CN111209839B - Face recognition method - Google Patents
Face recognition method Download PDFInfo
- Publication number
- CN111209839B CN111209839B CN201911422660.8A CN201911422660A CN111209839B CN 111209839 B CN111209839 B CN 111209839B CN 201911422660 A CN201911422660 A CN 201911422660A CN 111209839 B CN111209839 B CN 111209839B
- Authority
- CN
- China
- Prior art keywords
- face
- face recognition
- feature vectors
- far
- recognition method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 239000013598 vector Substances 0.000 claims abstract description 64
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000013528 artificial neural network Methods 0.000 claims abstract description 5
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 19
- 241000208340 Araliaceae Species 0.000 claims description 8
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 8
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 8
- 235000008434 ginseng Nutrition 0.000 claims description 8
- 238000013135 deep learning Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000712 assembly Effects 0.000 description 1
- 238000000429 assembly Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a face recognition method, which comprises the following steps: s1, acquiring a face picture; s2, extracting face features through a deep neural network; s3, acquiring a first loss function; s4, acquiring a second loss function; s5, acquiring a total loss function; s6, carrying out similarity calculation on the two face feature vectors; and S7, a network training step. The face recognition method provided by the invention can improve the recognition accuracy. The invention can make the distance of the face feature vectors among different people as far as possible and the distance of the face feature vectors among the same person as close as possible.
Description
Technical Field
The invention belongs to the technical field of face recognition, relates to a face recognition method, and in particular relates to a face recognition method capable of improving recognition accuracy.
Background
In the face recognition field, the loss function is a loss function which is generally used when deep learning is utilized to optimize classification and recognition models
(the disclosed general method) for face recognition tasks, an excellent model is to have the distance between features of the same person's photograph as close as possible, and the distance between features of different persons' photographs as far as possible. However, softmax is just a classification that does not place the photo features of the same person as close as possible, but the photo features of different persons as far as possible. Arcface has been proposed.
In the method, in the process of the invention,
x * representing the neural network output layer. In face recognition, it is generally used as a face feature.
Representing that the face features are normalized;
The weight vector is normalized. At x i And W is ji Theta between ji Adding an angle interval m, punishing the angle between the photo features of the human face and the corresponding weights in an addition mode, so that the included angle between the photo features of each human face and the corresponding weights is as small as possible, namely the photo features of each human face are as close to the corresponding weight vectors as possible, thereby enhancing the intra-class tightness, namely the photo features of the same human face are as close as possible; s is a super-parameter, formulated during training.
As known above, the arcface loss function is optimized to have the photo features of each person as close as possible to the corresponding weight vector, thereby making the same person photo features as close as possible. But does not explicitly require that the different people photo features be as far as possible. As shown in fig. 1.
Therefore, if the distance between the weight vectors is as far as possible, the feature vectors of the face photos between different people can be as far as possible, so that the feature vectors of the same face photo are as close as possible, and the feature vectors of the different face photos are as far as possible.
In view of this, there is an urgent need to design a face recognition method so as to overcome the above-mentioned drawbacks of the existing recognition methods.
Disclosure of Invention
The invention provides a face recognition method which can improve recognition accuracy.
In order to solve the technical problems, according to one aspect of the present invention, the following technical scheme is adopted:
a face recognition method, the face recognition method comprising:
s1, acquiring a face picture;
s2, extracting face features through a deep neural network;
s3, acquiring a first loss function;
wherein N is the number of each batch of photos; s is radix Ginseng Rubra, m is radix Ginseng Rubra; w is a weight vector matrix;
s4, acquiring a second loss function;
t is a super ginseng;
s5, acquiring a total loss function; l=l arcface +λL edu The method comprises the steps of carrying out a first treatment on the surface of the Lambda is the super-ginseng;
in the training process, cascading convolutional neural network structures and total loss functions, and optimizing network parameters through continuous iteration; finally obtaining a face recognition model based on deep learning; extracting face features of each face photo by using the face recognition model obtained in the training stage;
s6, carrying out similarity calculation on the two face feature vectors, and adopting a cosine included angle mode;
let two face feature vectors be: f (F) 1 ＝[f 1,1 ,f 1,2 ,…,f 1,n-1 ,f 1,n ]，F 2 ＝[f 2,1 ,f 2,2 ,…,f 2,n-1 ,f 2,n ]N is the dimension of the face feature;
the similarity of the two face feature vectors is:
the larger the similarity value is, the more similar the two face features are, and the greater the possibility of the same person is;
s7, a network training step; for training network parameters, a random gradient descent method SGD with small batches of Mini-batch is used for training.
As an embodiment of the present invention, the weight of arcface
Is L edc As an optimization object.
As one embodiment of the invention, lambda is the super parameter, 0.5 is taken; t is the radix scrophulariae, and 2 is taken. The specific values of the parameters may be adjusted during the training process.
As one embodiment of the present invention, L edc Is responsible for optimizing the W weight vectors in arcface so that the W weight vectors are as far apart as possible; and L is arcface And the face photo feature vectors of the same person are optimized to be as close to the corresponding feature vectors as possible.
In one embodiment of the present invention
In->
The meaning of (1) is for the classification weight vector in each arcface +.>
And other classification weight vectors->
Is the shortest distance of (2); the larger this shortest distance, the smaller the loss function value (illustratively: the process of deep learning training, i.e., the process of letting the loss function value be smaller and smaller, which is generally referred to as training convergence, if the value is larger and larger, the training divergence); thereby making the distance between the weight vectors as large as possible; because of the weight vectorEach component corresponds to each person, and as long as the components of the weight vector are separated as far as possible, that is, the feature vectors of the face photos between each person are separated as far as possible, so that photo features between the same person can be obtained as close as possible, and photo features between different persons are as far as possible.
The invention has the beneficial effects that: the face recognition method provided by the invention can improve the recognition accuracy. The invention can make the distance of the face feature vectors among different people as far as possible and the distance of the face feature vectors among the same person as close as possible. Since each component of the weight vector is made as far as possible and each component of the weight vector corresponds to each person by training, as long as the respective components of the weight vector are separated as far as possible, that is, the feature vectors of the face photo between each person are separated as far as possible; for different photos of the same person, the feature vector of the different photos of each person is continuously close to the weight vector of the corresponding person in the training process, so that the photo features among the same person can be obtained as close as possible. Finally, photo features among different persons are as far as possible, and photo features of the same person are as close as possible.
Drawings
Fig. 1 is a flowchart of a face recognition method according to an embodiment of the present invention.
Detailed Description
Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
For a further understanding of the present invention, preferred embodiments of the invention are described below in conjunction with the examples, but it should be understood that these descriptions are merely intended to illustrate further features and advantages of the invention, and are not limiting of the claims of the invention.
The description of this section is intended to be illustrative of only a few exemplary embodiments and the invention is not to be limited in scope by the description of the embodiments. It is also within the scope of the description and claims of the invention to interchange some of the technical features of the embodiments with other technical features of the same or similar prior art.
The invention discloses a face recognition method, and fig. 1 is a flow chart of the face recognition method in an embodiment of the invention; referring to fig. 1, in an embodiment of the present invention, the face recognition method includes:
step S1, obtaining a face picture I;
s2, extracting face features X through a deep neural network;
s3, acquiring a first loss function;
wherein N is the number of each batch of photos; s is radix Ginseng Rubra, m is radix Ginseng Rubra; w is a weight vector matrix. In one embodiment of the invention, the weight of arcface
Is L edc As an optimization object.
S4, acquiring a second loss function;
t is a super ginseng; in one embodiment, t is taken as 2.
S5, acquiring a total loss function; l=l arcface +λL edu The method comprises the steps of carrying out a first treatment on the surface of the Lambda is the super-ginseng. In one embodiment, λ is the radix Ginseng Rubra, 0.5.
In the training process, the CNN structure (such as ResNet 34) and the total loss function of the convolutional neural network are cascaded, and the network parameters are continuously and iteratively optimized through a random gradient descent algorithm SGD. Finally, a face recognition model based on deep learning, namely network parameters of Resnet34, is obtained. And extracting the face characteristics of each face photo by utilizing the face recognition model obtained in the training stage.
In one embodiment of the invention, L edc Is responsible for optimizing the W weight vectors in arcface so that the W weight vectors are as far apart as possible; and L is arcface And the face photo feature vectors of the same person are optimized to be as close to the corresponding feature vectors as possible.
In one embodiment, in
In->
The meaning of (1) is for the classification weight vector in each arcface +.>
And other classification weight vectors->
Is the shortest distance of (2); the smaller this shortest distance, the larger the loss function value; the weight vectors are given as large as possible; since each component of the weight vector corresponds to each person, as long as the components of the weight vector are separated as far as possible, that is, the feature vectors of the face photos between each person are separated as far as possible, the photo features between the same person can be obtained as close as possible, and the photo features between different persons are as far as possible.
S6, carrying out similarity calculation on the two face feature vectors, and adopting a cosine included angle mode;
let two face feature vectors be: f (F) 1 ＝[f 1,1 ,f 1,2 ,…,f 1,n-1 ,f 1,n ]，F 2 ＝[f 2,1 ,f 2,2 ,…,f 2,n-1 ,f 2,n ]N is the dimension of the face feature;
the similarity of the two face feature vectors is:
the larger the similarity value is, the more similar the two face features are, and the greater the possibility of the same person is;
s7, a network training step; for training network parameters, a random gradient descent method SGD with small batches of Mini-batch is used for training. The method is a common convolutional neural network CNN parameter optimization method; convolutional neural network CNN employs mainstream Resnet50, resnet34, resnet100.
In summary, the face recognition method provided by the invention can improve the recognition accuracy. The invention can make the distance of the face feature vectors among different people as far as possible and the distance of the face feature vectors among the same person as close as possible. Since each component of the weight vector is made as far as possible and each component of the weight vector corresponds to each person by training, as long as the respective components of the weight vector are separated as far as possible, that is, the feature vectors of the face photo between each person are separated as far as possible; for different photos of the same person, the feature vector of the different photos of each person is continuously close to the weight vector of the corresponding person in the training process, so that the photo features among the same person can be obtained as close as possible. Finally, photo features among different persons are as far as possible, and photo features of the same person are as close as possible.
The technical features of the above-described embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above-described embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The description and applications of the present invention herein are illustrative and are not intended to limit the scope of the invention to the embodiments described above. Variations and modifications of the embodiments disclosed herein are possible, and alternatives and equivalents of the various components of the embodiments are known to those of ordinary skill in the art. It will be clear to those skilled in the art that the present invention may be embodied in other forms, structures, arrangements, proportions, and with other assemblies, materials, and components, without departing from the spirit or essential characteristics thereof. Other variations and modifications of the embodiments disclosed herein may be made without departing from the scope and spirit of the invention.
Claims (5)
1. A face recognition method, characterized in that the face recognition method comprises:
s1, acquiring a face picture;
s2, extracting face features through a deep neural network;
s3, acquiring a first loss function;
wherein N is the number of each batch of photos; s is a super parameter, m is an angular margin, and is a super parameter; w is a weight vector matrix;
s4, acquiring a second loss function;
t is a super ginseng;
at the position of
In->
The meaning of (1) is for the classification weight vector in each arcface +.>
And other classification weight vectors->
Is the shortest distance of (2); the larger this shortest distance, the smaller the loss function value; so that the distance between the weight vectors is as far as possible; because each component of the weight vector corresponds to each person, as long as the components of the weight vector are separated as far as possible, namely, the feature vectors of the face photos among each person are separated as far as possible, so that photo features among the same person can be obtained as close as possible, and photo features among different persons are as far as possible;
s5, acquiring a total loss function; l=l arcface +λL edu The method comprises the steps of carrying out a first treatment on the surface of the Lambda is the super-ginseng;
in the training process, cascading convolutional neural network structures and total loss functions, and optimizing network parameters through continuous iteration; finally obtaining a face recognition model based on deep learning; extracting face features of each face photo by using the face recognition model obtained in the training stage;
s6, carrying out similarity calculation on the two face feature vectors, and adopting a cosine included angle mode;
let two face feature vectors be: f (F) 1 ＝[f 1,1 ,f 1,2 ,…,f 1,n-1 ,f 1,n ]，F 2 ＝[f 2,1 ,f 2,2 ,…,f 2,n-1 ,f 2,n ]；
The similarity of the two face feature vectors is:
the larger the similarity value is, the more similar the two face features are, and the greater the possibility of the same person is;
s7, a network training step; for training network parameters, a random gradient descent method SGD with small batches of Mini-batch is used for training.
2. The face recognition method according to claim 1, wherein:
weights of arcface
Is L edc As an optimization object.
3. The face recognition method according to claim 1, wherein:
lambda is 0.5; t is the radix scrophulariae, and 2 is taken.
4. The face recognition method according to claim 1, wherein:
s is set to 30 and m is set to 0.1.
5. The face recognition method according to claim 1, wherein:
L edc is responsible for optimizing the W weight vectors in arcface so that the W weight vectors are as far apart as possible; and L is arcface And the face photo feature vectors of the same person are optimized to be as close to the corresponding feature vectors as possible.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911422660.8A CN111209839B (en) | 2019-12-31 | 2019-12-31 | Face recognition method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911422660.8A CN111209839B (en) | 2019-12-31 | 2019-12-31 | Face recognition method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111209839A CN111209839A (en) | 2020-05-29 |
| CN111209839B true CN111209839B (en) | 2023-05-23 |
Family
ID=70789467
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911422660.8A Active CN111209839B (en) | 2019-12-31 | 2019-12-31 | Face recognition method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111209839B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113069080B (en) * | 2021-03-22 | 2021-12-21 | 上海交通大学医学院附属第九人民医院 | Difficult airway assessment method and device based on artificial intelligence |
| CN112949618A (en) * | 2021-05-17 | 2021-06-11 | 成都市威虎科技有限公司 | Face feature code conversion method and device and electronic equipment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017215240A1 (en) * | 2016-06-14 | 2017-12-21 | 广州视源电子科技股份有限公司 | Neural network-based method and device for face feature extraction and modeling, and face recognition |
| WO2017219391A1 (en) * | 2016-06-24 | 2017-12-28 | 深圳市唯特视科技有限公司 | Face recognition system based on three-dimensional data |
| CN108647583A (en) * | 2018-04-19 | 2018-10-12 | 浙江大承机器人科技有限公司 | A kind of face recognition algorithms training method based on multiple target study |
| CN109033938A (en) * | 2018-06-01 | 2018-12-18 | 上海阅面网络科技有限公司 | A kind of face identification method based on ga s safety degree Fusion Features |
| CN109165566A (en) * | 2018-08-01 | 2019-01-08 | 中国计量大学 | A kind of recognition of face convolutional neural networks training method based on novel loss function |
| CN109214360A (en) * | 2018-10-15 | 2019-01-15 | 北京亮亮视野科技有限公司 | A kind of construction method of the human face recognition model based on ParaSoftMax loss function and application |
| WO2019128367A1 (en) * | 2017-12-26 | 2019-07-04 | 广州广电运通金融电子股份有限公司 | Face verification method and apparatus based on triplet loss, and computer device and storage medium |
| WO2019192121A1 (en) * | 2018-04-04 | 2019-10-10 | 平安科技（深圳）有限公司 | Dual-channel neural network model training and human face comparison method, and terminal and medium |
-
2019
- 2019-12-31 CN CN201911422660.8A patent/CN111209839B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017215240A1 (en) * | 2016-06-14 | 2017-12-21 | 广州视源电子科技股份有限公司 | Neural network-based method and device for face feature extraction and modeling, and face recognition |
| WO2017219391A1 (en) * | 2016-06-24 | 2017-12-28 | 深圳市唯特视科技有限公司 | Face recognition system based on three-dimensional data |
| WO2019128367A1 (en) * | 2017-12-26 | 2019-07-04 | 广州广电运通金融电子股份有限公司 | Face verification method and apparatus based on triplet loss, and computer device and storage medium |
| WO2019192121A1 (en) * | 2018-04-04 | 2019-10-10 | 平安科技（深圳）有限公司 | Dual-channel neural network model training and human face comparison method, and terminal and medium |
| CN108647583A (en) * | 2018-04-19 | 2018-10-12 | 浙江大承机器人科技有限公司 | A kind of face recognition algorithms training method based on multiple target study |
| CN109033938A (en) * | 2018-06-01 | 2018-12-18 | 上海阅面网络科技有限公司 | A kind of face identification method based on ga s safety degree Fusion Features |
| CN109165566A (en) * | 2018-08-01 | 2019-01-08 | 中国计量大学 | A kind of recognition of face convolutional neural networks training method based on novel loss function |
| CN109214360A (en) * | 2018-10-15 | 2019-01-15 | 北京亮亮视野科技有限公司 | A kind of construction method of the human face recognition model based on ParaSoftMax loss function and application |
Non-Patent Citations (1)
| Title |
|---|
| 刘施乐 ; .基于深度学习的人脸识别技术研究.电子制作.2018,(24),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111209839A (en) | 2020-05-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107748895B (en) | Unmanned aerial vehicle landing landform image classification method based on DCT-CNN model | |
| CN110458750B (en) | Unsupervised image style migration method based on dual learning | |
| CN112614077B (en) | Unsupervised low-illumination image enhancement method based on generation countermeasure network | |
| CN109064423B (en) | Intelligent image repairing method for generating antagonistic loss based on asymmetric circulation | |
| CN111209839B (en) | Face recognition method | |
| CN107330355B (en) | Deep pedestrian re-identification method based on positive sample balance constraint | |
| CN110660020B (en) | Image super-resolution method of antagonism generation network based on fusion mutual information | |
| CN111738303B (en) | Long-tail distribution image recognition method based on hierarchical learning | |
| CN108846822B (en) | Fusion method of visible light image and infrared light image based on hybrid neural network | |
| CN110969089B (en) | Lightweight face recognition system and recognition method in noise environment | |
| CN108960043A (en) | A kind of personage's family relationship construction method for electron album management | |
| CN111339931A (en) | Face recognition network model construction method combined with gender attribute training | |
| CN107357761A (en) | A kind of minimal error entropy computational methods of quantization | |
| CN110021049A (en) | A kind of highly concealed type antagonism image attack method based on space constraint towards deep neural network | |
| CN113361387A (en) | Face image fusion method and device, storage medium and electronic equipment | |
| CN113989405B (en) | Image generation method based on small sample continuous learning | |
| Huang et al. | Self-supervised medical image denoising based on wista-net for human healthcare in metaverse | |
| CN111291193B (en) | Application method of knowledge graph in zero-time learning | |
| Krushkal | Strengthened Moser’s conjecture, geometry of Grunsky coefficients and Fredholm eigenvalues | |
| JP2020123337A (en) | On-device continuous learning method and device of neural network for analyzing input data by optimal sampling of training image for smart phone, drone, ship, or military purpose, and test method and device using it | |
| CN114373205B (en) | Face detection and recognition method based on convolution width network | |
| CN110599403A (en) | Image super-resolution reconstruction method with good high-frequency visual effect | |
| JP2020170496A (en) | Age privacy protection method and system for face recognition | |
| CN116137023A (en) | Low-illumination image enhancement method based on background modeling and detail enhancement | |
| Yang et al. | Image super resolution using deep convolutional network based on topology aggregation structure |
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 |