CN110363107A - Face forehead point Quick Extended method, apparatus, storage medium and processor - Google Patents
Face forehead point Quick Extended method, apparatus, storage medium and processor Download PDFInfo
- Publication number
- CN110363107A CN110363107A CN201910562255.XA CN201910562255A CN110363107A CN 110363107 A CN110363107 A CN 110363107A CN 201910562255 A CN201910562255 A CN 201910562255A CN 110363107 A CN110363107 A CN 110363107A
- Authority
- CN
- China
- Prior art keywords
- face
- point
- forehead
- direction vector
- horizontal direction
- 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
- 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
-
- 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)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to technical field of face recognition, it discloses and a kind of face forehead point Quick Extended method is provided, device, storage medium and processor, wherein, the face point that this method is got by recognition of face procedure identification, then face Height Standard value is calculated by the face got the point, horizontal direction vector sum vertical direction vector, finally the symmetrical face point of forehead is indicated in the face point at the forehead center for generating expression respectively and in the central point two sides, the method obtained compared to existing face forehead key point or characteristic point, the present invention is not only quick and easy, but also it is not take up additional hardware resource.
Description
Technical field
The present invention relates to face recognition technologies, more particularly to a kind of face forehead point Quick Extended method, apparatus, storage
Deposit medium and processor.
Background technique
Face recognition technology is quite mature at present, therefore the software tool development kit based on recognition of face
(Software Development Kit, i.e. SDK) is also more universal.Therefore, software engineer also starts according to based on people
The software tool development kit of face identification also starts to be gradually added into face identification functions in various application programs.But technical staff
It was found that do not include the point at forehead position in the characteristic point that the software tool development kit offer of existing recognition of face recognizes, but
The feature point of forehead position is extremely important (such as camera software) in some application software.
However in the prior art, there is no the technologies that forehead characteristic point is directly obtained according to software tool development kit, and
It is the identification that forehead characteristic point is realized using the forehead identification model or algorithm of stand-alone development.For example, application No. is
2018114938557 prior art discloses a kind of recognition methods of face forehead key point, which is exactly to adopt
Sample is formed with being labeled to facial image, network training is then carried out and obtains neural network model, it can be according to the mind
The forehead key point in face is identified through network model.Although the prior art may be implemented crucial to face forehead
The identification of point, but its implementation is excessively complicated, and when operation needs to occupy hardware resource larger.
So those skilled in the art wish that carrying out rapidly extracting on the basis of existing software tool development kit recognizes people
The characteristic point of face forehead, and accomplish to be not take up hardware resource.
Summary of the invention
Technical purpose
For solve how additionally occupy hardware resource on the basis of quick obtaining to expression face forehead feature people
The technical issues of face point.
Technical solution
First aspect
There is provided a kind of face forehead point Quick Extended method based on recognition of face, which is characterized in that the face volume
Head point Quick Extended method the following steps are included: S101, using face recognizer identify current face's image with two
The face point that dimensional feature point indicates；S102 is calculated separately according to the face point and is obtained the high scale of face in current face's image
Quasi- value, horizontal direction vector sum vertical direction vector, wherein the horizontal direction vector be people's face horizontal direction from left to right
Direction vector, the vertical direction vector are the direction vector of face vertical direction from bottom to top；S103 is getting Vertical Square
The one first forehead center indicated with reference to the position generation one of the calibrated altitude of multiple with two dimensional character point on the direction of vector
Point；S104, the position of the horizontal direction vector of one second reference multiple on getting the forehead central point left and right horizontal direction
Set the symmetrical forehead point that each generation one is indicated with two dimensional character point.
In some embodiments, the calculating acquisition methods of the face Height Standard value include: the face that will identify that
Topmost two dimensional character point progress difference operation corresponding with face point where the both ends of nose bottom obtains middle expression nose.
In some embodiments, the calculating acquisition methods of the horizontal direction vector include: in the face that will identify that
The corresponding two dimensional character of face point where indicating canthus clicks through row vector operation and obtains.
In some embodiments, the calculating acquisition methods of the vertical direction vector include: in the face that will identify that
Indicate that the corresponding two dimensional character click-through row vector operation of face point where both ends obtains under bridge of the nose upper end and bridge of the nose lower end.
Second aspect
A kind of face forehead point Quick Extended device based on recognition of face, including point identification module are provided, are used for
The face point of current face's image indicated with two dimensional character point is identified using face recognizer；Computing module, according to institute
It states face point and calculates separately face Height Standard value, horizontal direction vector sum vertical direction vector in acquisition current face's image,
Wherein, the horizontal direction vector is the direction vector of people's face horizontal direction from left to right, and the vertical direction vector is face
The direction vector of vertical direction from bottom to top；First point expansion module, in the direction for getting vertical direction vector
The one first forehead central point indicated with reference to the position generation one of the calibrated altitude of multiple with two dimensional character point；Second point Bits Expanding
Module, the position for the horizontal direction vector of one second reference multiple on getting the forehead central point left and right horizontal direction
Set the symmetrical forehead point that each generation one is indicated with two dimensional character point.
In some embodiments, the calculating acquisition methods of the face Height Standard value include: the face that will identify that
The corresponding two dimensional character point of face point where middle expression nose most upper and lower ends carries out difference operation and obtains.
In some embodiments, the calculating acquisition methods of the horizontal direction vector include: in the face that will identify that
The corresponding two dimensional character of face point where indicating canthus clicks through row vector operation and obtains.
In some embodiments, the calculating acquisition methods of the vertical direction vector include: in the face that will identify that
The corresponding two dimensional character of face point where indicating bridge of the nose upper and lower ends clicks through row vector operation and obtains.
The third aspect
A kind of storage medium is provided, the storage medium includes the program of storage, and described program executes first party when running
Face forehead point Quick Extended method described in any one of face.
Fourth aspect
A kind of processor is provided, the processor executes in first aspect when described program is run and appoints for running program
Face forehead point Quick Extended method described in one.
Technical effect
Simple calculations can be passed through on the basis of the face point that existing recognition of face SDK is got through the invention
It can quickly obtain indicating the human face characteristic point of face forehead indicated with two dimensional character point, it is crucial compared to existing face forehead
The method that point or characteristic point obtain, the present invention is not only quick and easy, but also is not take up additional hardware resource.
Detailed description of the invention
Fig. 1 is the process the present invention is based on the face forehead point Quick Extended method of recognition of face in one embodiment
Figure.
Fig. 2 is the face point got according to existing recognition of face SDK and obtains the effect of forehead inflexion point according to the present invention
Fruit figure.
Principle framework of the Fig. 3 the present invention is based on the face forehead point Quick Extended device of recognition of face in one example
Figure.
Drawing reference numeral explanation
200 face forehead point Quick Extended devices
210 point identification modules
220 computing modules
230 first point expansion modules
240 second point Bits Expanding modules
S101-S104 method and step
Specific embodiment
Embodiment 1
See Fig. 1, shows that the present invention is based on the face forehead point Quick Extended methods of recognition of face in one embodiment
Flow chart, as shown, the face forehead point Quick Extended method the following steps are included:
Step S101 identifies the face of current face's image indicated with two dimensional character point using face recognizer
Point；
Step S102 is calculated separately according to the face point and is obtained face Height Standard value, level in current face's image
Direction vector and vertical direction vector, wherein the horizontal direction vector is the direction vector of people's face horizontal direction from left to right,
The vertical direction vector is the direction vector of face vertical direction from bottom to top；
Step S103, the position of the calibrated altitude of one first reference multiple is raw on the direction for getting vertical direction vector
The forehead central point indicated at one with two dimensional character point；
Step S104, the horizontal direction of one second reference multiple on getting the forehead central point left and right horizontal direction
The position of vector respectively generates a symmetrical forehead point indicated with two dimensional character point.
It can be on the basis of the face point that existing recognition of face SDK is got, by simply counting by the above method
The human face characteristic point indicated with two dimensional character point for calculating and can quickly obtaining indicating face forehead is learned, compared to existing face volume
The method that head key point or characteristic point obtain, the present invention is not only quick and easy, but also is not take up additional hardware resource.
Specifically, face Height Standard value refers in face that face recognition software is identified in above-mentioned steps S102
The distance between nose the top and nose bottom value, calculating acquisition methods, which can be, indicates nose in the face that will identify that
Sub two dimensional character point corresponding with face point where nose bottom topmost carries out difference operation and obtains.It should be appreciated that face
The reference standard of Height Standard value is not limited in a kind of this method, can also specifically be set according to the face point got
Meter, such as difference operation can be carried out according to the upper end of the bridge of the nose and the lower end of the bridge of the nose and obtained.As long as calculated distance institute
Line segment where at a distance from extended line and the calculation in the present embodiment of line segment obtain, which is overlapped, or principle is identical all should
It is considered as equivalent calculating means.
Specifically, the calculating acquisition methods of horizontal direction vector include: in the face that will identify that in above-mentioned steps S102
The corresponding two dimensional character of face point where indicating canthus clicks through row vector operation and obtains.Wherein, which can be left in face
The canthus of eye and right eye can also be left eye canthus profile and right eye canthus profile in face.
Specifically, the calculating acquisition methods of vertical direction vector include: in the face that will identify that in above-mentioned steps S102
The corresponding two dimensional character of face point where indicating bridge of the nose upper end and bridge of the nose lower end both ends clicks through row vector operation and obtains.
Specifically, described first includes constant 2.7 with reference to multiple, the second reference multiple includes constant 0.7.Wherein,
In the present invention first with reference to multiple and second with reference to multiple and not exclusive, it is disclosed herein with reference to being based on technical staff's root
The relatively reasonable value obtained according to a large amount of experiment.Certainly in specific implementation, it is similar often that other also can be used
Number as this refer to multiple, but its only solve the technology of the present invention purpose in technical problem optimization technique means and not must
Want technological means.
For above-mentioned technical proposal can be more clearly understood convenient for those skilled in the art, below with an existing face
For identifying SDK, identification can be carried out to facial image and gets 106 two dimensional character points, is detailed in Fig. 2, wherein each two
Dimensional feature point includes the coordinate value indicated with x on two-dimensional coordinate and y, and the two dimensional character point of following presentation face point is with Px table
Show, x is the sequence in the 106 two dimensional character points got.
It wherein, can be according to these since the two dimensional character point got is all the coordinate points for indicating face characteristic
The two dimensional character point of acquisition performs mathematical calculations.
Specifically, the calculating about face Height Standard value, here by between nose bottom and nose the top away from
From as face Height Standard value, in conjunction with Fig. 2 it is found that indicating that nose is most upper in the face point got according to recognition of face SDK
The face point at end is P43, and expression nose the lowermost face point is P49, then calculating P43 and P49 according to distance calculation formula
The distance between two face points obtain face Height Standard value.
Specifically, the calculating about horizontal direction vector, in conjunction with Fig. 2 it is found that the face got according to recognition of face SDK
The face point that right eye inner eye corner profile is indicated in point is P78, indicates that the face point of left eye inner eye corner profile is P79, then by people
Face point P79 subtracts face point P78 and gets the direction vector of current face's horizontal direction from left to right.It should be appreciated that level side
Be not limited in preceding method to the calculation of vector, for example, can also be the P58 and P55 that respectively indicate left and right canthus come into
Row vector is calculated.
Specifically, the calculating about vertical direction vector, in conjunction with Fig. 2 it is found that the face got according to recognition of face SDK
The face point that bridge of the nose the top is indicated in point is P43, and expression bridge of the nose the lowermost face point is P46, then can be by face point
P46 subtracts face point P43 and gets the direction vector of current face's vertical direction from bottom to top.
It is after obtaining above-mentioned face Height Standard value, horizontal direction vector sum vertical direction vector, then with P46 point
Point obtains a new people in the vertical direction vector direction got at the absolute distance of one 2.3 times of vertical direction vectors
Face point is indicated using the face point as the central point of forehead, such as Fig. 2 with P108.
Then, then by the obtained P108 horizontal direction vector for subtracting 0.7 times obtain a face point, i.e., left forehead point, such as
Fig. 2 is indicated with P107.
Finally, obtained P108 is obtained into a face point plus 0.7 times of horizontal direction vector again, i.e., right forehead point, such as
Fig. 2 is indicated with PP106.It so both can be on having got the forehead central point left and right horizontal direction one second with reference to multiple
The position of horizontal direction vector respectively generates a symmetrical forehead point indicated with two dimensional character point.
It can be clearly understood that by above-mentioned example, be led on the basis of the face point got according to recognition of face SDK
It crosses and calculates quick obtaining to the face point for indicating forehead feature.
Embodiment 2
See Fig. 3, shows the present invention is based on the face forehead point Quick Extended device of recognition of face in one example
Principle framework figure, as shown, the device 200 includes point identification module 210, computing module 220, the first point expansion module
230 and second point Bits Expanding module 240, wherein point identification module 210 works as forefathers for identifying using face recognizer
The face point of face image indicated with two dimensional character point；Computing module 220 calculates separately acquisition according to the face point and works as forefathers
Face Height Standard value, horizontal direction vector sum vertical direction vector in face image, wherein the horizontal direction vector is face
The direction vector of horizontal direction from left to right, the vertical direction vector are the direction vector of face vertical direction from bottom to top；
First point expansion module 230 is for the calibrated altitude of one first reference multiple on the direction for getting vertical direction vector
Position generates a forehead central point indicated with two dimensional character point；Second point Bits Expanding module 240 is for getting the volume
The position of the horizontal direction vector of one second reference multiple generates one respectively with two dimensional character point on head central point left and right horizontal direction
The symmetrical forehead point indicated.
Specifically, in some embodiments, the calculating acquisition methods of the face Height Standard value include: that will identify that
Indicate that topmost two dimensional character point progress difference operation corresponding with face point where the both ends of nose bottom obtains nose in face
It arrives.
Specifically, in some embodiments, the calculating acquisition methods of the horizontal direction vector include: the people that will identify that
The corresponding two dimensional character of face point where indicating canthus in face clicks through row vector operation and obtains.
Specifically, in some embodiments, the calculating acquisition methods of the vertical direction vector include: the people that will identify that
The corresponding two dimensional character of face point where indicating bridge of the nose upper end and bridge of the nose lower end both ends in face clicks through row vector operation and obtains.
Specifically, in some embodiments, the first reference multiple includes constant 2.7, and described second refers to multiple packet
Include constant 0.7.
It is to be appreciated that it is the standard software list obtained by computer program that each module in the present embodiment, which can have,
Member is also possible to the standard hardware unit realized by hardware, passes through these standard software units or standard hardware unit
It is of coupled connections and may be constructed the device for being able to solve technical problem in the technology of the present invention purpose.
Embodiment 3
The present embodiment additionally provides a kind of storage medium, and the storage medium includes the program of storage, described program operation
Any enforceable face forehead point Quick Extended method in Shi Zhihang embodiment 1.
In specific implementation, the storage medium can be random access memory, read-only memory or server etc.
Storage medium.
Embodiment 4
The present embodiment additionally provides a kind of processor, and the processor executes when running for running program, described program
Any enforceable face forehead point Quick Extended method in embodiment 1.
In specific implementation, the processor can be CPU, or may be the microcomputer systems such as single-chip microcontroller.
Claims (10)
1. a kind of face forehead point Quick Extended method based on recognition of face, which is characterized in that the face forehead point
Quick Extended method the following steps are included:
S101 identifies the face point of current face's image indicated with two dimensional character point using face recognizer；
S102 is calculated separately according to the face point and is obtained face Height Standard value, horizontal direction vector in current face's image
With vertical direction vector, wherein the horizontal direction vector is the direction vector of people's face horizontal direction from left to right, described vertical
Direction vector is the direction vector of face vertical direction from bottom to top；
S103, the position of the calibrated altitude of one first reference multiple generates one with two on the direction for getting vertical direction vector
The forehead central point that dimensional feature point indicates；
S104, the position of the horizontal direction vector of one second reference multiple on getting the forehead central point left and right horizontal direction
Set the symmetrical forehead point that each generation one is indicated with two dimensional character point.
2. the face forehead point Quick Extended method according to claim 1 based on recognition of face, which is characterized in that institute
The calculating acquisition methods for stating face Height Standard value include: that nose the top and nose bottom are indicated in the face that will identify that
The corresponding two dimensional character point of face point where both ends carries out difference operation and obtains.
3. the face forehead point Quick Extended method according to claim 1 based on recognition of face, which is characterized in that its
In, the calculating acquisition methods of the horizontal direction vector include: that face point where indicating canthus in the face that will identify that is corresponding
Two dimensional character click through row vector operation obtain.
4. the face forehead point Quick Extended method according to claim 1 based on recognition of face, which is characterized in that institute
The calculating acquisition methods for stating vertical direction vector include: that bridge of the nose upper end and bridge of the nose lower end both ends institute are indicated in the face that will identify that
Row vector operation is clicked through in the corresponding two dimensional character of face point to obtain.
5. a kind of face forehead point Quick Extended device based on recognition of face characterized by comprising
Point identification module, for identifying the people of current face's image indicated with two dimensional character point using face recognizer
Face point；
Computing module calculates separately according to the face point and obtains face Height Standard value, horizontal direction in current face's image
Vector sum vertical direction vector, wherein the horizontal direction vector is the direction vector of people's face horizontal direction from left to right, described
Vertical direction vector is the direction vector of face vertical direction from bottom to top；
First point expansion module, the calibrated altitude for the one first reference multiple on the direction for getting vertical direction vector
Position generate the forehead central point that indicates with two dimensional character point；
Second point Bits Expanding module, for the one second reference multiple on getting the forehead central point left and right horizontal direction
The position of horizontal direction vector respectively generates a symmetrical forehead point indicated with two dimensional character point.
6. the face forehead point Quick Extended device according to claim 5 based on recognition of face, which is characterized in that institute
The calculating acquisition methods for stating face Height Standard value include: that nose the top and nose bottom are indicated in the face that will identify that
The corresponding two dimensional character point of face point where both ends carries out difference operation and obtains.
7. the face forehead point Quick Extended installation method according to claim 5 based on recognition of face, which is characterized in that
Wherein, the calculating acquisition methods of the horizontal direction vector include: face point pair where indicating canthus in the face that will identify that
The two dimensional character answered clicks through row vector operation and obtains.
8. the face forehead point Quick Extended device according to claim 5 based on recognition of face, which is characterized in that institute
The calculating acquisition methods for stating vertical direction vector include: that both ends under bridge of the nose upper end and bridge of the nose lower end are indicated in the face that will identify that
Face point corresponding two dimensional character in place clicks through row vector operation and obtains.
9. a kind of storage medium, which is characterized in that the storage medium includes the program of storage, right of execution when described program is run
Benefit requires the described in any item face forehead point Quick Extended methods of 1-4.
10. a kind of processor, which is characterized in that the processor is for running program, and perform claim is wanted when described program is run
Seek the described in any item face forehead point Quick Extended methods of 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910562255.XA CN110363107A (en) | 2019-06-26 | 2019-06-26 | Face forehead point Quick Extended method, apparatus, storage medium and processor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910562255.XA CN110363107A (en) | 2019-06-26 | 2019-06-26 | Face forehead point Quick Extended method, apparatus, storage medium and processor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110363107A true CN110363107A (en) | 2019-10-22 |
Family
ID=68216946
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910562255.XA Pending CN110363107A (en) | 2019-06-26 | 2019-06-26 | Face forehead point Quick Extended method, apparatus, storage medium and processor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110363107A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111126344A (en) * | 2019-12-31 | 2020-05-08 | 杭州趣维科技有限公司 | Method and system for generating key points of forehead of human face |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021550A (en) * | 2014-05-22 | 2014-09-03 | 西安理工大学 | Automatic positioning and proportion determining method for proportion of human face |
| CN104157001A (en) * | 2014-08-08 | 2014-11-19 | 中科创达软件股份有限公司 | Method and device for drawing head caricature |
| CN106934375A (en) * | 2017-03-15 | 2017-07-07 | 中南林业科技大学 | The facial expression recognizing method of distinguished point based movement locus description |
| CN108009470A (en) * | 2017-10-20 | 2018-05-08 | 深圳市朗形网络科技有限公司 | A kind of method and apparatus of image zooming-out |
| CN109410298A (en) * | 2018-11-02 | 2019-03-01 | 北京恒信彩虹科技有限公司 | A kind of production method and expression shape change method of dummy model |
-
2019
- 2019-06-26 CN CN201910562255.XA patent/CN110363107A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104021550A (en) * | 2014-05-22 | 2014-09-03 | 西安理工大学 | Automatic positioning and proportion determining method for proportion of human face |
| CN104157001A (en) * | 2014-08-08 | 2014-11-19 | 中科创达软件股份有限公司 | Method and device for drawing head caricature |
| CN106934375A (en) * | 2017-03-15 | 2017-07-07 | 中南林业科技大学 | The facial expression recognizing method of distinguished point based movement locus description |
| CN108009470A (en) * | 2017-10-20 | 2018-05-08 | 深圳市朗形网络科技有限公司 | A kind of method and apparatus of image zooming-out |
| CN109410298A (en) * | 2018-11-02 | 2019-03-01 | 北京恒信彩虹科技有限公司 | A kind of production method and expression shape change method of dummy model |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111126344A (en) * | 2019-12-31 | 2020-05-08 | 杭州趣维科技有限公司 | Method and system for generating key points of forehead of human face |
| CN111126344B (en) * | 2019-12-31 | 2023-08-01 | 杭州趣维科技有限公司 | Method and system for generating key points of forehead of human face |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | SaliencyGAN: Deep learning semisupervised salient object detection in the fog of IoT | |
| US11487995B2 (en) | Method and apparatus for determining image quality | |
| CN111339990B (en) | Face recognition system and method based on dynamic update of face features | |
| WO2016023264A1 (en) | Fingerprint identification method and fingerprint identification device | |
| WO2019105163A1 (en) | Target person search method and apparatus, device, program product and medium | |
| WO2016138838A1 (en) | Method and device for recognizing lip-reading based on projection extreme learning machine | |
| WO2021208601A1 (en) | Artificial-intelligence-based image processing method and apparatus, and device and storage medium | |
| WO2022228252A1 (en) | Human behavior detection method and apparatus, electronic device and storage medium | |
| US20230230305A1 (en) | Online streamer avatar generation method and apparatus | |
| CN111680550B (en) | Emotion information identification method and device, storage medium and computer equipment | |
| WO2021223738A1 (en) | Method, apparatus and device for updating model parameter, and storage medium | |
| CN110675469A (en) | Image description method for detecting spatial relationship between targets in construction scene | |
| CN111754637A (en) | Large-scale three-dimensional face synthesis system with suppressed sample similarity | |
| CN113095421A (en) | Method for generating font database, and training method and device of neural network model | |
| CN112580666A (en) | Image feature extraction method, training method, device, electronic equipment and medium | |
| CN107330363B (en) | Rapid internet billboard detection method | |
| TW202201275A (en) | Device and method for scoring hand work motion and storage medium | |
| CN110363107A (en) | Face forehead point Quick Extended method, apparatus, storage medium and processor | |
| CN110377526A (en) | A kind of test method of AI intelligence software | |
| CN106778925A (en) | A kind of super complete face automatic registration method of the attitude of recognition of face and its device | |
| CN111126344B (en) | Method and system for generating key points of forehead of human face | |
| CN110826501B (en) | Face key point detection method and system based on sparse key point calibration | |
| CN110363110A (en) | Face forward reference method for fast establishing, device, storage medium and processor | |
| CN113903071A (en) | Face recognition method and device, electronic equipment and storage medium | |
| CN107506400A (en) | A kind of image search method based on cognitive characteristics and manifold ranking |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191022 |
|
| RJ01 | Rejection of invention patent application after publication |