CN109902616A - Face three-dimensional feature point detecting method and system based on deep learning - Google Patents
Face three-dimensional feature point detecting method and system based on deep learning Download PDFInfo
- Publication number
- CN109902616A CN109902616A CN201910138641.6A CN201910138641A CN109902616A CN 109902616 A CN109902616 A CN 109902616A CN 201910138641 A CN201910138641 A CN 201910138641A CN 109902616 A CN109902616 A CN 109902616A
- Authority
- CN
- China
- Prior art keywords
- face
- feature point
- dimensional feature
- dimensional
- picture
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Abstract
The invention discloses a kind of face three-dimensional feature point detecting method and system based on deep learning, wherein this method comprises: establishing human face data collection, processing is carried out to the face picture that human face data is concentrated by face three-dimensional reconstruction and obtains face geometry；Certain vertex is demarcated on face template is characterized the data set for a little establishing the corresponding face three-dimensional feature point composition of face picture；Training input is face picture, and exports the deep neural network of the distribution temperature figure for face three-dimensional feature point coordinate；Use generation confrontation network to utilize discrimination natwork when training, input is that face picture and three-dimensional feature point are distributed temperature figure, exporting true-false value indicates whether the face picture of input and three-dimensional feature point distribution temperature figure are mating, to obtain testing result by the neural network after training.This method can detecte the three-dimensional coordinate of human face characteristic point in picture, and contact between face marginal point and faceform with very strong, so that human face rebuilding result is more accurate.
Description
Technical field
The present invention relates to computer vision and graphics techniques field, in particular to a kind of face three based on deep learning
Dimensional feature point detecting method and system.
Background technique
The concept of deep learning is derived from the research of artificial neural network, and the multilayer perceptron containing more hidden layers is exactly a kind of depth
Learning structure.Deep learning, which forms more abstract high level by combination low-level feature, indicates attribute classification or feature, with discovery
The distributed nature of data indicates.
Facial feature points detection has important application in recognition of face, human face rebuilding and face tracking.In face weight
It builds, in face tracking and the non-rigid registration of face, generally requires pair in specific characteristic point and faceform's template between vertex
It should be related to, among application, the use of two-dimension human face characteristic point has many inconvenience, and not with corresponding relationship in faceform's template
Determining face Edge Feature Points will lead to human face rebuilding result inaccuracy, bring certain difficulty to application.
Summary of the invention
The present invention is directed to solve at least some of the technical problems in related technologies.
For this purpose, an object of the present invention is to provide a kind of face three-dimensional feature point detection side based on deep learning
Method.This method only needs to input individual face picture, so that it may export the three-dimensional feature point of face, and this feature point and face are true
There is stronger corresponding relationship between geometry of reals.
It is another object of the present invention to propose a kind of face three-dimensional feature point detection system based on deep learning.
In order to achieve the above objectives, one aspect of the present invention proposes the face three-dimensional feature point detection side based on deep learning
Method, comprising the following steps: establish human face data collection, the face picture that the human face data is concentrated by face three-dimensional reconstruction into
Row processing, to obtain face geometry；Certain vertex is demarcated in face three-dimensional template to be characterized a little, and it is right with it to establish face picture
The data set that the face three-dimensional feature point answered is constituted；Training input is face picture, and exporting is face three-dimensional feature point coordinate
Distribution temperature figure deep neural network；In training, use generation confrontation network to utilize discrimination natwork, wherein input
It is distributed temperature figure for face picture and three-dimensional feature point, exports the face picture and three-dimensional feature point minute for indicating input for true or false
Whether cloth temperature figure is mating, to obtain testing result by the neural network after training.
The face three-dimensional feature point detecting method based on deep learning of the embodiment of the present invention, by utilizing deep learning
Method is trained to obtain the network of face three-dimensional feature point detection, obtains and faceform's template corresponding relationship and determination
Face Edge Feature Points so that human face rebuilding result is more accurate, and apply very simple.
