two arbitrary face recognition algorithms may be set, and the preset face recognition algorithm is not specifically limited herein.
Step S101: and performing feature extraction on the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain a feature vector of the face image to be recognized and a feature vector of the sample face image corresponding to each face recognition algorithm.
In this step, the length of the feature vector extracted according to different face recognition algorithms may be any length, but the extracted feature vector must satisfy the following condition:
Norm(F)＝1
wherein F is the characteristic vector, and norm (F) is the module length of the characteristic vector F.
Step S102: and respectively calculating the contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image.
In this step, the contrast similarity value of each face recognition algorithm may be normalized to a [0, 1] interval to obtain a uniform contrast similarity value for subsequent classification and recognition.
Step S103: and respectively determining the recognition result of each face recognition algorithm according to the contrast similarity value of each face recognition algorithm.
In the step, the contrast similarity value obtained by each face recognition algorithm is compared with a preset contrast similarity threshold value;
if the contrast similarity value of the face recognition algorithm is larger than a preset contrast similarity threshold value, determining the recognition result of the face recognition algorithm as successful recognition;
and if the contrast similarity value of the face recognition algorithm is smaller than or equal to a preset contrast similarity threshold, determining the recognition result of the face recognition algorithm as recognition failure.
Here, the contrast similarity threshold may be preset according to actual situations and requirements, and is not specifically limited herein.
Step S104: performing fusion processing on the feature vectors of the face images to be recognized extracted by each face recognition algorithm, performing fusion processing on the feature vectors of the sample face images extracted by each face recognition algorithm, and determining a final recognition result according to the fusion feature vectors of the face images to be recognized and the fusion feature vectors of the sample face images; or, fusing the contrast similarity values of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity values; or, the recognition result of each face recognition algorithm is subjected to decision processing, and the final recognition result is determined according to the processing result.
In this step, how to specifically perform fusion processing on the feature vectors of the face images to be recognized extracted by each face recognition algorithm and the feature vectors of the sample face images extracted by each face recognition algorithm is specifically performed, and a final recognition result is determined according to the fusion feature vectors of the face images to be recognized and the fusion feature vectors of the sample face images, which is described in detail:
firstly, the feature vectors of the face image to be recognized extracted by each face recognition algorithm are fused to obtain the fusion feature vector f of the face image to be recognizedA：
fA＝[f1；f2；…；fn]Wherein f isAFor fusing feature vectors, f1～fnExtracting feature vectors of the face image to be recognized for each face recognition algorithm;
fusion feature vector f is subjected to Principal Component Analysis (PCA)ADimension reduction processing is carried out to obtain a new fusion characteristic vector FA：
similarly, the feature vectors of the sample face images extracted by each face recognition algorithm are subjected to fusion processing to obtain fusion feature vectors f of the sample face imagesB：
fB＝[f1；f2；…；fn]Wherein f isBFor fusing feature vectors, f1～fnExtracting a feature vector of a sample face image for each face recognition algorithm;
fusing feature vectors f by PCA pairsBDimension reduction processing is carried out to obtain a new fusion characteristic vector FB：
then, a new fusion feature vector F is calculated according to the Probabilistic Linear Discriminant Analysis (PLDA)AWith the new fused feature vector FBTo obtain a fused contrast similarity value s (ab):
wherein P is the intra-class variance of the PLDA, and Q is the inter-class variance of the PLDA;
finally, determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold;
specifically, the fused contrast similarity value is compared with a preset contrast similarity threshold value;
if the fused contrast similarity value is larger than a preset contrast similarity threshold value, determining that the final recognition result is successful in recognition;
and if the fused contrast similarity value is smaller than or equal to a preset contrast similarity threshold, determining that the final recognition result is recognition failure.
Here, the contrast similarity threshold may be preset according to actual situations and requirements, and is not specifically limited herein.
In this step, how to specifically perform fusion processing on the contrast similarity value of each face recognition algorithm is described in detail, and a final recognition result is determined according to the fused contrast similarity value:
firstly, presetting a weight coefficient of each face recognition algorithm;
then, according to the contrast similarity value and the weight coefficient of each face recognition algorithm, obtaining a fused contrast similarity value:
wherein S (A, B) is the fused contrast similarity value, omegaiIs the weight coefficient of the ith face recognition algorithm, FiThe contrast similarity value of the ith face recognition algorithm is 1, 2, …, n, n is a positive integer;
finally, determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold;
specifically, the fused contrast similarity value is compared with a preset contrast similarity threshold value;
if the fused contrast similarity value is larger than a preset contrast similarity threshold value, determining that the final recognition result is successful in recognition;
and if the fused contrast similarity value is smaller than or equal to a preset contrast similarity threshold, determining that the final recognition result is recognition failure.
Here, the contrast similarity threshold may be preset according to actual situations and requirements, and is not specifically limited herein.
In this step, how to specifically perform decision processing on the recognition result of each face recognition algorithm is determined, and a final recognition result is determined according to the processing result, which is described in detail as follows:
mode one, decision strategy based on voting
Firstly, determining the number R of successfully-identified identification results and the number G of all identification results;
then, judging whether the number R of the successfully recognized recognition results is greater than the number G of all half recognition results;
if the number R of successfully recognized recognition results is larger than the number G of all half recognition results, determining that the final recognition result is successfully recognized;
and if the number R of successfully recognized recognition results is less than or equal to one half of the number G of all recognition results, determining that the final recognition result is a recognition failure.
Each face recognition algorithm obtains a recognition result, the number G is equal to the number of the face recognition algorithms, and the number R is smaller than or equal to the number G.
Mode two, decision strategy based on prior information
Considering that each face recognition algorithm has a recognition object type which is good in itself, for example, some face recognition algorithms have a good recognition rate for a child face image, but have a poor recognition rate for a certificate photo. Therefore, each type of face image can be identified in advance according to each type of face identification algorithm, and an identification rate table of each type of face image by each type of face identification algorithm can be obtained; when a group of images needs to be recognized, the characteristics of the group of face images are obtained first, for example, the face images of children are determined according to the characteristics of the face images, or the face images of minority nationalities are determined according to the characteristics of the face images, then the determined object types and the pre-obtained recognition rates are used as prior information, and a face recognition algorithm with the best recognition rate is selected according to the prior information.
Specifically, each type of face image is identified in advance according to each type of face identification algorithm, and the identification rate of each type of face image by each type of face identification algorithm is obtained, as shown in table 1:
| face recognition algorithm 1 | Face recognition algorithm | …… | Face recognition algorithm N | |
| Class A | 0.980 | 0.965 | …… | 0.910 |
| Class B | 0.923 | 0.995 | …… | 0.945 |
| Class C | 0.921 | 0.936 | …… | 0.988 |
| …… | …… | …… | …… | …… |
TABLE 1
Determining the object types of the face image to be recognized and the sample face image;
selecting a face recognition algorithm with the highest recognition rate corresponding to the object type according to the pre-obtained recognition rate of each face recognition algorithm to each type of face image;
and determining the recognition result obtained by the selected face recognition algorithm as a final recognition result.
Specifically, the face image is firstly divided into a plurality of object types, for example, a type a: acquiring a face image and authenticating a certificate photo on site; b type: child authentication; class C: the minority is authenticated; then determining the object type according to the characteristics of the face image to be recognized and the sample face image; selecting a face recognition algorithm with the highest recognition rate corresponding to the object type according to the pre-obtained recognition rate of each face recognition algorithm to each type of face image; and if the face recognition algorithm with the highest recognition rate exists, determining the face recognition algorithm as a final recognition result, and if a plurality of face recognition algorithms with the highest recognition rates exist, determining the recognition result obtained by any one face recognition algorithm with the highest recognition rate as the final recognition result.
In order to implement the method, the embodiment of the present invention further provides an identification system fusing multiple face identification algorithms, and because the principle of solving the problem of the system is similar to the method, the implementation process and the implementation principle of the system can be described by referring to the implementation process and the implementation principle of the method, and repeated details are not repeated.
The embodiment of the invention provides an identification system fusing a plurality of face identification algorithms, as shown in fig. 2, the system comprises: the system comprises a presetting module 200, a feature extraction module 201, a calculation module 202, a first determination module 203 and a second determination module 204; wherein,
the presetting module 200 is used for presetting at least two face recognition algorithms;
the feature extraction module 201 is configured to perform feature extraction on the face image to be recognized and the sample face image through each face recognition algorithm, and obtain a feature vector of the face image to be recognized and a feature vector of the sample face image corresponding to each face recognition algorithm respectively;
the calculating module 202 is configured to calculate a contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image;
the first determining module 203 is configured to determine the recognition result of each face recognition algorithm according to the contrast similarity value of each face recognition algorithm;
the second determining module 204 is configured to perform fusion processing on the feature vectors of the to-be-recognized face images extracted by each face recognition algorithm, perform fusion processing on the feature vectors of the sample face images extracted by each face recognition algorithm, and determine a final recognition result according to the fusion feature vectors of the to-be-recognized face images and the fusion feature vectors of the sample face images; or, fusing the contrast similarity values of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity values; or, the recognition results of each face recognition algorithm are fused, and the final recognition result is determined according to the fused recognition results.
In a specific implementation, the second determining module 204 is specifically configured to:
the feature vectors of the face image to be recognized extracted by each face recognition algorithm are subjected to fusion processing to obtain fusion feature vectors f of the face image to be recognizedAAnd fusing the feature vectors f by PCA pairsADimension reduction processing is carried out to obtain a new fusion characteristic vector FA：
The characteristic vectors of the sample face images extracted by each face recognition algorithm are fused to obtain the fusion characteristic vector f of the sample face imagesBAnd fusing the feature vectors f by PCA pairsBDimension reduction processing is carried out to obtain a new fusion characteristic vector FB：
Calculating a new fusion eigenvector F according to the PLDAAWith the new fused feature vector FBObtaining a fused contrast similarity value;
and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
In a specific implementation, the second determining module 204 is specifically configured to:
presetting a weight coefficient of each face recognition algorithm;
obtaining a fused contrast similarity value according to the contrast similarity value and the weight coefficient of each face recognition algorithm:
and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
In a specific implementation, the second determining module 204 is specifically configured to:
determining the number of successfully identified identification results and the number of all identification results;
judging whether the number of the successfully identified identification results is greater than one half of the number of all the identification results;
if the number of the successfully identified identification results is greater than one half of the number of all the identification results, determining that the final identification result is successfully identified;
and if the number of the successfully-identified identification results is less than or equal to one half of the number of all the identification results, determining that the final identification result is identification failure.
In a specific implementation, the second determining module 204 is specifically configured to:
identifying each type of face image according to each face identification algorithm in advance to obtain the identification rate of each type of face image by each face identification algorithm;
determining the object types of the face image to be recognized and the sample face image;
selecting a face recognition algorithm with the highest recognition rate corresponding to the object type according to the determined object type and the recognition rate of each type of face recognition algorithm to each type of face image obtained in advance;
and determining the recognition result obtained by the selected face recognition algorithm as a final recognition result.
The above division manner of the functional modules is only one preferred implementation manner given in the embodiment of the present invention, and the division manner of the functional modules does not limit the present invention. For convenience of description, the parts of the system described above are separately described as functionally divided into various modules or units. Of course, the functionality of the various modules or units may be implemented in the same one or more pieces of software or hardware in practicing the invention.
In summary, in the embodiment of the present invention, at least two face recognition algorithms are preset first; extracting the features of the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain the feature vector of the face image to be recognized and the feature vector of the sample face image corresponding to each face recognition algorithm; respectively calculating the contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image; respectively determining the recognition result of each face recognition algorithm according to the contrast similarity value of each face recognition algorithm; performing fusion processing on the feature vectors of the face images to be recognized extracted by each face recognition algorithm, performing fusion processing on the feature vectors of the sample face images extracted by each face recognition algorithm, and determining a final recognition result according to the fusion feature vectors of the face images to be recognized and the fusion feature vectors of the sample face images; or, fusing the contrast similarity values of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity values; or, the recognition results of each face recognition algorithm are fused, and the final recognition result is determined according to the fused recognition results. Therefore, the embodiment of the invention integrates a plurality of face recognition algorithms to perform face recognition, respectively performs the fusion processing on the feature vector, the contrast similarity value or the recognition result, and determines the final recognition result according to the fusion processing result, thereby solving the problem that the accuracy and the accuracy of the face recognition are limited by using a single face recognition algorithm, and improving the accuracy of the face recognition.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.
