CN111723755A - Optimization method and system of face recognition base - Google Patents

Optimization method and system of face recognition base Download PDF

Info

Publication number
CN111723755A
CN111723755A CN202010586533.8A CN202010586533A CN111723755A CN 111723755 A CN111723755 A CN 111723755A CN 202010586533 A CN202010586533 A CN 202010586533A CN 111723755 A CN111723755 A CN 111723755A
Authority
CN
China
Prior art keywords
face
glasses
base
face image
image
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010586533.8A
Other languages
Chinese (zh)
Other versions
CN111723755B (en
Inventor
杨帆
朱莹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xiaoshi Technology Jiangsu Co ltd
Original Assignee
Nanjing Zhenshi Intelligent Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing Zhenshi Intelligent Technology Co Ltd filed Critical Nanjing Zhenshi Intelligent Technology Co Ltd
Priority to CN202010586533.8A priority Critical patent/CN111723755B/en
Publication of CN111723755A publication Critical patent/CN111723755A/en
Application granted granted Critical
Publication of CN111723755B publication Critical patent/CN111723755B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G06V40/161—Detection; Localisation; Normalisation
    • G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/168—Feature extraction; Face representation
    • G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships

Landscapes

  • Engineering & Computer Science (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Human Computer Interaction (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Engineering & Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Geometry (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Collating Specific Patterns (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

The invention provides an optimization method and a system of a face recognition base library, which comprises the following steps: acquiring a face image from a face recognition base; detecting whether a human face in the human face image wears glasses or not; responding to the situation that the glasses are not worn on the face in the extracted face image, synthesizing the face image with a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image; respectively extracting the original face image and the corresponding synthesized features of the face image wearing the glasses to generate multi-dimensional base features; and adopting a PCA algorithm to perform dimensionality reduction processing on the multi-dimensional base features to obtain final base features. The invention aims to optimize a face recognition base by synthesizing images worn with glasses, synthesize a plurality of pictures worn with different glasses according to an original base to form a multi-dimensional base, reduce the dimension of the features of the multi-dimensional base by PCA and reduce the influence of wearing glasses and different types of glasses on face recognition of a mask.

Description

Optimization method and system of face recognition base
Technical Field
The invention relates to the technical field of face recognition, in particular to a method and a system for optimizing a face recognition base.
Background
The human face recognition is a biological recognition technology for carrying out identity recognition based on human face feature information, the features of an input human face image are compared with the features of a bottom library human face image one by one, a bottom library image with the highest similarity with the features of the input image is found out, if the similarity is larger than a preset similarity threshold value, the bottom library image and the input image are the same person, otherwise, the identity of the input image cannot be determined.
The quality of the face image of the bottom library directly affects the face recognition effect, and a high-quality face photo needs to be adopted as the bottom library in a face recognition scene without wearing a mask. In a face recognition scene of wearing a mask, the mask covers important information such as the mouth and the nose of a human face, the recognition is mainly performed in an eye region, a recognition algorithm (particularly eye features) based on local attention enhancement is facilitated, and the wearing condition of glasses can generate large interference on face recognition. Because the human face base library is obtained by collecting human face pictures with higher quality to extract features, especially under the conditions of no glasses and no mask, under the human face recognition scene of wearing the mask, when the glasses information of the actually obtained human face images and the human face images in the base library are not matched, the phenomenon of false recognition or recognition failure is easy to occur.
Disclosure of Invention
In order to achieve the above object, a first aspect of the present invention provides a method for optimizing a face recognition base, including the following steps:
acquiring a face image from a face recognition base;
detecting whether a human face in the human face image wears glasses or not;
responding to the situation that the glasses are not worn on the face in the extracted face image, synthesizing the face image with a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image;
respectively extracting the original face image and the corresponding synthesized features of the face image wearing the glasses to generate multi-dimensional base features; and
and performing dimensionality reduction on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
Preferably, the synthesis of the face image and the glasses image comprises the steps of firstly extracting eye key points of the face, aligning glasses with the eyes according to the eye key points, and then synthesizing the glasses and the face according to a mask image of the glasses to form the face image with the glasses, so as to generate the multi-dimensional face base image.
Preferably, the dimension reduction process includes:
for the multi-dimensional base features corresponding to one face feature extracted from the face recognition base, the multi-dimensional base features comprise (m-1) face image features with glasses and 1 original face image feature, and the multi-dimensional base feature is X ∈ Rm×nEach base library feature
Figure BDA0002554031060000021
The covariance matrix of X is calculated as follows:
Figure BDA0002554031060000022
wherein the content of the first and second substances,
Figure BDA0002554031060000023
performing characteristic decomposition on the covariance matrix C to obtain m eigenvalues lambda1,λ2...λmAnd m corresponding feature vectors u1,u2...umAt the maximum eigenvalue λmaxCorresponding feature vector u ∈ R1×mIs the first main component, and Y is the single-dimensional characteristic Y ∈ R after dimensionality reduction1×n。
According to the second aspect of the present invention, there is provided a system for optimizing a face recognition base, including:
the image acquisition module is used for acquiring a face image from the face recognition base;
the detection module is used for detecting whether the human face in the human face image wears glasses or not;
a synthesizing module for synthesizing the face image and a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image in response to that the glasses are not worn on the face in the extracted face image;
the characteristic extraction module is used for respectively extracting the characteristics of the original face image and the corresponding synthesized face image wearing the glasses and generating the characteristic of the multi-dimensional base; and
and the dimension reduction module is used for performing dimension reduction processing on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
According to a third aspect of the present invention, there is provided a system for optimizing a face recognition base, including:
one or more processors;
a memory storing instructions that are operable, when executed by the one or more processors, to cause the one or more processors to perform operations comprising:
acquiring a face image from a face recognition base;
detecting whether a human face in the human face image wears glasses or not;
responding to the situation that the glasses are not worn on the face in the extracted face image, synthesizing the face image with a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image;
respectively extracting the original face image and the corresponding synthesized features of the face image wearing the glasses to generate multi-dimensional base features; and
and performing dimensionality reduction on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
It should be understood that all combinations of the foregoing concepts and additional concepts described in greater detail below can be considered as part of the inventive subject matter of this disclosure unless such concepts are mutually inconsistent. In addition, all combinations of claimed subject matter are considered a part of the presently disclosed subject matter.
The foregoing and other aspects, embodiments and features of the present teachings can be more fully understood from the following description taken in conjunction with the accompanying drawings. Additional aspects of the present invention, such as features and/or advantages of exemplary embodiments, will be apparent from the description which follows, or may be learned by practice of specific embodiments in accordance with the teachings of the present invention.
Drawings
The drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component that is illustrated in various figures may be represented by a like numeral. For purposes of clarity, not every component may be labeled in every drawing. Embodiments of various aspects of the present invention will now be described, by way of example, with reference to the accompanying drawings, in which:
FIG. 1 is an exemplary flow of a method for optimization of a face recognition base in an exemplary embodiment of the invention;
FIG. 2 is a block diagram of an optimization system for a face recognition base of an exemplary embodiment of the present invention;
fig. 3 is a block diagram of a computer system on which an optimization method of a face recognition base is based according to an exemplary embodiment of the present invention.
Fig. 4a is a schematic view of the location of the eye-nose key points, and fig. 4b is a schematic view of a selected nasal bridge frame region.
Fig. 5a is a schematic diagram of a bridge frame region, fig. 5b is a schematic diagram of edge binarization, and fig. 5c is a schematic diagram of edge detection.
Detailed Description
In order to better understand the technical content of the present invention, specific embodiments are described below with reference to the accompanying drawings.
In this disclosure, aspects of the present invention are described with reference to the accompanying drawings, in which a number of illustrative embodiments are shown. Embodiments of the present disclosure are not necessarily intended to include all aspects of the invention. It should be appreciated that the various concepts and embodiments described above, as well as those described in greater detail below, may be implemented in any of numerous ways, as the disclosed concepts and embodiments are not limited to any one implementation. In addition, some aspects of the present disclosure may be used alone, or in any suitable combination with other aspects of the present disclosure.
The optimization method of the face recognition base library according to the exemplary embodiment of the invention aims to optimize the face recognition base library by synthesizing images worn with glasses, synthesize a plurality of pictures worn with different glasses according to an original base library to form a multi-dimensional base library, and reduce the influence of wearing glasses and different types of glasses on face recognition worn with a mask by reducing the dimension of the features of the multi-dimensional base library through PCA.
With reference to fig. 1 to 3, the method for optimizing the face recognition base includes the following steps: acquiring a face image from a face recognition base; detecting whether a human face in the human face image wears glasses or not; responding to the situation that the glasses are not worn on the face in the extracted face image, synthesizing the face image with a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image; respectively extracting the original face image and the corresponding synthesized features of the face image wearing the glasses to generate multi-dimensional base features; and adopting a PCA algorithm to perform dimensionality reduction processing on the multi-dimensional base features to obtain final base features.
Therefore, in the optimization process of the embodiment of the invention, the main face base database characteristics are obtained by synthesizing the face base database pictures with a plurality of glasses and PCA dimension reduction, so that the influence of the glasses on face recognition of the mask can be greatly reduced.
Referring to fig. 1, corresponding processing is performed according to a detection result of whether glasses are worn by a face in a face image, wherein if glasses are worn by the face in the face image taken out, face features are directly extracted from the face image taken out from the base library to serve as final base library features. And if the detected result is that the glasses are not worn, synthesizing the human face and the glasses to perform multi-maintenance image generation and feature extraction.
In order to ensure the accuracy of the optimization of the face bottom library, since the lenses are different in size and difficult to determine the region, the embodiment of the invention avoids the traditional dichotomy method for detecting the wearing of glasses (such as the traditional glasses recognizer of the mobilenet v3 two-classification network), and performs the glasses recognition by judging whether the bridge of the nose has a glasses frame. In the preferred embodiment, the state of wearing the glasses is detected through the edge information of the glasses frame, and the specific detection and identification process is as follows:
nose bridge frame region (x, y, w, h) is represented by left eye keypoint coordinates (x)1，y1) Coordinates of key points for the right eye (x)2，y2) And nose key point coordinates (x)3，y3) Determining that the coordinates of the central points of the key points of the left eye and the right eye are (x)c，yc) The calculation is shown as follows:
xc＝(x2+x1)/2
yc＝(y2+y1)/2
determining the width w according to the horizontal distance of key points of the two eyes, determining the height h according to the vertical distance between the central point of the key points of the left eye and the right eye and the nose, and finally calculating the coordinates (x, y) of the starting point of the nose bridge frame region:
w＝(x2-xc)/2
h＝y3-yc
x＝xc-w/2
y＝yc-h/3
and (3) detecting the edge information of the nose bridge frame region by using a canny edge detection method by combining the positions of the eye-nose key points in the graph of fig. 4a and the selected nose bridge frame region shown in fig. 4b to obtain an edge binary image of the region. As shown in fig. 5a-5c, wherein 5a is a nose bridge frame region schematic, fig. 5b is an edge binarization schematic, and fig. 5c is an edge detection schematic.
In combination with the drawings, the area with the glasses frame is white, and the rest is black. In the embodiment of the invention, in order to enhance the information of the edge of the spectacle frame, the edge binary image is subjected to morphological dilation processing to obtain a final edge detection image.
And then scanning a nose bridge frame region from top to bottom by adopting a sliding window with the width of w and the height of h/10, calculating the proportion of white pixel points in the sliding window, judging to wear glasses if the proportion exceeds half, and judging not to wear glasses and needing to perform face and glasses synthesis processing on the picture if the proportion does not exceed half.
Preferably, the synthesis of the face image and the glasses image comprises the steps of firstly extracting eye key points of the face, aligning the glasses with the eyes according to the eye key points, and then synthesizing the glasses and the face according to a mask image of the glasses to form the face image with the glasses, so that a multi-dimensional face base image is generated.
The extraction of the eye key points of the face image is realized by adopting a multi-scale face detection model of multi-task learning, and the eye key points are detected and extracted. Then, an alignment matrix is determined by matching the eye with the features of the key points of the glasses, and the synthesis of the two is performed based on the mask map of the glasses.
Preferably, the alignment matrix may be determined based on an affine transformation/perspective transformation.
Preferably, the dimension reduction process comprises:
for the multi-dimensional base features corresponding to one face feature extracted from the face recognition base, the multi-dimensional base features comprise (m-1) face image features with glasses and 1 original face image feature, and the multi-dimensional base feature is X ∈ Rm×nEach base library feature
Figure BDA0002554031060000051
The covariance matrix of X is calculated as follows:
Figure BDA0002554031060000052
wherein the content of the first and second substances,
Figure BDA0002554031060000053
performing characteristic decomposition on the covariance matrix C to obtain m eigenvalues lambda1,λ2...λmAnd m corresponding feature vectors u1,u2...umAt the maximum eigenvalue λmaxCorresponding feature vector u ∈ R1×mIs the first main component, and Y is the single-dimensional characteristic Y ∈ R after dimensionality reduction1×n。
With reference to fig. 2, according to the improvement of the present invention, there is provided an optimization system for a face recognition base, including:
the image acquisition module is used for acquiring a face image from the face recognition base;
the detection module is used for detecting whether the human face in the human face image wears glasses or not;
a synthesizing module for synthesizing the face image and a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image in response to that the glasses are not worn on the face in the extracted face image;
the characteristic extraction module is used for respectively extracting the characteristics of the original face image and the corresponding synthesized face image wearing the glasses and generating the characteristic of the multi-dimensional base; and
and the dimension reduction module is used for performing dimension reduction processing on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
Preferably, the synthesis module comprises:
a module for extracting eye key points of a human face;
means for aligning the eyewear with the eyes according to the eye keypoints;
and the module is used for synthesizing the glasses and the human face according to the mask image of the glasses to form a human face image wearing the glasses, so that a multi-dimensional human face base image is generated.
Preferably, the dimension reduction module is configured to perform the dimension reduction process in the following manner:
for the multi-dimensional base features corresponding to one face feature extracted from the face recognition base, the multi-dimensional base features comprise (m-1) face image features with glasses and 1 original face image feature, and the multi-dimensional base feature is X ∈ Rm×nEach base library feature
Figure BDA0002554031060000061
The covariance matrix of X is calculated as follows:
Figure BDA0002554031060000062
wherein the content of the first and second substances,
Figure BDA0002554031060000063
performing characteristic decomposition on the covariance matrix C to obtain m eigenvalues lambda1,λ2...λmAnd m corresponding feature vectors u1,u2...umAt the maximum eigenvalue λmaxCorresponding feature vector u ∈ R1×mIs the first main component, and Y is the single-dimensional characteristic Y ∈ R after dimensionality reduction1×n。
FIG. 3 is an exemplary computer system showing the implementation of an optimization scheme for a face recognition base. With reference to fig. 1, 2, and 3, according to an embodiment of the disclosure, a system for optimizing a face recognition base is further provided, including:
one or more processors;
a memory storing instructions that are operable, when executed by the one or more processors, to cause the one or more processors to perform operations comprising:
acquiring a face image from a face recognition base;
detecting whether a human face in the human face image wears glasses or not;
responding to the situation that the glasses are not worn on the face in the extracted face image, synthesizing the face image with a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image;
respectively extracting the original face image and the corresponding synthesized features of the face image wearing the glasses to generate multi-dimensional base features; and
and performing dimensionality reduction on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
Although the present invention has been described with reference to the preferred embodiments, it is not intended to be limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, the protection scope of the present invention should be determined by the appended claims.

Claims (10)

1. A method for optimizing a face recognition base library is characterized by comprising the following steps:
acquiring a face image from a face recognition base;
detecting whether a human face in the human face image wears glasses or not;
responding to the situation that the glasses are not worn on the face in the extracted face image, synthesizing the face image with a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image;
respectively extracting the original face image and the corresponding synthesized features of the face image wearing the glasses to generate multi-dimensional base features; and
and performing dimensionality reduction on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
2. The method for optimizing a face recognition base according to claim 1, wherein in response to the face in the extracted face image wearing glasses, the face features are extracted directly for the face image extracted from the base as final base features.
3. The optimization method of the face recognition base library according to claim 1, wherein the synthesis of the face image and the glasses image comprises the steps of firstly extracting eye key points of a face, aligning glasses with the eyes according to the eye key points, and then synthesizing the glasses and the face according to a mask image of the glasses to form a face image with the glasses, thereby generating a multi-dimensional face base library image.
4. The method for optimizing the face recognition base according to claim 1, wherein the process of dimension reduction processing comprises:
for the multi-dimensional base features corresponding to one face feature extracted from the face recognition base, the multi-dimensional base features comprise (m-1) face image features with glasses and 1 original face image feature, and the multi-dimensional base feature is X ∈ Rm×nEach base library feature
Figure FDA0002554031050000011
n is less than or equal to m-1, and calculating the covariance matrix of X as shown in the following formula:
Figure FDA0002554031050000012
wherein the content of the first and second substances,
Figure FDA0002554031050000013
performing characteristic decomposition on the covariance matrix C to obtain m eigenvalues lambda1,λ2...λmAnd m corresponding feature vectors u1,u2...umAt the maximum eigenvalue λmaxCorresponding feature vector u ∈ R1×mIs the first main component, and Y is the single-dimensional characteristic Y ∈ R after dimensionality reduction1×n。
5. An optimization system for a face recognition base, comprising:
the image acquisition module is used for acquiring a face image from the face recognition base;
the detection module is used for detecting whether the human face in the human face image wears glasses or not;
a synthesizing module for synthesizing the face image and a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image in response to that the glasses are not worn on the face in the extracted face image;
the characteristic extraction module is used for respectively extracting the characteristics of the original face image and the corresponding synthesized face image wearing the glasses and generating the characteristic of the multi-dimensional base; and
and the dimension reduction module is used for performing dimension reduction processing on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
6. The system for optimizing a face recognition base library of claim 5, wherein the synthesis module comprises:
a module for extracting eye key points of a human face;
means for aligning the eyewear with the eyes according to the eye keypoints;
and the module is used for synthesizing the glasses and the human face according to the mask image of the glasses to form a human face image wearing the glasses, so that a multi-dimensional human face base image is generated.
7. The system for optimizing a face recognition base stock according to claim 6, wherein the dimension reduction module is configured to perform the dimension reduction process in the following manner:
for the multi-dimensional base features corresponding to one face feature extracted from the face recognition base, the multi-dimensional base features comprise (m-1) face image features with glasses and 1 original face image feature, and the multi-dimensional base feature is X ∈ Rm×nEach base library feature
Figure FDA0002554031050000021
n is less than or equal to m-1, and calculating the covariance matrix of X as shown in the following formula:
Figure FDA0002554031050000022
wherein the content of the first and second substances,
Figure FDA0002554031050000023
performing characteristic decomposition on the covariance matrix C to obtain m eigenvalues lambda1,λ2...λmAnd m corresponding feature vectors u1,u2...umAt the maximum eigenvalue λmaxCorresponding feature vector u ∈ R1×mIs the first main component, and Y is the single-dimensional characteristic Y ∈ R after dimensionality reduction1×n。
8. An optimization system for a face recognition base, comprising:
one or more processors;
a memory storing instructions that are operable, when executed by the one or more processors, to cause the one or more processors to perform operations comprising:
acquiring a face image from a face recognition base;
detecting whether a human face in the human face image wears glasses or not;
responding to the situation that the glasses are not worn on the face in the extracted face image, synthesizing the face image with a plurality of randomly selected glasses images to generate a corresponding multi-dimensional base face image;
respectively extracting the original face image and the corresponding synthesized features of the face image wearing the glasses to generate multi-dimensional base features; and
and performing dimensionality reduction on the multi-dimensional base features by adopting a PCA algorithm to obtain final base features.
9. The system for optimizing a human face recognition base library according to claim 8, wherein the operation of synthesizing the human face image and the glasses image comprises the following processes:
firstly, extracting eye key points of a human face, aligning glasses with the eyes according to the eye key points, and then synthesizing the glasses and the human face according to a mask image of the glasses to form a human face image wearing the glasses, thereby generating a multi-dimensional human face base image.
10. The system for optimizing a face recognition base library according to claim 8, wherein the process of dimension reduction processing comprises:
for the multi-dimensional base features corresponding to one face feature extracted from the face recognition base, the multi-dimensional base features comprise (m-1) face image features with glasses and 1 original face image feature, and the multi-dimensional base feature is X ∈ Rm×nEach base library feature
Figure FDA0002554031050000031
n is less than or equal to m-1, calculating XCovariance matrix, as shown in the following equation:
Figure FDA0002554031050000032
wherein the content of the first and second substances,
Figure FDA0002554031050000033
performing characteristic decomposition on the covariance matrix C to obtain m eigenvalues lambda1,λ2...λmAnd m corresponding feature vectors u1,u2...umAt the maximum eigenvalue λmaxCorresponding feature vector u ∈ R1×mIs the first main component, and Y is the single-dimensional characteristic Y ∈ R after dimensionality reduction1×n。
CN202010586533.8A 2020-07-19 2020-07-19 Optimization method and system of face recognition base Active CN111723755B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010586533.8A CN111723755B (en) 2020-07-19 2020-07-19 Optimization method and system of face recognition base

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010586533.8A CN111723755B (en) 2020-07-19 2020-07-19 Optimization method and system of face recognition base

Publications (2)

Publication Number Publication Date
CN111723755A true CN111723755A (en) 2020-09-29
CN111723755B CN111723755B (en) 2022-09-06

Family

ID=72568618

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010586533.8A Active CN111723755B (en) 2020-07-19 2020-07-19 Optimization method and system of face recognition base

Country Status (1)

Country Link
CN (1) CN111723755B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113435248A (en) * 2021-05-18 2021-09-24 武汉天喻信息产业股份有限公司 Mask face recognition base enhancement method, device, equipment and readable storage medium
CN114429663A (en) * 2022-01-28 2022-05-03 北京百度网讯科技有限公司 Updating method of human face base, human face recognition method, device and system

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105184253A (en) * 2015-09-01 2015-12-23 北京旷视科技有限公司 Face identification method and face identification system
CN108319943A (en) * 2018-04-25 2018-07-24 北京优创新港科技股份有限公司 A method of human face recognition model performance under the conditions of raising is worn glasses
CN111062328A (en) * 2019-12-18 2020-04-24 中新智擎科技有限公司 Image processing method and device and intelligent robot

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105184253A (en) * 2015-09-01 2015-12-23 北京旷视科技有限公司 Face identification method and face identification system
CN108319943A (en) * 2018-04-25 2018-07-24 北京优创新港科技股份有限公司 A method of human face recognition model performance under the conditions of raising is worn glasses
CN111062328A (en) * 2019-12-18 2020-04-24 中新智擎科技有限公司 Image processing method and device and intelligent robot

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113435248A (en) * 2021-05-18 2021-09-24 武汉天喻信息产业股份有限公司 Mask face recognition base enhancement method, device, equipment and readable storage medium
CN114429663A (en) * 2022-01-28 2022-05-03 北京百度网讯科技有限公司 Updating method of human face base, human face recognition method, device and system
CN114429663B (en) * 2022-01-28 2023-10-20 北京百度网讯科技有限公司 Updating method of face base, face recognition method, device and system

Also Published As

Publication number Publication date
CN111723755B (en) 2022-09-06

Similar Documents

Publication Publication Date Title
US10198623B2 (en) Three-dimensional facial recognition method and system
US9098760B2 (en) Face recognizing apparatus and face recognizing method
US8064653B2 (en) Method and system of person identification by facial image
US9626553B2 (en) Object identification apparatus and object identification method
Tsao et al. A data mining approach to face detection
CN109740572B (en) Human face living body detection method based on local color texture features
CN105956578A (en) Face verification method based on identity document information
Chan et al. Sparse representation of (multiscale) histograms for face recognition robust to registration and illumination problems
CN114359998B (en) Identification method of face mask in wearing state
CN111723755B (en) Optimization method and system of face recognition base
CN110991258B (en) Face fusion feature extraction method and system
Bąk et al. Person re-identification using deformable patch metric learning
Barbu An automatic face detection system for RGB images
JP2012160001A (en) Face image authentication apparatus
Devadethan et al. Face detection and facial feature extraction based on a fusion of knowledge based method and morphological image processing
Mohamed et al. Automated face recogntion system: Multi-input databases
CN111950403A (en) Iris classification method and system, electronic device and storage medium
Kwon Face recognition using depth and infrared pictures
Tao et al. Automatic localization of human eyes in complex background
CN109145875B (en) Method and device for removing black frame glasses in face image
Paul et al. Automatic adaptive facial feature extraction using CDF analysis
CN111723612A (en) Face recognition and face recognition network training method and device, and storage medium
US10438061B2 (en) Adaptive quantization method for iris image encoding
Paul et al. Extraction of facial feature points using cumulative distribution function by varying single threshold group
Thomas et al. Real Time Face Mask Detection and Recognition using Python

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
CP01 Change in the name or title of a patent holder
CP01 Change in the name or title of a patent holder

Address after: No.568 longmian Avenue, gaoxinyuan, Jiangning District, Nanjing City, Jiangsu Province, 211000

Patentee after: Xiaoshi Technology (Jiangsu) Co.,Ltd.

Address before: No.568 longmian Avenue, gaoxinyuan, Jiangning District, Nanjing City, Jiangsu Province, 211000

Patentee before: NANJING ZHENSHI INTELLIGENT TECHNOLOGY Co.,Ltd.