In addition, the face three-dimensional feature point detecting method according to the above embodiment of the present invention based on deep learning can be with
With following additional technical characteristic:
Further, in one embodiment of the invention, the characteristic point is three-dimensional feature point, and the three-dimensional feature
There are corresponding relationships with the face three-dimensional template for point.
Further, in one embodiment of the invention, it is described by face three-dimensional reconstruction to the human face data collection
In face picture carry out processing be face endpoint detections, to obtain the face edge shown in the not described face picture
Point, but the marginal point of the face geometrically two sides.
Further, in one embodiment of the invention, further includes: in training, export as true or false to constitute life
Temperature figure structure is distributed at confrontation error, and by the three-dimensional feature point of face three-dimensional feature point and generation network in the data set
The error term built is trained.
Further, in one embodiment of the invention, the error term needs in generation confrontation network, with
Training for the neural network.
In order to achieve the above objectives, another aspect of the present invention proposes a kind of face three-dimensional feature point inspection based on deep learning
Examining system, comprising: processing module is for establishing human face data collection, the people concentrated by face three-dimensional reconstruction to the human face data
Face picture is handled, and to obtain face geometry, and is demarcated certain vertex in face three-dimensional template and is characterized a little, establishes face
The data set that the corresponding face three-dimensional feature point of picture is constituted；Default training module is face picture for training input,
And output is the deep neural network of the distribution temperature figure of face three-dimensional feature point coordinate；Dual training module is generated for instructing
When practicing, use generation confrontation network to utilize discrimination natwork, wherein to input and be distributed temperature for face picture and three-dimensional feature point
Figure, exporting indicates whether the face picture of input and three-dimensional feature point distribution temperature figure are mating for true or false, after through training
Neural network obtain testing result.
The face three-dimensional feature point detection system based on deep learning of the embodiment of the present invention, by utilizing deep learning
Method is trained to obtain the network of face three-dimensional feature point detection, obtains and faceform's template corresponding relationship and determination
Face Edge Feature Points so that human face rebuilding result is more accurate, and apply very simple.
In addition, the face three-dimensional feature point detection system according to the above embodiment of the present invention based on deep learning can be with
With following additional technical characteristic:
Further, in one embodiment of the invention, the characteristic point is three-dimensional feature point, and the three-dimensional feature
There are corresponding relationships with the face three-dimensional template for point.
Further, in one embodiment of the invention, it is described by face three-dimensional reconstruction to the human face data collection
In face picture carry out processing be face endpoint detections, to obtain the face edge shown in the not described face picture
Point, but the marginal point of the face geometrically two sides.
Further, in one embodiment of the invention, further includes: in training, export as true or false to constitute life
Temperature figure structure is distributed at confrontation error, and by the three-dimensional feature point of face three-dimensional feature point and generation network in the data set
The error term built is trained.
Further, in one embodiment of the invention, the error term needs in generation confrontation network, with
Training for the neural network.
The additional aspect of the present invention and advantage will be set forth in part in the description, and will partially become from the following description
Obviously, or practice through the invention is recognized.
Detailed description of the invention
Above-mentioned and/or additional aspect and advantage of the invention will become from the following description of the accompanying drawings of embodiments
Obviously and it is readily appreciated that, in which:
Fig. 1 is the face three-dimensional feature point detecting method flow chart based on deep learning according to the embodiment of the present invention；
Fig. 2 is the face three-dimensional rebuilding method flow chart according to the embodiment of the present invention；
Fig. 3 is to detect network training method flow chart according to the face three-dimensional feature point of the embodiment of the present invention；
Fig. 4 is the face three-dimensional feature point overhaul flow chart according to the embodiment of the present invention；
Fig. 5 is the face three-dimensional feature point detection system structural representation based on deep learning according to the embodiment of the present invention
Figure.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to is used to explain the present invention, and is not considered as limiting the invention.
The face three-dimensional feature point inspection based on deep learning proposed according to embodiments of the present invention is described with reference to the accompanying drawings
Method and system are surveyed, it is three-dimensional special to describe the face based on deep learning proposed according to embodiments of the present invention with reference to the accompanying drawings first
Levy point detecting method.
Fig. 1 is the face three-dimensional feature point detecting method flow chart based on deep learning of one embodiment of the invention.
As shown in Figure 1, should face three-dimensional feature point detecting method based on deep learning the following steps are included:
In step s101, human face data collection is established, the face picture concentrated by face three-dimensional reconstruction to human face data
It is handled, to obtain face geometry.
Wherein, characteristic point is three-dimensional feature point, and there are corresponding relationships with face three-dimensional template for three-dimensional feature point.
It should be noted that carrying out processing to the face picture that human face data is concentrated by face three-dimensional reconstruction is face side
The detection of edge point, to obtain the face marginal point shown in not face picture, but the marginal point of face geometrically two sides.
In step s 102, certain vertex is demarcated in face three-dimensional template to be characterized a little, it is right with it to establish face picture
The data set that the face three-dimensional feature point answered is constituted.
That is, handled by the face picture that face three-dimensional reconstruction concentrates data, to obtain face geometry,
It geometrically specifies certain vertex to be characterized a little in face in advance, thus obtains human face characteristic point coordinate.
Specifically, it as shown in Fig. 2, being handled by face picture of the face three-dimensional rebuilding method to input, utilizes
It obtains input picture to deform faceform's template, the geometrical model of face in input picture is obtained, on face template
Specified certain vertex is characterized a little, thus obtains face three-dimensional feature point coordinate.
In step s 103, training input is face picture, and output is the distribution temperature of face three-dimensional feature point coordinate
The deep neural network of figure.
Wherein, deep neural network is trained in advance, and not applicable when pre-training to generate countercheck, only training data constructs
Training error is trained.
Specifically, as shown in figure 3, training input is face picture, and exporting is that face three-dimensional feature point is distributed temperature figure,
Each temperature figure includes the probability of three-dimensional feature point distribution and the depth of corresponding points.Wherein, the generation network is trained in advance,
It is face picture that it, which is inputted, exports the temperature figure for the distribution of face three-dimensional feature point, using the result of the pre-training as generation pair
The initialization of anti-network generates network.It outputs it to input with it and combine as negative sample, it is corresponding with data set with its input
Standard output combination be used as positive sample, the input as discrimination natwork.
In step S104, in training, use generation confrontation network to utilize discrimination natwork, wherein to input as face
Picture and three-dimensional feature point are distributed temperature figure, and exporting indicates that the face picture of input and three-dimensional feature point are distributed temperature for true or false
Whether figure is mating, to obtain testing result by the neural network after training.
It should be noted that exporting in training and generating confrontation error for true or false to constitute, and by the people in data set
Face three-dimensional feature point and the error term for the three-dimensional feature point distribution temperature figure building for generating network are trained.Wherein, error term
It needs in generating confrontation network, with the training for neural network.
In other words, the network obtained after pre-training makes a living into one discrimination natwork of network struction, and input is face picture
And the temperature figure of face three-dimensional feature point distribution, when input is face picture and three-dimensional feature point distribution temperature in data set
When figure combination, output should be true, be otherwise it is false, generation confrontation error is constructed with this, what when pre-training, used, face in data set
Three-dimensional feature point and the error term for generating the three-dimensional feature point distribution temperature figure building that network generates will also generate dual training
Among be used for network training.
Particularly, as shown in figure 4, the network can be used after network training is good, a face picture is inputted, i.e.,
The distribution temperature figure of the corresponding each three-dimensional feature point of the face picture can be obtained, each pixel of temperature figure includes corresponding
The probability and depth information being distributed at this after obtaining the temperature distribution map, can therefrom extract the highest pixel of probability, knot
Depth information is closed, the position of this feature point is calculated.
The face three-dimensional feature point detecting method based on deep learning proposed according to embodiments of the present invention, by using deeply
The method of degree study is trained to obtain the network of face three-dimensional feature point detection, obtains close corresponding with faceform's template
System and the face Edge Feature Points determined so that human face rebuilding result is more accurate, and apply very simple.
The face three-dimensional feature point inspection based on deep learning proposed according to embodiments of the present invention referring next to attached drawing description
Examining system.
Fig. 5 is the face three-dimensional feature point detection system structural representation based on deep learning of one embodiment of the invention
Figure.
As shown in figure 5, the system 10 includes: processing module 100, default training module 200 and generates dual training module
300。
Wherein, processing module 100 is for establishing human face data collection, the people concentrated by face three-dimensional reconstruction to human face data
Face picture is handled, and to obtain face geometry, and is demarcated certain vertex in face three-dimensional template and is characterized a little, establishes face
The data set that the corresponding face three-dimensional feature point of picture is constituted.
Further, in one embodiment of the invention, characteristic point is three-dimensional feature point, and three-dimensional feature point and face
There are corresponding relationships for three-dimensional template, and carrying out processing to the face picture that human face data is concentrated by face three-dimensional reconstruction is people
Face endpoint detections, to obtain the face marginal point shown in not face picture, but the marginal point of face geometrically two sides.
Minute that default training module 200 is used to train input for face picture, and export as face three-dimensional feature point coordinate
The deep neural network of cloth temperature figure.
It generates dual training module 300 to be used in training, uses generation confrontation network to utilize discrimination natwork, wherein
Input is that face picture and three-dimensional feature point are distributed temperature figure, exports the face picture and three-dimensional feature that input is indicated for true or false
Whether point distribution temperature figure is mating, to obtain testing result by the neural network after training.
Further, in one embodiment of the invention, further includes: in training, export as true or false to constitute life
At confrontation error, and the three-dimensional feature point distribution temperature figure of face three-dimensional feature point and generation network in data set is constructed
Error term is trained.Wherein, error term needs in generating confrontation network, with the training for neural network.
It should be noted that aforementioned explaining to the face three-dimensional feature point detecting method embodiment based on deep learning
Bright to be also applied for the system, details are not described herein again.
The face three-dimensional feature point detection system based on deep learning proposed according to embodiments of the present invention, by using deeply
The method of degree study is trained to obtain the network of face three-dimensional feature point detection, obtains close corresponding with faceform's template
System and the face Edge Feature Points determined so that human face rebuilding result is more accurate, and apply very simple.
In addition, term " first ", " second " are used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance
Or implicitly indicate the quantity of indicated technical characteristic.Define " first " as a result, the feature of " second " can be expressed or
Implicitly include at least one this feature.In the description of the present invention, the meaning of " plurality " is at least two, such as two, three
It is a etc., unless otherwise specifically defined.
In the present invention unless specifically defined or limited otherwise, term " installation ", " connected ", " connection ", " fixation " etc.
Term shall be understood in a broad sense, for example, it may be being fixedly connected, may be a detachable connection, or integral；It can be mechanical connect
It connects, is also possible to be electrically connected；It can be directly connected, can also can be in two elements indirectly connected through an intermediary
The interaction relationship of the connection in portion or two elements, unless otherwise restricted clearly.For those of ordinary skill in the art
For, the specific meanings of the above terms in the present invention can be understood according to specific conditions.
In the present invention unless specifically defined or limited otherwise, fisrt feature in the second feature " on " or " down " can be with
It is that the first and second features directly contact or the first and second features pass through intermediary mediate contact.Moreover, fisrt feature exists
Second feature " on ", " top " and " above " but fisrt feature be directly above or diagonally above the second feature, or be merely representative of
First feature horizontal height is higher than second feature.Fisrt feature can be under the second feature " below ", " below " and " below "
One feature is directly under or diagonally below the second feature, or is merely representative of first feature horizontal height less than second feature.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
It must be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be in office
It can be combined in any suitable manner in one or more embodiment or examples.In addition, without conflicting with each other, the skill of this field
Art personnel can tie the feature of different embodiments or examples described in this specification and different embodiments or examples
It closes and combines.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art within the scope of the invention can be to above-mentioned
Embodiment is changed, modifies, replacement and variant.
Claims (10)
1. a kind of face three-dimensional feature point detecting method based on deep learning, which comprises the following steps:
Human face data collection is established, is handled by the face picture that face three-dimensional reconstruction concentrates the human face data, to obtain
Take face geometry；
Calibration certain vertex, which is characterized, in face three-dimensional template a little establishes the corresponding face three-dimensional feature point of face picture
The data set of composition；
Training input is face picture, and exports the deep neural network of the distribution temperature figure for face three-dimensional feature point coordinate；
And
In training, use generation confrontation network to utilize discrimination natwork, wherein to input as face picture and three-dimensional feature point minute
Cloth temperature figure, exporting indicates whether the face picture of input and three-dimensional feature point distribution temperature figure are mating for true or false, to pass through
Neural network after training obtains testing result.
2. the face three-dimensional feature point detecting method according to claim 1 based on deep learning, which is characterized in that described
Characteristic point is three-dimensional feature point, and there are corresponding relationships with the face three-dimensional template for three-dimensional feature point.
3. the face three-dimensional feature point detecting method according to claim 1 based on deep learning, which is characterized in that described
Carrying out processing to the face picture that the human face data is concentrated by face three-dimensional reconstruction is face endpoint detections, to obtain simultaneously
The face marginal point shown in the non-face picture, but the marginal point of the face geometrically two sides.
4. the face three-dimensional feature point detecting method according to claim 1 based on deep learning, which is characterized in that also wrap
It includes:
In training, exports and generate confrontation error for true or false to constitute, and by the face three-dimensional feature point in the data set
It is trained with the error term for the three-dimensional feature point distribution temperature figure building for generating network.
5. the face three-dimensional feature point detecting method according to claim 4 based on deep learning, which is characterized in that described
Error term needs in generation confrontation network, with the training for the neural network.
6. a kind of face three-dimensional feature point detection system based on deep learning characterized by comprising
Processing module, for establishing human face data collection, the human face data is concentrated by face three-dimensional reconstruction face picture
Handled, to obtain face geometry, and in face three-dimensional template demarcate certain vertex be characterized a little, establish face picture with
The data set that its corresponding face three-dimensional feature point is constituted；
Default training module, is used for the distribution temperature trained input for face picture, and exported as face three-dimensional feature point coordinate
The deep neural network of figure；And
Dual training module is generated, in training, using generation confrontation network with using discrimination natwork, wherein input and be
Face picture and three-dimensional feature point are distributed temperature figure, export the face picture and the distribution of three-dimensional feature point that input is indicated for true or false
Whether temperature figure is mating, to obtain testing result by the neural network after training.
7. the face three-dimensional feature point detection system according to claim 6 based on deep learning, which is characterized in that described
Characteristic point is three-dimensional feature point, and there are corresponding relationships with the face three-dimensional template for three-dimensional feature point.
8. the face three-dimensional feature point detection system according to claim 6 based on deep learning, which is characterized in that described
Carrying out processing to the face picture that the human face data is concentrated by face three-dimensional reconstruction is face endpoint detections, to obtain simultaneously
The face marginal point shown in the non-face picture, but the marginal point of the face geometrically two sides.
9. the face three-dimensional feature point detection system according to claim 6 based on deep learning, which is characterized in that also wrap
It includes:
In training, exports and generate confrontation error for true or false to constitute, and by the face three-dimensional feature point in the data set
It is trained with the error term for the three-dimensional feature point distribution temperature figure building for generating network.
10. the face three-dimensional feature point detection system according to claim 9 based on deep learning, which is characterized in that institute
It states error term to need in generation confrontation network, with the training for the neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910138641.6A CN109902616B (en) | 2019-02-25 | 2019-02-25 | Human face three-dimensional feature point detection method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910138641.6A CN109902616B (en) | 2019-02-25 | 2019-02-25 | Human face three-dimensional feature point detection method and system based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109902616A true CN109902616A (en) | 2019-06-18 |
| CN109902616B CN109902616B (en) | 2020-12-01 |
Family
ID=66945627
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910138641.6A Active CN109902616B (en) | 2019-02-25 | 2019-02-25 | Human face three-dimensional feature point detection method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109902616B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110516643A (en) * | 2019-08-30 | 2019-11-29 | 电子科技大学 | A kind of face 3D critical point detection method and system based on joint thermodynamic chart |
| CN110569768A (en) * | 2019-08-29 | 2019-12-13 | 四川大学 | construction method of face model, face recognition method, device and equipment |
| CN110765976A (en) * | 2019-11-01 | 2020-02-07 | 重庆紫光华山智安科技有限公司 | Generation method of human face characteristic points, training method of data network and related device |
| CN111583422A (en) * | 2020-04-17 | 2020-08-25 | 清华大学 | Heuristic editing method and device for three-dimensional human body model |
| CN112883920A (en) * | 2021-03-22 | 2021-06-01 | 清华大学 | Point cloud deep learning-based three-dimensional face scanning feature point detection method and device |
| CN113128292A (en) * | 2019-12-31 | 2021-07-16 | Tcl集团股份有限公司 | Image identification method, storage medium and terminal equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140185924A1 (en) * | 2012-12-27 | 2014-07-03 | Microsoft Corporation | Face Alignment by Explicit Shape Regression |
| CN106980812A (en) * | 2016-12-14 | 2017-07-25 | 四川长虹电器股份有限公司 | Three-dimensional face features' independent positioning method based on concatenated convolutional neutral net |
| CN107045631A (en) * | 2017-05-25 | 2017-08-15 | 北京华捷艾米科技有限公司 | Facial feature points detection method, device and equipment |
| CN107122705A (en) * | 2017-03-17 | 2017-09-01 | 中国科学院自动化研究所 | Face critical point detection method based on three-dimensional face model |
| CN109063695A (en) * | 2018-09-18 | 2018-12-21 | 图普科技（广州）有限公司 | A kind of face critical point detection method, apparatus and its computer storage medium |
| CN109241910A (en) * | 2018-09-07 | 2019-01-18 | 高新兴科技集团股份有限公司 | A kind of face key independent positioning method returned based on the cascade of depth multiple features fusion |
-
2019
- 2019-02-25 CN CN201910138641.6A patent/CN109902616B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140185924A1 (en) * | 2012-12-27 | 2014-07-03 | Microsoft Corporation | Face Alignment by Explicit Shape Regression |
| CN106980812A (en) * | 2016-12-14 | 2017-07-25 | 四川长虹电器股份有限公司 | Three-dimensional face features' independent positioning method based on concatenated convolutional neutral net |
| CN107122705A (en) * | 2017-03-17 | 2017-09-01 | 中国科学院自动化研究所 | Face critical point detection method based on three-dimensional face model |
| CN107045631A (en) * | 2017-05-25 | 2017-08-15 | 北京华捷艾米科技有限公司 | Facial feature points detection method, device and equipment |
| CN109241910A (en) * | 2018-09-07 | 2019-01-18 | 高新兴科技集团股份有限公司 | A kind of face key independent positioning method returned based on the cascade of depth multiple features fusion |
| CN109063695A (en) * | 2018-09-18 | 2018-12-21 | 图普科技（广州）有限公司 | A kind of face critical point detection method, apparatus and its computer storage medium |
Non-Patent Citations (3)
| Title |
|---|
| ADRIAN BULAT AND GEORGIOS TZIMIROPOULOS: ""Two-stage Convolutional Part Heatmap Regression for the 1st 3D Face Alignment in the Wild (3DFAW) Challenge"", 《ARXIV》 * |
| WAYNE WU，CHEN QIAN ET AL.: ""Look at Boundary: A Boundary-Aware Face Alignment Algorithm"", 《ARXIV》 * |
| 杨俊钦，张雨楠，林实锋，徐伟臻: ""基于深度学习的人脸多属性识别系统"", 《图形图像》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110569768A (en) * | 2019-08-29 | 2019-12-13 | 四川大学 | construction method of face model, face recognition method, device and equipment |
| CN110569768B (en) * | 2019-08-29 | 2022-09-02 | 四川大学 | Construction method of face model, face recognition method, device and equipment |
| CN110516643A (en) * | 2019-08-30 | 2019-11-29 | 电子科技大学 | A kind of face 3D critical point detection method and system based on joint thermodynamic chart |
| CN110765976A (en) * | 2019-11-01 | 2020-02-07 | 重庆紫光华山智安科技有限公司 | Generation method of human face characteristic points, training method of data network and related device |
| CN113128292A (en) * | 2019-12-31 | 2021-07-16 | Tcl集团股份有限公司 | Image identification method, storage medium and terminal equipment |
| CN111583422A (en) * | 2020-04-17 | 2020-08-25 | 清华大学 | Heuristic editing method and device for three-dimensional human body model |
| CN111583422B (en) * | 2020-04-17 | 2023-03-28 | 清华大学 | Heuristic editing method and device for three-dimensional human body model |
| CN112883920A (en) * | 2021-03-22 | 2021-06-01 | 清华大学 | Point cloud deep learning-based three-dimensional face scanning feature point detection method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109902616B (en) | 2020-12-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109902616A (en) | Face three-dimensional feature point detecting method and system based on deep learning | |
| CN107437092B (en) | The classification method of retina OCT image based on Three dimensional convolution neural network | |
| O'Toole et al. | 3D shape and 2D surface textures of human faces: The role of “averages” in attractiveness and age | |
| CN105574518B (en) | Method and device for detecting living human face | |
| CN109543526B (en) | True and false facial paralysis recognition system based on depth difference characteristics | |
| CN106295568A (en) | The mankind's naturalness emotion identification method combined based on expression and behavior bimodal | |
| CN111528859B (en) | Child ADHD screening and evaluating system based on multi-modal deep learning technology | |
| CN110287805A (en) | Micro- expression recognition method and system based on three stream convolutional neural networks | |
| CN106796449A (en) | Eye-controlling focus method and device | |
| CN106529442A (en) | Pedestrian identification method and apparatus | |
| Wang et al. | Investigation into recognition algorithm of helmet violation based on YOLOv5-CBAM-DCN | |
| CN106777953A (en) | The analysis method and system of medical image data | |
| CN101329722A (en) | Human face recognition method for performing recognition algorithm based on neural network | |
| CN108846343B (en) | Multi-task collaborative analysis method based on three-dimensional video | |
| CN110175501A (en) | More people's scene focus recognition methods based on recognition of face | |
| CN110458005A (en) | It is a kind of based on the progressive invariable rotary method for detecting human face with pseudo-crystalline lattice of multitask | |
| Li et al. | Non-intrusive comfort sensing: Detecting age and gender from infrared images for personal thermal comfort | |
| CN107392151A (en) | Face image various dimensions emotion judgement system and method based on neutral net | |
| Poggi et al. | Confidence estimation for ToF and stereo sensors and its application to depth data fusion | |
| Dosso et al. | NICUface: Robust neonatal face detection in complex NICU scenes | |
| CN106725341A (en) | A kind of enhanced lingual diagnosis system | |
| CN105844243B (en) | A kind of finger multi-modal biological characteristic granulation fusion method based on geometry | |
| Nasir et al. | Fuzzy triangulation signature for detection of change in human emotion from face video image sequence | |
| CN114067185A (en) | Film evaluation system based on facial expression recognition | |
| CN108304800A (en) | A kind of method of Face datection and face alignment |
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 |