Claims (2)
1. A recognition method and system fusing multiple face recognition algorithms are characterized in that at least two face recognition algorithms are preset, and the method further comprises the following steps:
extracting the features of the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain the feature vector of the face image to be recognized and the feature vector of the sample face image corresponding to each face recognition algorithm;
respectively calculating the contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image;
presetting a weight coefficient of each face recognition algorithm, obtaining a fused contrast similarity value according to the contrast similarity value and the weight coefficient of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
2. A recognition system that incorporates multiple face recognition algorithms, the system comprising: the device comprises a presetting module, a feature extraction module and a calculation module; wherein,
the preset module is used for presetting at least two face recognition algorithms;
the feature extraction module is used for extracting features of the face image to be recognized and the sample face image through each face recognition algorithm to respectively obtain a feature vector of the face image to be recognized and a feature vector of the sample face image corresponding to each face recognition algorithm;
the computing module is used for respectively computing the contrast similarity value of each face recognition algorithm according to the feature vector of the face image to be recognized extracted by each face recognition algorithm and the feature vector of the sample face image;
presetting a weight coefficient of each face recognition algorithm, obtaining a fused contrast similarity value according to the contrast similarity value and the weight coefficient of each face recognition algorithm, and determining a final recognition result according to the fused contrast similarity value and a preset contrast similarity threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610634586.6A CN106250858B (en) | 2016-08-05 | 2016-08-05 | Recognition method and system fusing multiple face recognition algorithms |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610634586.6A CN106250858B (en) | 2016-08-05 | 2016-08-05 | Recognition method and system fusing multiple face recognition algorithms |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106250858A CN106250858A (en) | 2016-12-21 |
| CN106250858B true CN106250858B (en) | 2021-08-13 |
Family
ID=58077677
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610634586.6A Active CN106250858B (en) | 2016-08-05 | 2016-08-05 | Recognition method and system fusing multiple face recognition algorithms |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106250858B (en) |
Families Citing this family (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106650693B (en) * | 2016-12-30 | 2020-07-03 | 河北三川科技有限公司 | Multi-feature fusion recognition algorithm for face comparison |
| CN107563280A (en) * | 2017-07-24 | 2018-01-09 | 南京道熵信息技术有限公司 | Face identification method and device based on multi-model |
| CN107729300B (en) * | 2017-09-18 | 2021-12-24 | 百度在线网络技术（北京）有限公司 | Text similarity processing method, device and equipment and computer storage medium |
| CN109918976B (en) * | 2017-12-13 | 2021-04-02 | 航天信息股份有限公司 | Portrait comparison algorithm fusion method and device thereof |
| CN108197532B (en) * | 2017-12-18 | 2019-08-16 | 深圳励飞科技有限公司 | The method, apparatus and computer installation of recognition of face |
| CN108491794B (en) * | 2018-03-22 | 2023-04-07 | 腾讯科技（深圳）有限公司 | Face recognition method and device |
| CN108776768A (en) * | 2018-04-19 | 2018-11-09 | 广州视源电子科技股份有限公司 | Image recognition method and device |
| CN108764350A (en) * | 2018-05-30 | 2018-11-06 | 苏州科达科技股份有限公司 | Target identification method, device and electronic equipment |
| CN109145758A (en) * | 2018-07-25 | 2019-01-04 | 武汉恩智电子科技有限公司 | A kind of recognizer of the face based on video monitoring |
| CN108921140A (en) * | 2018-08-07 | 2018-11-30 | 安徽云森物联网科技有限公司 | Pedestrian's recognition methods again |
| CN110908289A (en) * | 2018-09-17 | 2020-03-24 | 珠海格力电器股份有限公司 | Smart home control method and device |
| CN111177436B (en) * | 2018-11-09 | 2023-08-22 | 浙江宇视科技有限公司 | Face feature retrieval method, device and equipment |
| CN109377624B (en) * | 2018-11-23 | 2020-09-08 | 深圳市宝诚通科技发展有限公司 | Intelligent door opening system based on face image recognition |
| CN109624844A (en) * | 2018-12-05 | 2019-04-16 | 电子科技大学成都学院 | A kind of bus driving protection system based on image recognition and voice transmission control |
| CN110245573B (en) * | 2019-05-21 | 2023-05-26 | 平安科技（深圳）有限公司 | Sign-in method and device based on face recognition and terminal equipment |
| CN110309744B (en) * | 2019-06-21 | 2023-07-25 | 武汉市公安局视频侦查支队 | Method and device for identifying suspects |
| CN110532856A (en) * | 2019-07-16 | 2019-12-03 | 公安部第一研究所 | A kind of face identification method of more algorithm fusions |
| CN110334695A (en) * | 2019-07-18 | 2019-10-15 | 国网电子商务有限公司 | A kind of identity identifying method and device based on face characteristic |
| CN110688892B (en) * | 2019-08-20 | 2023-01-03 | 武汉烽火众智数字技术有限责任公司 | Portrait identification alarm method and system based on data fusion technology |
| CN110728522A (en) * | 2019-10-11 | 2020-01-24 | 上海可充电子商务有限公司 | Face brushing public traffic system |
| CN111209937B (en) * | 2019-12-27 | 2024-03-29 | 深圳智链物联科技有限公司 | Classification method and device for charging curve model and server |
| CN111178259A (en) * | 2019-12-30 | 2020-05-19 | 八维通科技有限公司 | Recognition method and system supporting multi-algorithm fusion |
| CN112308070B (en) * | 2020-10-30 | 2024-04-26 | 深圳前海微众银行股份有限公司 | Identification method and device for certificate information, equipment and computer readable storage medium |
| CN112597823A (en) * | 2020-12-07 | 2021-04-02 | 深延科技(北京)有限公司 | Attention recognition method and device, electronic equipment and storage medium |
| CN116434313B (en) * | 2023-04-28 | 2023-11-14 | 北京声迅电子股份有限公司 | Face recognition method based on multiple face recognition modules |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102411708A (en) * | 2011-12-02 | 2012-04-11 | 湖南大学 | Face recognition method combining dual-tree complex wavelet transform and discrete wavelet transform |
| CN102521561A (en) * | 2011-11-16 | 2012-06-27 | 湖南大学 | Face identification method on basis of multi-scale weber local features and hierarchical decision fusion |
| CN102831413A (en) * | 2012-09-11 | 2012-12-19 | 上海中原电子技术工程有限公司 | Face identification method and face identification system based on fusion of multiple classifiers |
| CN103136504A (en) * | 2011-11-28 | 2013-06-05 | 汉王科技股份有限公司 | Face recognition method and device |
| CN104899579A (en) * | 2015-06-29 | 2015-09-09 | 小米科技有限责任公司 | Face recognition method and face recognition device |
-
2016
- 2016-08-05 CN CN201610634586.6A patent/CN106250858B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521561A (en) * | 2011-11-16 | 2012-06-27 | 湖南大学 | Face identification method on basis of multi-scale weber local features and hierarchical decision fusion |
| CN103136504A (en) * | 2011-11-28 | 2013-06-05 | 汉王科技股份有限公司 | Face recognition method and device |
| CN102411708A (en) * | 2011-12-02 | 2012-04-11 | 湖南大学 | Face recognition method combining dual-tree complex wavelet transform and discrete wavelet transform |
| CN102831413A (en) * | 2012-09-11 | 2012-12-19 | 上海中原电子技术工程有限公司 | Face identification method and face identification system based on fusion of multiple classifiers |
| CN104899579A (en) * | 2015-06-29 | 2015-09-09 | 小米科技有限责任公司 | Face recognition method and face recognition device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106250858A (en) | 2016-12-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106250858B (en) | Recognition method and system fusing multiple face recognition algorithms | |
| Shekhar et al. | Joint sparse representation for robust multimodal biometrics recognition | |
| CN100356388C (en) | Biocharacteristics fusioned identity distinguishing and identification method | |
| US11062120B2 (en) | High speed reference point independent database filtering for fingerprint identification | |
| US9911027B2 (en) | Fingerprint authentication system, fingerprint authentication program and fingerprint authentication method | |
| JP2016157420A (en) | Image template masking | |
| Perumal et al. | A multimodal biometric system based on palmprint and finger knuckle print recognition methods. | |
| CN102891751A (en) | Method and equipment for generating business code from fingerprint image | |
| KR101854258B1 (en) | Memory optimization method for applying deep learning to pc | |
| Kant | A multimodal approach to improve the performance of biometric system | |
| CN109598235B (en) | Finger vein image authentication method and device | |
| CN112036316B (en) | Finger vein recognition method, device, electronic equipment and readable storage medium | |
| CN112836629A (en) | Image classification method | |
| Neha et al. | Biometric re-authentication: An approach towards achieving transparency in user authentication | |
| Booysens et al. | Exploration of ear biometrics using EfficientNet | |
| CN107292146B (en) | User feature vector selection method and system | |
| Triantafyllou et al. | Iris authentication utilizing co-occurrence matrices and textile features | |
| CN110851808A (en) | Identity authentication method and device, electronic equipment and computer readable storage medium | |
| Ayodele et al. | Current practices in information fusion for multimodal biometrics | |
| Shu et al. | Multi-biometrics fusion for identity verification | |
| CN108596112A (en) | A kind of choice of dynamical method of extensive iris feature identification matching threshold | |
| CN113626469B (en) | Internet of things equipment matching method and device | |
| TW201227541A (en) | Face feature recognition method | |
| CN111160306B (en) | Three-dimensional palmprint recognition method integrating multi-feature and principal component analysis | |
| CN111368665B (en) | High-anti-spoofing multi-mode identity authentication method based on competition and cooperation mechanism |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information | ||
| CB02 | Change of applicant information |
Address after: 401120 No. 1 and 2, floors 3-8, building 11, No. 2, Huizhu Road, Yubei District, Chongqing Applicant after: CHONGQING ZHONGKE YUNCONG TECHNOLOGY Co.,Ltd. Address before: 401122 Central Sixth Floor of Mercury Science and Technology Building B, Central Section of Huangshan Avenue, Northern New District of Chongqing Applicant before: CHONGQING ZHONGKE YUNCONG TECHNOLOGY Co.,Ltd. |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |