CN110852150A - Face verification method, system, equipment and computer readable storage medium - Google Patents
Face verification method, system, equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN110852150A CN110852150A CN201910913242.2A CN201910913242A CN110852150A CN 110852150 A CN110852150 A CN 110852150A CN 201910913242 A CN201910913242 A CN 201910913242A CN 110852150 A CN110852150 A CN 110852150A
- Authority
- CN
- China
- Prior art keywords
- face
- image
- information
- verified
- face feature
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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
Abstract
The application relates to a face verification method, a face verification system, face verification equipment and a computer-readable storage medium, wherein the method comprises the following steps: acquiring an image to be verified and identification information of the image, wherein the image comprises a human face; inputting the image and the identification information into a first depth learning network model trained in advance to obtain face region information in the image; performing image processing on at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area; inputting the face images into a pre-trained second deep learning network model to obtain face feature combination information of each face image; and acquiring a face verification result of the image to be verified according to the face feature combination information of each face image. The embodiment of the application adopts the improved deep learning network model, so that the face recognition rate and the face verification precision can be effectively improved.
Description
Technical Field
The present application relates to the field of face recognition technologies, and in particular, to a face verification method, a face verification system, a face verification device, and a computer-readable storage medium.
Background
The face recognition technology has wide application prospect, and has important application value in the fields of national security, military security, public security and the like, one of the cores of the face recognition is face verification, the traditional face verification method is to verify a detected face based on the face characteristics of a person, but along with the gradual evolution of an application scene of the face detection from indoor to outdoor, the application scene is developed from a single limited scene to various complex detection scenes such as squares, stations, subway windows and the like, the requirements of the face detection are higher and higher, and under the complex detection environment, the traditional face detection is not fully satisfactory in performance.
With the development of deep learning, the face detection technology based on deep learning has achieved great success, and in the traditional method based on deep learning, because pictures input by a user are easily affected by uncontrollable factors such as background environment, illumination change, face emotion and the like, discrimination of face verification is interfered, the face recognition rate is greatly reduced, and the face verification precision is low, so that the improvement of the face recognition rate and the face verification precision under complex conditions has great significance.
Disclosure of Invention
In order to solve the technical problems described above or at least partially solve the technical problems, the present application provides a face verification method, a system, a device and a computer-readable storage medium.
In view of the above, in a first aspect, the present application provides a face verification method, including the following steps:
acquiring an image to be verified and identification information of the image, wherein the image comprises a human face;
inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises a coordinate position of at least one face region in the image;
performing image processing on the at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area;
inputting a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area into a pre-trained second deep learning network model to obtain face feature combination information of each face image, wherein the face feature combination information is feature information obtained by combining face feature information corresponding to each face image;
and acquiring a face verification result of the image to be verified according to the face feature combination information of each face image.
With reference to the first aspect, in a first possible implementation manner of the first aspect, the performing image processing on the at least one face region according to a preset rule to obtain a plurality of to-be-verified face images with different face angles, where the to-be-verified face images correspond to each face region in the at least one face region includes:
determining a reference line of the image according to the coordinate position of at least one face area in the image;
and rotating the image along the reference line according to a preset rotation angle and/or preset rotation times to obtain a plurality of to-be-verified face images with different face angles, which correspond to each face area in the at least one face area.
With reference to the first aspect, in a second possible implementation manner of the first aspect, the inputting, into a second deep learning network model trained in advance, a plurality of to-be-verified face images with different face angles, which correspond to each of the at least one face region, to obtain face feature combination information of each of the face images includes:
acquiring a face feature vector of each face image;
combining the face feature vectors to obtain the face feature combined vector;
and taking the face feature combination vector as the face feature combination information.
With reference to the first aspect, in a third possible implementation manner of the first aspect, the combining the face feature vectors to obtain the face feature combination vector includes:
combining the face feature vectors according to a preset combination mode to obtain the face feature combination vectors, wherein the preset combination mode comprises at least one of the following combinations: affine combinations, linear combinations, and convex combinations.
With reference to the first aspect, in a fourth possible implementation manner of the first aspect, the obtaining, according to the face feature combination information of each of the face images, a face verification result of the image to be verified includes:
searching sample identification information corresponding to the identification information and sample face feature information corresponding to the sample identification information from a database;
matching the sample face feature information with each face feature combination information to obtain a matching result;
and acquiring a face verification result of the image to be verified according to the matching result.
With reference to the first aspect, in a fifth possible implementation manner of the first aspect, the matching the sample face feature information with each piece of face feature combination information to obtain a matching result includes:
respectively calculating difference values of the face feature combination information and the sample face feature information;
comparing each difference value with a set difference threshold value;
wherein, the obtaining the face verification result of the image to be verified according to the matching result comprises:
when any difference value is smaller than the difference threshold value, determining that sample face feature information matched with the face feature combination information exists in the preset database, and the face verification of the image is successful;
and when the difference values are all larger than or equal to the difference threshold value, determining that sample face feature information matched with the face feature combination information does not exist in the preset database, and failing to verify the face of the image.
With reference to the first aspect, in a sixth possible implementation manner of the first aspect, the separately calculating a difference value between each piece of the face feature combination information and the sample face feature information includes:
calculating Euclidean distance between each piece of face feature combination information and the sample face feature information;
and taking the value of the Euclidean distance as the difference value.
In a second aspect, the present application provides a face verification system, comprising:
the system comprises an acquisition unit, a verification unit and a verification unit, wherein the acquisition unit is used for acquiring an image to be verified and identification information of the image, and the image comprises a human face;
the input unit is used for inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises the coordinate position of at least one face region in the image;
the image processing unit is used for carrying out image processing on the at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area;
the input unit is further configured to input the multiple to-be-verified face images with different face angles, which correspond to each of the at least one face region, into a second deep learning network model trained in advance to obtain face feature combination information of each face image, where the face feature combination information is feature information obtained by combining face feature information corresponding to each face image;
the obtaining unit is further configured to obtain a face verification result of the image to be verified according to the face feature combination information of each face image.
In a third aspect, the present application provides a computer-readable storage medium having stored thereon a face verification program, which when executed by a processor implements the steps of the face verification method according to the first aspect.
In a fourth aspect, the present application provides an electronic device, comprising: a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory communicating via the bus when the electronic device is operating, the machine-readable instructions when executed by the processor performing the face verification method according to the first aspect.
The technical scheme provided by the embodiment of the application has the following advantages:
the embodiment of the application provides a face verification method, a system and a computer readable storage medium, wherein the face verification method comprises the steps of obtaining an image to be verified (the image comprises a face) and identification information of the image; inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises a coordinate position of at least one face region in the image; performing image processing on the at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area; inputting a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area into a pre-trained second deep learning network model to obtain face feature combination information of each face image, wherein the face feature combination information is feature information obtained by combining face feature information corresponding to each face image; and acquiring a face verification result of the image to be verified according to the face feature combination information of each face image.
According to the method and the device for verifying the face of the image, the image to be verified and the identification information of the image are input into the two improved deep learning network models (the first deep learning network model and the second deep learning network model), the face verification result of the image to be verified is obtained, the influence and interference of the image to be verified by the traditional verification method due to uncontrollable factors such as background environment, illumination change, face inclination angle and face emotion can be effectively avoided, and the face recognition rate and the face verification precision of the image to be verified are further improved.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without inventive exercise.
Fig. 1 is a schematic flow chart of a face verification method according to an embodiment of the present application;
fig. 2 is a schematic flow chart of a face verification method according to another embodiment of the present application;
fig. 3 is a schematic flowchart of a face verification method according to another embodiment of the present application;
fig. 4 is a schematic flowchart of a face verification method according to another embodiment of the present application;
fig. 5 is a schematic structural diagram of a face verification system according to an embodiment of the present application;
fig. 6 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
A server implementing various embodiments of the present invention will now be described with reference to the accompanying drawings. In the following description, suffixes such as "module", "component", or "unit" used to denote elements are used only for facilitating the explanation of the present invention, and have no specific meaning in themselves. Thus, "module" and "component" may be used in a mixture.
Because the traditional face verification method based on deep learning, the image to be verified input by a user is easily influenced by uncontrollable factors such as background environment, illumination change, face emotion, face angle and the like, and the face recognition rate and the face verification precision are greatly reduced, so that the improvement of the face recognition rate and the face verification precision under complex conditions has important significance, the embodiment of the application provides a face verification method, as shown in fig. 1, the method can comprise the following steps:
s101, obtaining an image to be verified and identification information of the image, wherein the image comprises a human face.
Optionally, the identification information of the image to be verified includes an ID of the image (ID of the user), and the image to be verified includes one or more faces.
S102, inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises a coordinate position of at least one face region in the image.
According to the embodiment of the application, the first deep learning network model is used for extracting the face region in the image to obtain the face region information, for example, the face region information comprises the rectangular frame coordinates, the eyes, the nose and the coordinate positions of two mouth corners occupied by the face region.
The embodiment of the application further comprises: firstly, uploading an image to be verified and identification information (such as an ID of the image) of the image to a face recognition server, wherein the face recognition server is provided with a pre-trained deep learning network model, firstly, checking sample identification information (such as a sample ID) in a database, judging whether the ID of the image is occupied, if so, feeding back an error code 1 to identify that the ID of the image is registered, and storing a record of registration failure in a table REGISTER _ FAIL; if the ID of the image is found to be failed (namely the database reports errors), an error code 4 is fed back to identify the database error reporting and the registration failure record is stored in a REGISTER _ FAIL; and if the ID of the image is not occupied, entering a face detection step of a first deep learning network model to detect the face in the image.
The first deep learning network model can detect whether a face exists in the image or not and several faces exist in the image, and can extract face region information. If the image does not have the face, feeding back an error code 2, and storing the registration failure record into a list REGISTER _ FAIL; if more than two faces are detected in the image, feeding back an error code 3, and storing the registration failure record into a list REGISTER _ FAIL; and if the detection result is that a human face is detected, entering a subsequent image processing step, wherein the human face region information is known.
The embodiment of the application further comprises the following steps after the face region information is extracted from the image:
based on the face region information, judging whether the coordinate position of each face region in the image meets a preset condition, wherein the preset condition comprises that: whether the coordinate position of the face area is within a set coordinate position range, for example, whether the coordinate positions of the two eyes are on the same horizontal line; if any one of the coordinate positions does not accord with a preset condition, adjusting the coordinate position until the coordinate position accords with the preset condition; and if all the coordinate positions meet preset conditions, directly outputting the face region information.
S103, performing image processing on the at least one face area according to a preset rule to obtain a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area.
For example, an image to be verified and an ID of the image are input to a face recognition server, a plurality of face images to be verified with different face angles are output through a first deep learning network model trained in the face recognition server in advance, and the face image is input to a second deep learning network model trained in advance to obtain face feature combination information of each face image, wherein the face feature combination information is feature information obtained by combining face feature information corresponding to each face image.
The following describes, by using a specific embodiment, the acquisition of a plurality of face images at different face angles and the acquisition of face feature combination information of each face image, which are not described herein again.
And S104, inputting the face images into a pre-trained second deep learning network model to obtain the face feature combination information of each face image.
And S105, acquiring a face verification result of the image to be verified according to the face feature combination information of each face image.
According to the method and the device for verifying the face of the image, the image to be verified and the identification information of the image are input into the two improved deep learning network models (the first deep learning network model and the second deep learning network model), the face verification result of the image to be verified is obtained, the influence and interference of the image to be verified by the traditional verification method due to uncontrollable factors such as background environment, illumination change, face inclination angle and face emotion can be effectively avoided, and the face recognition rate and the face verification precision of the image to be verified are further improved.
Referring to fig. 2, an embodiment of the present application further provides a face verification method, where in step S103, image processing is performed on the at least one face area according to a preset rule, so as to obtain a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area, and the method includes the following steps:
s201, determining a reference line of the image according to the coordinate position of at least one face area in the image.
S202, rotating the image along the datum line according to a preset rotation angle and/or preset rotation times to obtain a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area.
For example, the coordinate position 1 of the middle point of the two eyes, the coordinate position 2 of the nose and the coordinate position 3 of the mouth can be obtained based on the face region information, the coordinate position 1, the coordinate position 2 and the coordinate position 3 are connected into a line, taking the line as a reference line, rotating the image along the reference line by pixels, wherein the rotation angle is 1 degree, the rotation times are 29 times, the image is rotated by 29 times, each time the image is rotated by 1 degree, 30 face images with deviation angles of plus or minus 15 degrees are obtained, the 30 human face images have different human face angles, the influence of non-controllable factors such as human face emotion and human face angles on human face recognition and human face verification is overcome, pixels of the images are rotated for n times by using a specific rotation mode to obtain n +1 human face images (including original images), the multi-angle characteristics of human faces can be mined, and therefore the human face recognition rate is improved.
Referring to fig. 3, an embodiment of the present application further provides a face verification method, where in step S104, inputting the face image into a second deep learning network model trained in advance to obtain face feature combination information of each face image, includes:
s301, obtaining a face feature vector of each face image.
The human face image is a pixel point set of a region in the image, and a human face feature vector obtained after the human face image is input into the human face feature extraction model is a digital feature vector, namely, the inside of the human face feature vector is a number which is different from a pixel.
For example, 30 face images obtained by pixel rotation are input into the second deep learning network model, and face feature vectors of the 30 face images are output, wherein the face feature vectors are 512-dimensional digital feature vectors.
And S302, combining the face feature vectors to obtain the face feature combined vector.
Optionally, the face feature vectors are combined according to a preset combination mode to obtain the face feature combination vector, where the preset combination mode includes at least one of the following combinations: affine combinations, linear combinations, and convex combinations.
Inputting the obtained face FEATURE vectors of 30 face images into a face FEATURE combination model, outputting the combination of the 30 face FEATURE vectors to obtain a 512-dimensional face FEATURE combination vector, storing the face FEATURE combination vector and the ID of the image into a table FEATURE _ LIBRARY, and feeding back code information 0 to indicate that the face FEATURE vectors are successfully obtained.
For example, 30 face feature vectors of 512 dimensions are combined in the form of affine combination, wherein face feature vectors with coefficients of [0.032, …,0.032, 0.038,0.06,0.038,0.032, … 0.032.032 ] corresponding to [ -15 °, … -2 °, 1 °,0 °, 1 °, 2 °, … 15 ° ] are used. The embodiment of the present application is only used for explaining a combination manner of face feature vectors, and is not used for limiting the same.
Face verification: when the face recognition server receives an image to be verified and the ID of the image, checking a database, if the ID of the image is not registered, feeding back an error code 3, and storing the record into a table VALIDATE _ FAIL; if the database is in error in the searching process, feeding back an error code 4, and storing the record into a table VALIDATE _ FAIL; if the ID of the image is registered, detecting whether a face exists in the image, if the face is not detected, feeding back an error code 2, and storing the record into a table VALIDATE _ FAIL; if the human face is detected (only the human face exists in the image during verification), the process enters a human face feature vector obtaining process, and human face feature combination vectors of all the human faces in the image are obtained.
And comparing the face feature combination vector and the ID of the image to be verified with the sample face feature vector in the database, and calculating the difference value between the face feature combination vector and the sample face feature vector to obtain a comparison result. For example, by calculating the euclidean distance between the face feature combination vector and the sample face feature vector, when the calculated euclidean distance is smaller than a set euclidean distance threshold (for example, the euclidean distance is set to 0.67), it is determined that the faces corresponding to the two feature vectors are the same face, the face verification of the image is successful, the code information 0 is fed back, and the record is stored in the table valid _ SUCCESS; otherwise, when the calculated euclidean distance is greater than or equal to the set euclidean distance threshold, the faces corresponding to the two feature vectors are determined to be different faces, that is, no face corresponding to the ID of the image exists in the faces in the image, the code information 1 is fed back, and the record is stored in the table valid _ FAIL, and the face verification of the image FAILs.
Optionally, when the face recognition server receives the image to be verified and the ID of the image, there may be multiple faces in the image, that is, multiple face images, and as long as one of the multiple face images matches with a sample face image corresponding to the ID of the image in the preset database, the image is considered to be verified to be passed, that is, the face verification is successful.
And S303, taking the face feature combination vector as the face feature combination information.
In this embodiment, a specific rotation mode is used to perform pixel rotation on an initial face image n times to obtain n +1 face images (including an original image), a multi-angle face feature vector combination mode is used to pass the obtained n +1 face images through a face feature combination model to obtain n +1 face feature vectors representing the face, and then the combination of the n +1 face feature vectors is used to obtain a face feature combination vector representing the face finally.
Referring to fig. 4, an embodiment of the present application further provides a face verification method, where in step S105, a face verification result of the image to be verified is obtained according to the face feature combination information of each face image, and the method includes the following steps:
s401, searching sample identification information corresponding to the identification information and sample face feature information corresponding to the sample identification information from a database.
Optionally, the database stores sample identification information of a plurality of sample images and sample facial feature information corresponding to each sample identification information.
S402, matching the sample face feature information with the face feature combination information to obtain a matching result.
And S403, acquiring a face verification result of the image to be verified according to the matching result.
Optionally, a matching result of the face feature combination information and the sample face feature information is obtained by respectively calculating a difference value between the face feature combination information and the sample face feature information, and a face verification result of the image to be verified is obtained according to the matching result of the face feature combination information and the sample face feature information; comparing each difference value with a set difference threshold value; when any difference value is smaller than the difference threshold value, determining that sample face feature information matched with the face feature combination information exists in the preset database, and the face verification of the image is successful; and when the difference values are all larger than or equal to the difference threshold value, determining that sample face feature information matched with the face feature combination information does not exist in the preset database, and failing to verify the face of the image.
Optionally, calculating difference values between each of the facial feature combination information and the sample facial feature information respectively includes:
calculating Euclidean distance between each piece of face feature combination information and the sample face feature information;
and taking the value of the Euclidean distance as the difference value.
The cosine similarity between each piece of face feature combination information and the sample face feature information may also be calculated, and the value of the cosine similarity is used as the difference value, which is not limited in the embodiment of the present application.
Referring to fig. 5, an embodiment of the present application further provides a face verification system, where the system includes:
the device comprises an acquisition unit 11, a verification unit and a verification unit, wherein the acquisition unit is used for acquiring an image to be verified and identification information of the image, and the image comprises a human face;
an input unit 12, configured to input the image and the identification information into a first deep learning network model trained in advance, so as to obtain face region information in the image, where the face region information includes a coordinate position of at least one face region in the image;
the image processing unit 13 is configured to perform image processing on the at least one face region according to a preset rule to obtain a plurality of to-be-verified face images with different face angles, where the to-be-verified face images correspond to each face region in the at least one face region;
the input unit 12 is further configured to input the multiple to-be-verified face images with different face angles corresponding to each of the at least one face region into a second deep learning network model trained in advance, so as to obtain face feature combination information of each face image, where the face feature combination information is feature information obtained by combining face feature information corresponding to each face image;
the obtaining unit 11 is further configured to obtain a face verification result of the image to be verified according to the face feature combination information of each face image.
An embodiment of the present application further provides a computer-readable storage medium, where a face verification program is stored on the computer-readable storage medium, and when executed by a processor, the face verification program implements the steps of the face verification method according to the embodiments of the methods, for example, the method includes:
acquiring an image to be verified and identification information of the image, wherein the image comprises a human face;
inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises a coordinate position of at least one face region in the image;
performing image processing on the at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area;
inputting a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area into a pre-trained second deep learning network model to obtain face feature combination information of each face image, wherein the face feature combination information is feature information obtained by combining face feature information corresponding to each face image;
and acquiring a face verification result of the image to be verified according to the face feature combination information of each face image.
Referring to fig. 6, an embodiment of the present application further provides an electronic device 610, including: a processor 611, a memory 612 and a bus 613, wherein the memory 612 stores machine-readable instructions executable by the processor 612, when the electronic device 610 is running, the processor 611 and the memory 612 communicate via the bus 613, and the machine-readable instructions when executed by the processor 611 perform the steps of the method for face verification according to the method embodiments, for example, the method includes:
acquiring an image to be verified and identification information of the image, wherein the image comprises a human face;
inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises a coordinate position of at least one face region in the image;
performing image processing on the at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area;
inputting a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area into a pre-trained second deep learning network model to obtain face feature combination information of each face image, wherein the face feature combination information is feature information obtained by combining face feature information corresponding to each face image;
and acquiring a face verification result of the image to be verified according to the face feature combination information of each face image.
For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functions of the units may be implemented in the same software and/or hardware or in a plurality of software and/or hardware when implementing the invention.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for apparatus or system embodiments, since they are substantially similar to method embodiments, they are described in relative terms, as long as they are described in partial descriptions of method embodiments. The above-described embodiments of the apparatus and system are merely illustrative, and the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
It is noted that, in this document, relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The foregoing are merely exemplary embodiments of the present invention, which enable those skilled in the art to understand or practice the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (10)
1. A face verification method, comprising the steps of:
acquiring an image to be verified and identification information of the image, wherein the image comprises a human face;
inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises a coordinate position of at least one face region in the image;
performing image processing on the at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area;
inputting a plurality of to-be-verified face images with different face angles corresponding to each face area in the at least one face area into a pre-trained second deep learning network model to obtain face feature combination information of each face image, wherein the face feature combination information is feature information obtained by combining face feature information corresponding to each face image;
and acquiring a face verification result of the image to be verified according to the face feature combination information of each face image.
2. The method of claim 1, wherein the image processing on the at least one face region according to a preset rule to obtain a plurality of face images to be verified with different face angles corresponding to each face region in the at least one face region comprises:
determining a reference line of the image according to the coordinate position of at least one face area in the image;
and rotating the image along the reference line according to a preset rotation angle and/or preset rotation times to obtain a plurality of to-be-verified face images with different face angles, which correspond to each face area in the at least one face area.
3. The method according to claim 1, wherein the inputting a plurality of to-be-verified face images with different face angles corresponding to each of the at least one face region into a second deep learning network model trained in advance to obtain the face feature combination information of each of the face images comprises:
acquiring a face feature vector of each face image;
combining the face feature vectors to obtain the face feature combined vector;
and taking the face feature combination vector as the face feature combination information.
4. The method of claim 3, wherein the combining the face feature vectors to obtain the face feature combined vector comprises:
combining the face feature vectors according to a preset combination mode to obtain the face feature combination vectors, wherein the preset combination mode comprises at least one of the following combinations: affine combinations, linear combinations, and convex combinations.
5. The method according to claim 1, wherein the obtaining a face verification result of the image to be verified according to the face feature combination information of each face image comprises:
searching sample identification information corresponding to the identification information and sample face feature information corresponding to the sample identification information from a database;
matching the sample face feature information with each face feature combination information to obtain a matching result;
and acquiring a face verification result of the image to be verified according to the matching result.
6. The method of claim 5, wherein the matching the sample face feature information with each of the face feature combination information to obtain a matching result comprises:
respectively calculating difference values of the face feature combination information and the sample face feature information;
comparing each difference value with a set difference threshold value;
wherein, the obtaining the face verification result of the image to be verified according to the matching result comprises:
when any difference value is smaller than the difference threshold value, determining that sample face feature information matched with the face feature combination information exists in the preset database, and the face verification of the image is successful;
and when the difference values are all larger than or equal to the difference threshold value, determining that sample face feature information matched with the face feature combination information does not exist in the preset database, and failing to verify the face of the image.
7. The method of claim 6, wherein the calculating the difference between each of the face feature combination information and the sample face feature information comprises:
calculating Euclidean distance between each piece of face feature combination information and the sample face feature information;
and taking the value of the Euclidean distance as the difference value.
8. A face verification system, the system comprising:
the system comprises an acquisition unit, a verification unit and a verification unit, wherein the acquisition unit is used for acquiring an image to be verified and identification information of the image, and the image comprises a human face;
the input unit is used for inputting the image and the identification information into a first deep learning network model trained in advance to obtain face region information in the image, wherein the face region information comprises the coordinate position of at least one face region in the image;
the image processing unit is used for carrying out image processing on the at least one face area according to a preset rule to obtain a plurality of face images to be verified with different face angles, wherein the face images correspond to each face area in the at least one face area;
the input unit is further configured to input the multiple to-be-verified face images with different face angles, which correspond to each of the at least one face region, into a second deep learning network model trained in advance to obtain face feature combination information of each face image, where the face feature combination information is feature information obtained by combining face feature information corresponding to each face image;
the obtaining unit is further configured to obtain a face verification result of the image to be verified according to the face feature combination information of each face image.
9. A computer-readable storage medium, having stored thereon a face authentication program which, when executed by a processor, performs the face authentication method according to any one of claims 1 to 7.
10. An electronic device, comprising: a processor, a memory and a bus, the memory storing machine-readable instructions executable by the processor, the processor and the memory communicating over the bus when the electronic device is operating, the machine-readable instructions when executed by the processor performing the face verification method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910913242.2A CN110852150B (en) | 2019-09-25 | 2019-09-25 | Face verification method, system, equipment and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910913242.2A CN110852150B (en) | 2019-09-25 | 2019-09-25 | Face verification method, system, equipment and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110852150A true CN110852150A (en) | 2020-02-28 |
| CN110852150B CN110852150B (en) | 2022-12-20 |
Family
ID=69595933
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910913242.2A Active CN110852150B (en) | 2019-09-25 | 2019-09-25 | Face verification method, system, equipment and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110852150B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401242A (en) * | 2020-03-16 | 2020-07-10 | Oppo广东移动通信有限公司 | Certificate photo detection method and device, electronic equipment and storage medium |
| CN112836660A (en) * | 2021-02-08 | 2021-05-25 | 上海卓繁信息技术股份有限公司 | Face library generation method and device for monitoring field and electronic equipment |
Citations (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101378444A (en) * | 2007-08-30 | 2009-03-04 | 精工爱普生株式会社 | Image processing device, image processing method, and image processing program |
| CN101771539A (en) * | 2008-12-30 | 2010-07-07 | 北京大学 | Face recognition based method for authenticating identity |
| CN102799901A (en) * | 2012-07-10 | 2012-11-28 | 辉路科技（北京）有限公司 | Method for multi-angle face detection |
| CN103227877A (en) * | 2012-01-31 | 2013-07-31 | 京瓷办公信息系统株式会社 | Image forming apparatus and image forming method |
| CN103873755A (en) * | 2012-12-14 | 2014-06-18 | 鸿富锦精密工业（深圳）有限公司 | System and method for shooting leaping people |
| CN105718863A (en) * | 2016-01-15 | 2016-06-29 | 北京海鑫科金高科技股份有限公司 | Living-person face detection method, device and system |
| CN106056059A (en) * | 2016-05-20 | 2016-10-26 | 合肥工业大学 | Multidirectional SLGS characteristic description and performance cloud weight fusion face recognition method |
| WO2017031886A1 (en) * | 2015-08-26 | 2017-03-02 | 北京奇虎科技有限公司 | Method for obtaining picture by means of remote control, and server |
| CN106503687A (en) * | 2016-11-09 | 2017-03-15 | 合肥工业大学 | The monitor video system for identifying figures of fusion face multi-angle feature and its method |
| CN107609459A (en) * | 2016-12-15 | 2018-01-19 | 平安科技（深圳）有限公司 | A kind of face identification method and device based on deep learning |
| CN108268864A (en) * | 2018-02-24 | 2018-07-10 | 达闼科技（北京）有限公司 | Face identification method, system, electronic equipment and computer program product |
| CN108563992A (en) * | 2018-03-13 | 2018-09-21 | 苏州奥科德瑞智能科技有限公司 | A kind of vision survey system based on recognition of face |
| CN109002767A (en) * | 2018-06-22 | 2018-12-14 | 恒安嘉新（北京）科技股份公司 | A kind of face verification method and system based on deep learning |
| CN109522786A (en) * | 2018-09-26 | 2019-03-26 | 珠海横琴现联盛科技发展有限公司 | Dynamic human face method for registering based on 3D camera |
| CN109784243A (en) * | 2018-12-29 | 2019-05-21 | 网易（杭州）网络有限公司 | Identity determines method and device, neural network training method and device, medium |
| CN109800648A (en) * | 2018-12-18 | 2019-05-24 | 北京英索科技发展有限公司 | Face datection recognition methods and device based on the correction of face key point |
| CN109948586A (en) * | 2019-03-29 | 2019-06-28 | 北京三快在线科技有限公司 | Method, apparatus, equipment and the storage medium of face verification |
-
2019
- 2019-09-25 CN CN201910913242.2A patent/CN110852150B/en active Active
Patent Citations (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101378444A (en) * | 2007-08-30 | 2009-03-04 | 精工爱普生株式会社 | Image processing device, image processing method, and image processing program |
| CN102096900A (en) * | 2007-08-30 | 2011-06-15 | 精工爱普生株式会社 | Image processing device, image processing method, and image processing program |
| CN101771539A (en) * | 2008-12-30 | 2010-07-07 | 北京大学 | Face recognition based method for authenticating identity |
| CN103227877A (en) * | 2012-01-31 | 2013-07-31 | 京瓷办公信息系统株式会社 | Image forming apparatus and image forming method |
| CN102799901A (en) * | 2012-07-10 | 2012-11-28 | 辉路科技（北京）有限公司 | Method for multi-angle face detection |
| CN103873755A (en) * | 2012-12-14 | 2014-06-18 | 鸿富锦精密工业（深圳）有限公司 | System and method for shooting leaping people |
| WO2017031886A1 (en) * | 2015-08-26 | 2017-03-02 | 北京奇虎科技有限公司 | Method for obtaining picture by means of remote control, and server |
| CN105718863A (en) * | 2016-01-15 | 2016-06-29 | 北京海鑫科金高科技股份有限公司 | Living-person face detection method, device and system |
| CN106056059A (en) * | 2016-05-20 | 2016-10-26 | 合肥工业大学 | Multidirectional SLGS characteristic description and performance cloud weight fusion face recognition method |
| CN106503687A (en) * | 2016-11-09 | 2017-03-15 | 合肥工业大学 | The monitor video system for identifying figures of fusion face multi-angle feature and its method |
| CN107609459A (en) * | 2016-12-15 | 2018-01-19 | 平安科技（深圳）有限公司 | A kind of face identification method and device based on deep learning |
| CN108268864A (en) * | 2018-02-24 | 2018-07-10 | 达闼科技（北京）有限公司 | Face identification method, system, electronic equipment and computer program product |
| CN108563992A (en) * | 2018-03-13 | 2018-09-21 | 苏州奥科德瑞智能科技有限公司 | A kind of vision survey system based on recognition of face |
| CN109002767A (en) * | 2018-06-22 | 2018-12-14 | 恒安嘉新（北京）科技股份公司 | A kind of face verification method and system based on deep learning |
| CN109522786A (en) * | 2018-09-26 | 2019-03-26 | 珠海横琴现联盛科技发展有限公司 | Dynamic human face method for registering based on 3D camera |
| CN109800648A (en) * | 2018-12-18 | 2019-05-24 | 北京英索科技发展有限公司 | Face datection recognition methods and device based on the correction of face key point |
| CN109784243A (en) * | 2018-12-29 | 2019-05-21 | 网易（杭州）网络有限公司 | Identity determines method and device, neural network training method and device, medium |
| CN109948586A (en) * | 2019-03-29 | 2019-06-28 | 北京三快在线科技有限公司 | Method, apparatus, equipment and the storage medium of face verification |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401242A (en) * | 2020-03-16 | 2020-07-10 | Oppo广东移动通信有限公司 | Certificate photo detection method and device, electronic equipment and storage medium |
| CN112836660A (en) * | 2021-02-08 | 2021-05-25 | 上海卓繁信息技术股份有限公司 | Face library generation method and device for monitoring field and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110852150B (en) | 2022-12-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3954484B2 (en) | Image processing apparatus and program | |
| WO2021036436A1 (en) | Facial recognition method and apparatus | |
| EP2580711B1 (en) | Distinguishing live faces from flat surfaces | |
| US9984280B2 (en) | Object recognition system using left and right images and method | |
| CN111145214A (en) | Target tracking method, device, terminal equipment and medium | |
| WO2019071664A1 (en) | Human face recognition method and apparatus combined with depth information, and storage medium | |
| CN113823001A (en) | Method, device, equipment and medium for generating house type graph | |
| CN105554486A (en) | Projection calibration method and device | |
| CN105981046A (en) | Fingerprint authentication using stitch and cut | |
| CN110852150B (en) | Face verification method, system, equipment and computer readable storage medium | |
| US11816923B2 (en) | Face image candidate determination apparatus for authentication, face image candidate determination method for authentication, program, and recording medium | |
| CN111914775B (en) | Living body detection method, living body detection device, electronic equipment and storage medium | |
| CN107622246B (en) | Face recognition method and related product | |
| US11729511B2 (en) | Method for wall line determination, method, apparatus, and device for spatial modeling | |
| CN108335099A (en) | Method, apparatus, mobile terminal and the storage medium of mobile payment | |
| US20210312163A1 (en) | Face recognition method, device and electronic equipment, and computer non-volatile readable storage medium | |
| CN105912912A (en) | Method and system for user to log in terminal by virtue of identity information | |
| US20220270396A1 (en) | Palm print recognition method, method for training feature extraction model, device, and medium | |
| CN108229357A (en) | Method, apparatus, mobile terminal and the storage medium of recognition of face | |
| CN111222452A (en) | Face matching method and device, electronic equipment and readable storage medium | |
| Nieto et al. | Non-linear optimization for robust estimation of vanishing points | |
| CN110245573A (en) | A kind of register method, apparatus and terminal device based on recognition of face | |
| CN107832598B (en) | Unlocking control method and related product | |
| CN103984931A (en) | Information processing method and first electronic equipment | |
| CN111881740A (en) | Face recognition method, face recognition device, electronic equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |