CN109993039A - Portrait identification method and device, computer readable storage medium - Google Patents
Portrait identification method and device, computer readable storage medium Download PDFInfo
- Publication number
- CN109993039A CN109993039A CN201810003470.1A CN201810003470A CN109993039A CN 109993039 A CN109993039 A CN 109993039A CN 201810003470 A CN201810003470 A CN 201810003470A CN 109993039 A CN109993039 A CN 109993039A
- Authority
- CN
- China
- Prior art keywords
- facial image
- human facial
- image recognition
- recognition algorithm
- identification
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
A kind of portrait identification method and device, computer readable storage medium, the portrait identification method includes: to carry out first time identification to images to be recognized using the first human Facial Image Recognition Algorithm, and be compared with target person image, obtain the matching target image for meeting similarity threshold;The top N matching target image that last time is identified is as the images to be recognized of identification next time, until completing to carry out images to be recognized the m times identification using m human Facial Image Recognition Algorithm, and it is compared with the target person image, the matching target image of the similarity threshold will be met as final matching target image, N is positive integer, and m is the positive integer greater than 1.Using the above scheme, the misclassification rate of Identification of Images can be reduced.
Description
Technical field
The present embodiments relate to image identification technical field more particularly to a kind of portrait identification method and devices, calculating
Machine readable storage medium storing program for executing.
Background technique
In face recognition system, either the identification of portrait static state or portrait Dynamic Recognition, recognition effect are to measure people
As the technical indicator of the most important and core of identifying system.Recognition effect is mainly that following two specific targets is leaned on to form: just
True discrimination, misclassification rate.
Correct recognition rata is related with the attribute of itself to human Facial Image Recognition Algorithm.Misclassification rate is then with front-end collection image data
Irregular and backstage comparison library the increase of quality, misclassification rate can also increase accordingly, however, there is no effective at present
Solution.
Summary of the invention
The technical issues of embodiment of the present invention solves is how to reduce the misclassification rate of Identification of Images.
In order to solve the above technical problems, it includes: using the first portrait that the embodiment of the present invention, which provides a kind of portrait identification method,
Recognizer carries out first time identification to images to be recognized, and is compared with target person image, obtains meeting similarity threshold
The matching target image of value;The top N matching target image that last time is identified is as the images to be recognized of identification next time, directly
It completes to using m human Facial Image Recognition Algorithm to the m times identification of images to be recognized progress, and is compared with the target person image
It is right, the matching target image of the similarity threshold will be met as final matching target image, N is positive integer, m be greater than
1 positive integer.
Optionally, before carrying out first time identification to images to be recognized using the first human Facial Image Recognition Algorithm, further includes: root
According to the corresponding correct recognition rata of the similarity threshold of alternative human Facial Image Recognition Algorithm and misclassification rate, determine that first Identification of Images is calculated
Method is to the corresponding algorithm types of m human Facial Image Recognition Algorithm.
Optionally, the corresponding correct recognition rata of similarity threshold and misclassification rate of the alternative human Facial Image Recognition Algorithm of the basis,
Determine first human Facial Image Recognition Algorithm to the corresponding algorithm types of m human Facial Image Recognition Algorithm, comprising: to know using each alternative portrait
Other algorithm identifies that statistics obtains under each similarity threshold, corresponding correct recognition rata to sample image library respectively
And misclassification rate;In the correct recognition rata for meeting setting, and under same misclassification rate, according to the similarity area of each human Facial Image Recognition Algorithm
Between intersection accounting, determine first human Facial Image Recognition Algorithm to m from the alternative human Facial Image Recognition Algorithm for meeting default accounting
The corresponding algorithm types of human Facial Image Recognition Algorithm.
Optionally, the similarity threshold is determined in the following way: the alternative Identification of Images for meeting default accounting is calculated
The average value of the corresponding similarity threshold of method is as the similarity threshold.
Optionally, m value is 2.
Optionally, first human Facial Image Recognition Algorithm to the m human Facial Image Recognition Algorithm includes any of the following algorithm:
Based on KL transformation algorithm, it is based on integral image characteristic method, based on singularity characteristics method, probabilistic model method and deep neural network model
Method.
The embodiment of the present invention also provides a kind of Identification of Images device, comprising: the first recognition unit is suitable for using the first portrait
Recognizer carries out first time identification to images to be recognized, and is compared with target person image, obtains meeting similarity threshold
The matching target image of value;Second recognition unit, the top N matching target image suitable for identifying last time are known as next time
Other images to be recognized, until using m human Facial Image Recognition Algorithm complete to images to be recognized carry out the m time identify, and with it is described
Target person image is compared, and will meet the matching target image of the similarity threshold as final matching target figure
Picture, m are the positive integer greater than 1.
Optionally, the Identification of Images device further include: human Facial Image Recognition Algorithm confirmation unit is suitable in first identification
Before unit carries out first time identification to images to be recognized using the first human Facial Image Recognition Algorithm, according to alternative human Facial Image Recognition Algorithm
The corresponding correct recognition rata of similarity threshold and misclassification rate determine first human Facial Image Recognition Algorithm to m human Facial Image Recognition Algorithm
Corresponding algorithm types.
Optionally, the human Facial Image Recognition Algorithm confirmation unit is suitable for using each alternative human Facial Image Recognition Algorithm respectively to sample
Image library is identified that statistics obtains under each similarity threshold, corresponding correct recognition rata and misclassification rate;It is set in satisfaction
Fixed correct recognition rata, and under same misclassification rate, according to the intersection accounting in the similarity section of each human Facial Image Recognition Algorithm, from completely
Determine that first human Facial Image Recognition Algorithm is corresponding to m human Facial Image Recognition Algorithm in the alternative human Facial Image Recognition Algorithm of the default accounting of foot
Algorithm types.
Optionally, the human Facial Image Recognition Algorithm confirmation unit, suitable for determining the similarity threshold in the following way: will
Meet the average value of the corresponding similarity threshold of alternative human Facial Image Recognition Algorithm of default accounting as the similarity threshold.
Optionally, m value is 2.
Optionally, first human Facial Image Recognition Algorithm to the m human Facial Image Recognition Algorithm includes any of the following algorithm:
Based on KL transformation algorithm, it is based on integral image characteristic method, based on singularity characteristics method, probabilistic model method and deep neural network model
Method.
Compared with prior art, the technical solution of the embodiment of the present invention has the advantages that
After carrying out the first identification to images to be recognized using the first human Facial Image Recognition Algorithm, before last time is identified
Images to be recognized of the N matching target images as identification next time, until the m times identification is completed, it is multiple by being carried out to image
Identification screening, so as to reduce the misclassification rate of Identification of Images.
Detailed description of the invention
Fig. 1 is a kind of flow chart of portrait identification method in the embodiment of the present invention;
Fig. 2 is a kind of structural schematic diagram of Identification of Images device in the embodiment of the present invention.
Specific embodiment
As noted previously, as irregular and backstage comparison library the increase of front-end collection image data quality, accidentally knows
Rate can also increase accordingly, however, there is no effective solution schemes at present.
It, will after carrying out the first identification to images to be recognized using the first human Facial Image Recognition Algorithm in the embodiment of the present invention
Images to be recognized of the top N matching target image that last time identifies as identification next time is led to until completing the m times identification
It crosses and repeatedly identification screening is carried out to image, so as to reduce the misclassification rate of human Facial Image Recognition Algorithm.
It is understandable to enable the above-mentioned purpose, feature and beneficial effect of the embodiment of the present invention to become apparent, below with reference to attached
Figure is described in detail specific embodiments of the present invention.
Referring to Fig.1, a kind of flow chart of portrait identification method in the embodiment of the present invention provided, below by specific steps
It is described in detail.
Step 11, using the first human Facial Image Recognition Algorithm to images to be recognized carry out first time identification, and with target person figure
As being compared, the matching target image for meeting similarity threshold is obtained.
In specific implementation, images to be recognized can be still image, or dynamic image.For example, can be to deposit
The great amount of images stored in storage device, or the image that image collecting device such as camera etc. acquires in real time.
After getting images to be recognized, images to be recognized can be carried out for the first time using the first human Facial Image Recognition Algorithm
Identification, and be compared with target person image, obtain the matching target image for meeting similarity threshold.
Step 12, the images to be recognized that top N matching target image last time identified is identified as next time, until
It completes to carry out images to be recognized the m times identification using m human Facial Image Recognition Algorithm, and is compared with the target person image
It is right, the matching target image of similarity threshold will be met as final matching target image.
In specific implementation, the matching target image that the first Identification of Images obtains is carried out from using the first human Facial Image Recognition Algorithm
In extract top N matching target image.Using the top N extracted matching target image as second Identification of Images
Images to be recognized, use the second human Facial Image Recognition Algorithm to the preceding N extracted for matching target image identify, and with it is described
Target person image is compared, and obtains the matching target image for meeting the similarity threshold.
The matching target image obtained from last time Identification of Images extracts the matching target image of corresponding number as next people
As the images to be recognized of identification, until complete to carry out images to be recognized the m times identification using m human Facial Image Recognition Algorithm, and with
The target person image is compared, and will meet the matching target image of the similarity threshold as final matching target
Image.Wherein, N is positive integer, and m is the positive integer greater than 1.
It, will by above scheme it is found that after carrying out the first identification to images to be recognized using the first human Facial Image Recognition Algorithm
Images to be recognized of the top N matching target image that last time identifies as identification next time is led to until completing the m times identification
It crosses and repeatedly identification screening is carried out to image, so as to reduce the misclassification rate of Identification of Images.
In an embodiment of the present invention, m value is 2, namely 2 Identification of Images are carried out during Identification of Images, in this way
Not only the misclassification rate of Identification of Images can have been reduced, but also the algorithm complexity of Identification of Images can be taken into account.
It in specific implementation, can before carrying out first time identification to images to be recognized using the first human Facial Image Recognition Algorithm
To determine that first portrait is known according to the corresponding correct recognition rata of the similarity threshold of alternative human Facial Image Recognition Algorithm and misclassification rate
Other algorithm is to the corresponding algorithm types of m human Facial Image Recognition Algorithm.
In specific implementation, when determining first human Facial Image Recognition Algorithm to the corresponding class of algorithms of m human Facial Image Recognition Algorithm
After type, Identification of Images can be carried out using identified human Facial Image Recognition Algorithm during subsequent Identification of Images, and need not held
Row first human Facial Image Recognition Algorithm to the corresponding algorithm types of m human Facial Image Recognition Algorithm verification step.It is understood that
It in practical applications, according to actual needs can also be again to first human Facial Image Recognition Algorithm to m human Facial Image Recognition Algorithm pair
The algorithm types answered are updated.
Specifically, human Facial Image Recognition Algorithm used by during Identification of Images can be determined as follows:
Sample image library is identified respectively using each alternative human Facial Image Recognition Algorithm, statistics is obtained in each similarity threshold
Under, corresponding correct recognition rata and misclassification rate;
In the correct recognition rata for meeting setting, and under same misclassification rate, according to the similarity area of each human Facial Image Recognition Algorithm
Between intersection accounting, determine first human Facial Image Recognition Algorithm to m from the alternative human Facial Image Recognition Algorithm for meeting default accounting
The corresponding algorithm types of human Facial Image Recognition Algorithm.
In specific implementation, determining first human Facial Image Recognition Algorithm to the corresponding class of algorithms of m human Facial Image Recognition Algorithm
After type, it can will meet the average value of the corresponding similarity threshold of alternative human Facial Image Recognition Algorithm of default accounting as described similar
Spend threshold value.Also it is used using the average value of the corresponding similarity threshold of selected human Facial Image Recognition Algorithm as during Identification of Images
Similarity threshold.
It in specific implementation, can be according to the size of default accounting from the corresponding algorithm of high to low determining human Facial Image Recognition Algorithm
Type.
First human Facial Image Recognition Algorithm can be to convert algorithm based on KL, based on product to the m human Facial Image Recognition Algorithm
Partial image characteristic method, based on any one in singularity characteristics method, probabilistic model method and deep neural network model method etc..
In specific implementation, first human Facial Image Recognition Algorithm, the i-th human Facial Image Recognition Algorithm and m human Facial Image Recognition Algorithm can
With identical, can also be different, 1 < i < m.
For example, using two kinds of human Facial Image Recognition Algorithms during Identification of Images, the first human Facial Image Recognition Algorithm and the second portrait are known
Other algorithm.First human Facial Image Recognition Algorithm can be deep neural network model method, and the second human Facial Image Recognition Algorithm can be depth mind
Through network Model Method.First human Facial Image Recognition Algorithm may be based on integral image characteristic method, and the second human Facial Image Recognition Algorithm can be with
For based on singularity characteristics method.It is understood that in practical applications, the first human Facial Image Recognition Algorithm and the second human Facial Image Recognition Algorithm
There may also be other combinations.
It, can be in conjunction with the advantage and disadvantage of algorithms of different, by to be identified by the way of the fusion of a variety of human Facial Image Recognition Algorithms
Image carries out examination filtering, can substantially reduce the misclassification rate of image on the basis of not reducing correct recognition rata.
Better understand and realize that the embodiment of the present invention, the embodiment of the present invention also provide for the ease of those skilled in the art
A kind of Identification of Images device.
Referring to Fig. 2, the Identification of Images device 20 may include: the first recognition unit 21 and the second recognition unit 22,
In,
First recognition unit 21 is suitable for carrying out first time knowledge to images to be recognized using the first human Facial Image Recognition Algorithm
Not, and with target person image it is compared, obtains the matching target image for meeting similarity threshold;
Second recognition unit 22, the top N matching target image suitable for identifying last time are identified as next time
Images to be recognized, until using m human Facial Image Recognition Algorithm complete to images to be recognized carry out the m time identify, and with the mesh
Mark character image is compared, and will meet the matching target image of the similarity threshold as final matching target image, m
For the positive integer greater than 1.
In specific implementation, the Identification of Images device 20 can also include: that (Fig. 2 is not for human Facial Image Recognition Algorithm confirmation unit
It shows).The human Facial Image Recognition Algorithm confirmation unit is suitable for using the first human Facial Image Recognition Algorithm pair in first recognition unit 21
Before images to be recognized carries out first time identification, according to the corresponding correct recognition rata of the similarity threshold of alternative human Facial Image Recognition Algorithm
And misclassification rate, determine first human Facial Image Recognition Algorithm to the corresponding algorithm types of m human Facial Image Recognition Algorithm.
In specific implementation, the human Facial Image Recognition Algorithm confirmation unit is suitable for using each alternative human Facial Image Recognition Algorithm difference
Sample image library is identified, statistics obtains under each similarity threshold, corresponding correct recognition rata and misclassification rate;?
Meet the correct recognition rata of setting, and under same misclassification rate, is accounted for according to the intersection in the similarity section of each human Facial Image Recognition Algorithm
Than determining that first human Facial Image Recognition Algorithm to m Identification of Images is calculated from the alternative human Facial Image Recognition Algorithm for meeting default accounting
The corresponding algorithm types of method.
In specific implementation, the human Facial Image Recognition Algorithm confirmation unit, suitable for determining the similarity in the following way
Threshold value: the average value of the corresponding similarity threshold of alternative human Facial Image Recognition Algorithm of default accounting will be met as the similarity threshold
Value.
In an embodiment of the present invention, m value is 2.
In specific implementation, first human Facial Image Recognition Algorithm to the m human Facial Image Recognition Algorithm includes following any one
Kind algorithm: based on KL transformation algorithm, it is based on integral image characteristic method, neural based on singularity characteristics method, probabilistic model method and depth
Network Model Method.
In specific implementation, the working principle and workflow of the Identification of Images device can be with reference to the above-mentioned realities of the present invention
The description in the portrait identification method of example offer is applied, details are not described herein again.
The embodiment of the present invention also provides another Identification of Images device, including memory and processor, on the memory
It is stored with the computer instruction that can be run on the processor, the processor executes sheet when running the computer instruction
The step of inventing the portrait identification method that any of the above-described embodiment provides.
The embodiment of the present invention also provides a kind of computer readable storage medium, and computer readable storage medium is non-volatile
Storage medium or non-transitory storage media, are stored thereon with computer instruction, and the computer instruction executes the present invention when running
The step of portrait identification method that any of the above-described embodiment provides.
Those of ordinary skill in the art will appreciate that all or part of the steps in the various methods of above-described embodiment is can
It is completed with instructing relevant hardware by program, which can be stored in a computer readable storage medium, storage
Medium may include: ROM, RAM, disk or CD etc..
Although present disclosure is as above, present invention is not limited to this.Anyone skilled in the art are not departing from this
It in the spirit and scope of invention, can make various changes or modifications, therefore protection scope of the present invention should be with claim institute
Subject to the range of restriction.
Claims (14)
1. a kind of portrait identification method characterized by comprising
First time identification is carried out to images to be recognized using the first human Facial Image Recognition Algorithm, and is compared with target person image,
Obtain the matching target image for meeting similarity threshold;
The top N matching target image that last time is identified is as the images to be recognized of identification next time, until using m portrait
Recognizer is completed to carry out images to be recognized the m times identification, and is compared with the target person image, will be described in satisfaction
For the matching target image of similarity threshold as final matching target image, N is positive integer, and m is the positive integer greater than 1.
2. portrait identification method according to claim 1, which is characterized in that treating knowledge using the first human Facial Image Recognition Algorithm
Other image carries out before first time identification, further includes: according to the corresponding correct knowledge of the similarity threshold of alternative human Facial Image Recognition Algorithm
Not rate and misclassification rate determine first human Facial Image Recognition Algorithm to the corresponding algorithm types of m human Facial Image Recognition Algorithm.
3. portrait identification method according to claim 2, which is characterized in that the phase of the alternative human Facial Image Recognition Algorithm of basis
Like the corresponding correct recognition rata of degree threshold value and misclassification rate, determine first human Facial Image Recognition Algorithm to m human Facial Image Recognition Algorithm pair
The algorithm types answered, comprising:
Sample image library is identified respectively using each alternative human Facial Image Recognition Algorithm, statistics obtains under each similarity threshold,
Corresponding correct recognition rata and misclassification rate;
In the correct recognition rata for meeting setting, and under same misclassification rate, according to the similarity section of each human Facial Image Recognition Algorithm
Intersection accounting determines first human Facial Image Recognition Algorithm to m portrait from the alternative human Facial Image Recognition Algorithm for meeting default accounting
The corresponding algorithm types of recognizer.
4. portrait identification method according to claim 3, which is characterized in that determine the similarity threshold in the following way
Value: the average value of the corresponding similarity threshold of alternative human Facial Image Recognition Algorithm of default accounting will be met as the similarity threshold
Value.
5. portrait identification method according to claim 1, which is characterized in that m value is 2.
6. portrait identification method according to claim 1, which is characterized in that first human Facial Image Recognition Algorithm to described
M human Facial Image Recognition Algorithm includes any of the following algorithm:
Based on KL transformation algorithm, it is based on integral image characteristic method, based on singularity characteristics method, probabilistic model method and deep neural network
Modelling.
7. a kind of Identification of Images device characterized by comprising
First recognition unit is suitable for carrying out first time identification, and and target to images to be recognized using the first human Facial Image Recognition Algorithm
Character image is compared, and obtains the matching target image for meeting similarity threshold;
Second recognition unit, the figure to be identified that the top N matching target image suitable for identifying last time is identified as next time
Picture, until using m human Facial Image Recognition Algorithm complete to images to be recognized carry out the m time identify, and with the target person image
It is compared, the matching target image of the similarity threshold will be met as final matching target image, m is greater than 1
Positive integer.
8. Identification of Images device according to claim 7, which is characterized in that further include: human Facial Image Recognition Algorithm confirmation unit,
Suitable for first recognition unit using the first human Facial Image Recognition Algorithm to images to be recognized carry out first time identification before, according to
The corresponding correct recognition rata of similarity threshold and misclassification rate of alternative human Facial Image Recognition Algorithm, determine first human Facial Image Recognition Algorithm
To the corresponding algorithm types of m human Facial Image Recognition Algorithm.
9. Identification of Images device according to claim 8, which is characterized in that the human Facial Image Recognition Algorithm confirmation unit is fitted
Sample image library is identified respectively in using each alternative human Facial Image Recognition Algorithm, statistics obtains under each similarity threshold, point
Not corresponding correct recognition rata and misclassification rate;In the correct recognition rata for meeting setting, and under same misclassification rate, according to each portrait
The intersection accounting in the similarity section of recognizer determines described first from the alternative human Facial Image Recognition Algorithm for meeting default accounting
Human Facial Image Recognition Algorithm is to the corresponding algorithm types of m human Facial Image Recognition Algorithm.
10. Identification of Images device according to claim 9, which is characterized in that the human Facial Image Recognition Algorithm confirmation unit is fitted
In determining the similarity threshold in the following way: the corresponding similarity of alternative human Facial Image Recognition Algorithm of default accounting will be met
The average value of threshold value is as the similarity threshold.
11. Identification of Images device according to claim 7, which is characterized in that m value is 2.
12. Identification of Images device according to claim 7, which is characterized in that first human Facial Image Recognition Algorithm is to described
M human Facial Image Recognition Algorithm includes any of the following algorithm:
Based on KL transformation algorithm, it is based on integral image characteristic method, based on singularity characteristics method, probabilistic model method and deep neural network
Modelling.
13. a kind of Identification of Images device, including memory and processor, being stored on the memory can be in the processor
The computer instruction of upper operation, which is characterized in that perform claim requires 1 to 6 when the processor runs the computer instruction
The step of described in any item portrait identification methods.
14. a kind of computer readable storage medium, is stored thereon with computer instruction, which is characterized in that the computer instruction
Perform claim requires the step of 1 to 6 described in any item portrait identification methods when operation.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810003470.1A CN109993039A (en) | 2018-01-02 | 2018-01-02 | Portrait identification method and device, computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810003470.1A CN109993039A (en) | 2018-01-02 | 2018-01-02 | Portrait identification method and device, computer readable storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109993039A true CN109993039A (en) | 2019-07-09 |
Family
ID=67128870
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810003470.1A Pending CN109993039A (en) | 2018-01-02 | 2018-01-02 | Portrait identification method and device, computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109993039A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112286780A (en) * | 2019-07-23 | 2021-01-29 | 浙江宇视科技有限公司 | Method, device and equipment for testing recognition algorithm and storage medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012003623A (en) * | 2010-06-18 | 2012-01-05 | Toshiba Corp | Face image search device and face image search method |
CN102982165A (en) * | 2012-12-10 | 2013-03-20 | 南京大学 | Large-scale human face image searching method |
CN103745240A (en) * | 2013-12-20 | 2014-04-23 | 许雪梅 | Method and system for retrieving human face on the basis of Haar classifier and ORB characteristics |
US20140270370A1 (en) * | 2013-03-18 | 2014-09-18 | Kabushiki Kaisha Toshiba | Person recognition apparatus and person recognition method |
CN105184238A (en) * | 2015-08-26 | 2015-12-23 | 广西小草信息产业有限责任公司 | Human face recognition method and system |
CN107038400A (en) * | 2016-02-04 | 2017-08-11 | 索尼公司 | Face identification device and method and utilize its target person tracks of device and method |
CN107316315A (en) * | 2017-05-04 | 2017-11-03 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of object recognition and detection method based on template matches |
-
2018
- 2018-01-02 CN CN201810003470.1A patent/CN109993039A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012003623A (en) * | 2010-06-18 | 2012-01-05 | Toshiba Corp | Face image search device and face image search method |
CN102982165A (en) * | 2012-12-10 | 2013-03-20 | 南京大学 | Large-scale human face image searching method |
US20140270370A1 (en) * | 2013-03-18 | 2014-09-18 | Kabushiki Kaisha Toshiba | Person recognition apparatus and person recognition method |
CN103745240A (en) * | 2013-12-20 | 2014-04-23 | 许雪梅 | Method and system for retrieving human face on the basis of Haar classifier and ORB characteristics |
CN105184238A (en) * | 2015-08-26 | 2015-12-23 | 广西小草信息产业有限责任公司 | Human face recognition method and system |
CN107038400A (en) * | 2016-02-04 | 2017-08-11 | 索尼公司 | Face identification device and method and utilize its target person tracks of device and method |
CN107316315A (en) * | 2017-05-04 | 2017-11-03 | 佛山市南海区广工大数控装备协同创新研究院 | A kind of object recognition and detection method based on template matches |
Non-Patent Citations (1)
Title |
---|
崇玉海: "基于二维图像表示的人脸检测与识别技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112286780A (en) * | 2019-07-23 | 2021-01-29 | 浙江宇视科技有限公司 | Method, device and equipment for testing recognition algorithm and storage medium |
CN112286780B (en) * | 2019-07-23 | 2024-03-12 | 浙江宇视科技有限公司 | Method, device, equipment and storage medium for testing recognition algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110807385B (en) | Target detection method, target detection device, electronic equipment and storage medium | |
US10402627B2 (en) | Method and apparatus for determining identity identifier of face in face image, and terminal | |
WO2018188453A1 (en) | Method for determining human face area, storage medium, and computer device | |
Yan et al. | Learning the change for automatic image cropping | |
CN104143079B (en) | The method and system of face character identification | |
CN108288027A (en) | A kind of detection method of picture quality, device and equipment | |
CN109214298B (en) | Asian female color value scoring model method based on deep convolutional network | |
CN108805016B (en) | Head and shoulder area detection method and device | |
CN110135500A (en) | Method for tracking target under a kind of more scenes based on adaptive depth characteristic filter | |
CN106446754A (en) | Image identification method, metric learning method, image source identification method and devices | |
CN103473564B (en) | A kind of obverse face detection method based on sensitizing range | |
CN104102920A (en) | Pest image classification method and pest image classification system based on morphological multi-feature fusion | |
CN103810490A (en) | Method and device for confirming attribute of face image | |
CN113449704B (en) | Face recognition model training method and device, electronic equipment and storage medium | |
CN110120065A (en) | A kind of method for tracking target and system based on layering convolution feature and dimension self-adaption core correlation filtering | |
CN109753884A (en) | A kind of video behavior recognition methods based on key-frame extraction | |
CN107944403A (en) | Pedestrian's attribute detection method and device in a kind of image | |
WO2023000253A1 (en) | Climbing behavior early-warning method and apparatus, electrode device, and storage medium | |
CN107679469A (en) | A kind of non-maxima suppression method based on deep learning | |
CN112183356A (en) | Driving behavior detection method and device and readable storage medium | |
CN108734127A (en) | Age identifies value adjustment method, device, equipment and storage medium | |
CN109410138A (en) | Modify jowled methods, devices and systems | |
WO2015176502A1 (en) | Image feature estimation method and device | |
CN108875506A (en) | Face shape point-tracking method, device and system and storage medium | |
CN111968087B (en) | Plant disease area detection method |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190709 |
|
WD01 | Invention patent application deemed withdrawn after publication |