CN109840477B - Method and device for recognizing shielded face based on feature transformation - Google Patents
Method and device for recognizing shielded face based on feature transformation Download PDFInfo
- Publication number
- CN109840477B CN109840477B CN201910006884.4A CN201910006884A CN109840477B CN 109840477 B CN109840477 B CN 109840477B CN 201910006884 A CN201910006884 A CN 201910006884A CN 109840477 B CN109840477 B CN 109840477B
- Authority
- CN
- China
- Prior art keywords
- feature
- neural network
- image
- network model
- convolutional neural
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the invention provides a method and a device for identifying an occluded face based on feature transformation, wherein the method comprises the following steps: inputting a target face image into a preset convolutional neural network model, and acquiring a feature map of the target face image from the last convolutional layer of the convolutional neural network model; calculating the pixel product of the feature map and a feature mask acquired in advance; and inputting the sub-pixel product to a feature layer of the convolutional neural network model, and outputting the feature of the target face image for recognition. According to the method and the device for identifying the shielded face based on the feature transformation, the feature transformation mode of adding the feature mask is adopted for the image of the shielded face, the feature response of the common easily shielded area of the face is abandoned, the engineering is easier to realize, the calculation time is shorter, the network structure is simpler, and the identification precision of the shielded face is improved.
Description
Technical Field
The embodiment of the invention relates to the technical field of face recognition, in particular to a method and a device for recognizing an occluded face based on feature transformation.
Background
The application of face recognition is more and more extensive, the requirement on the face recognition accuracy is more and more high, and especially the recognition accuracy under the condition that the face part is shielded is more important.
In the prior art, a face recognition algorithm based on a convolutional neural network is generally used for face recognition, but the face recognition algorithm based on the convolutional neural network depends on the quality of a data set to a great extent, but the face in an actual scene has greater complexity in the aspects of occlusion, angle and the like. The manual labeling difficulty of the shielded face picture is higher, so that most of the existing face recognition technologies have the problem that the recognition accuracy of the shielded face picture is seriously reduced. Most of the existing algorithms for improving the human face shielding robustness adopt a method of respectively training a plurality of human face regions by a plurality of network structures to fuse the characteristics of different regions of the human face, but the methods have the defects of high consumption of computing resources, longer computing time and more complex preprocessing of human face pictures.
Disclosure of Invention
It is an object of embodiments of the present invention to provide a method and apparatus for feature transform based occluded face recognition that overcomes or at least partially solves the above mentioned problems.
In order to solve the above technical problem, in one aspect, an embodiment of the present invention provides an occluded face recognition method based on feature transformation, including:
inputting a target face image into a preset convolutional neural network model, and acquiring a feature map of the target face image from the last convolutional layer of the convolutional neural network model;
calculating the pixel product of the feature map and a feature mask acquired in advance;
and inputting the sub-pixel product to a feature layer of the convolutional neural network model, and outputting the feature of the target face image for recognition.
In another aspect, an embodiment of the present invention provides an occluded face recognition device based on feature transformation, including:
the characteristic diagram extraction module is used for inputting a target face image into a preset convolutional neural network model and acquiring a characteristic diagram of the target face image from the last convolutional layer of the convolutional neural network model;
the calculation module is used for calculating the fractional pixel product of the feature map and a feature mask acquired in advance;
and the feature acquisition module is used for inputting the sub-pixel product to a feature layer of the convolutional neural network model and outputting the feature of the target face image for identification.
In another aspect, an embodiment of the present invention provides an electronic device, including:
the processor and the memory are communicated with each other through a bus; the memory stores program instructions executable by the processor, which when called by the processor are capable of performing the methods described above.
In yet another aspect, the present invention provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the above-mentioned method.
According to the method and the device for identifying the shielded face based on the feature transformation, the feature transformation mode of adding the feature mask is adopted for the image of the shielded face, the feature response of the common easily shielded area of the face is abandoned, the engineering is easier to realize, the calculation time is shorter, the network structure is simpler, and the identification precision of the shielded face is improved.
Drawings
FIG. 1 is a schematic diagram of an occluded face recognition method based on feature transformation according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of an occluded face recognition device based on feature transformation according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic diagram of an occluded face recognition method based on feature transformation according to an embodiment of the present invention, and as shown in fig. 1, an occluded face recognition method based on feature transformation according to an embodiment of the present invention is provided, an execution subject of which is an occluded face recognition device based on feature transformation, and the method includes:
step S101, inputting a target face image into a preset convolutional neural network model, and acquiring a feature map of the target face image from the last convolutional layer of the convolutional neural network model;
step S102, calculating the fractional pixel product of the feature map and a feature mask acquired in advance;
and S103, inputting the sub-pixel product to a feature layer of the convolutional neural network model, and outputting the feature of the target face image for identification.
Specifically, the method for recognizing an occluded face based on feature transformation provided by the embodiment of the present invention includes two stages: the first, experiment stage, the second, application stage.
In the experimental phase, a feature mask is obtained from the experiment.
In the application stage, firstly, the target face image is input into a preset convolutional neural network model, and the feature map of the target face image is obtained from the last convolutional layer of the convolutional neural network model.
Before inputting the feature map of the target face image into the feature layer, calculating the fractional product of the feature map of the target face image and a feature mask acquired in advance. The feature mask is acquired in advance at an experimental stage.
And finally, inputting the calculated sub-pixel product serving as a transformed feature map into a feature layer of the convolutional neural network model, and outputting the feature used for recognition of the target face image so as to recognize the target face image according to the feature.
According to the method for identifying the shielded face based on the feature transformation, provided by the embodiment of the invention, the feature transformation mode of adding the feature mask is adopted for the image of the shielded face, the feature response of the common easily shielded area of the face is abandoned, the engineering is easier to realize, the calculation time is shorter, the network structure is simpler, and the identification precision of the shielded face is improved.
On the basis of the above embodiment, further, the specific steps of obtaining the feature mask are as follows:
acquiring a plurality of original sample images in a target data set, adding an image subjected to artificial shielding to the original sample images, taking the original sample images subjected to the artificial shielding as shielding sample images, wherein one original sample image corresponds to one shielding sample image;
acquiring an initial characteristic mask according to a pair of original sample images and a shielding sample image;
the average of all the initial feature masks is taken as the final feature mask.
Specifically, the method for recognizing an occluded face based on feature transformation provided by the embodiment of the present invention includes two stages: the first, experiment stage, the second, application stage.
In the experimental phase, a feature mask is obtained from the experiment.
The specific steps for obtaining the feature mask are as follows:
firstly, acquiring a plurality of original sample images in a target data set, aligning the original sample images according to human face characteristic points aiming at each original sample image, adding artificial shielding to the human face image according to the characteristic points, acquiring an image obtained after the artificial shielding is added to the original sample image, taking the original sample image obtained after the artificial shielding is added as a shielding sample image, and enabling one original sample image to correspond to one shielding sample image.
Then, an initial feature mask is obtained according to a pair of the original sample image and the occlusion sample image.
Finally, the average of all the initial feature masks is taken as the final feature mask.
According to the method for identifying the shielded face based on the feature transformation, provided by the embodiment of the invention, the feature transformation mode of adding the feature mask is adopted for the image of the shielded face, the feature response of the common easily shielded area of the face is abandoned, the engineering is easier to realize, the calculation time is shorter, the network structure is simpler, and the identification precision of the shielded face is improved.
On the basis of the foregoing embodiments, further, the obtaining an initial feature mask according to a pair of original sample images and an occlusion sample image specifically includes:
inputting the target original sample image into the convolutional neural network model aiming at the target original sample image, and acquiring a feature map of the target original sample image from the last convolutional layer of the convolutional neural network model as a first feature map; inputting a target shielding sample image corresponding to the target original sample image into the convolutional neural network model, and acquiring a feature map of the target shielding sample image from the last convolutional layer of the convolutional neural network model as a second feature map;
performing difference on the first characteristic diagram and the second characteristic diagram to obtain a difference value matrix;
and carrying out binarization processing on the elements in the difference value matrix to obtain an initial characteristic mask.
Specifically, the specific steps of obtaining the feature mask are as follows:
firstly, acquiring a plurality of original sample images in a target data set, aligning the original sample images according to human face characteristic points aiming at each original sample image, adding artificial shielding to the human face image according to the characteristic points, acquiring an image obtained after the artificial shielding is added to the original sample image, taking the original sample image obtained after the artificial shielding is added as a shielding sample image, and enabling one original sample image to correspond to one shielding sample image.
Then, an initial feature mask is obtained according to a pair of the original sample image and the occlusion sample image.
Finally, the average of all the initial feature masks is taken as the final feature mask.
The method includes the steps of obtaining an initial feature mask according to a pair of original sample images and a shielding sample image, and specifically includes the following steps:
and aiming at the target original sample image, inputting the target original sample image without occlusion into the convolutional neural network model, and acquiring a feature map of the target original sample image from the last convolutional layer of the convolutional neural network model to be used as a first feature map.
And inputting a target shielding sample image corresponding to the target original sample image into the convolutional neural network model, and acquiring a feature map of the target shielding sample image from the last convolutional layer of the convolutional neural network model as a second feature map.
And then, performing difference on the first characteristic diagram and the second characteristic diagram to obtain a difference value matrix. In the difference matrix, the point with a large absolute value is the characteristic response point sensitive to the shielded area of the human face.
And finally, performing binarization processing on the elements in the difference matrix by selecting a proper preset threshold value to obtain an initial characteristic mask. The binarization processing mode is as follows: when the absolute value of the element is greater than a preset threshold value, the corresponding position is 0; and when the absolute value of the element is smaller than the preset threshold value, the corresponding position is 1.
According to the method for identifying the shielded face based on the feature transformation, provided by the embodiment of the invention, the feature transformation mode of adding the feature mask is adopted for the image of the shielded face, the feature response of the common easily shielded area of the face is abandoned, the engineering is easier to realize, the calculation time is shorter, the network structure is simpler, and the identification precision of the shielded face is improved.
On the basis of the above embodiments, further, the target data set is an LFW data set.
Specifically, the LFW data set is a face data set in an unlimited environment, and is widely used for training face recognition. The target data set adopts an LFW data set, so that the acquired feature mask is more accurate.
According to the method for identifying the shielded face based on the feature transformation, provided by the embodiment of the invention, the feature transformation mode of adding the feature mask is adopted for the image of the shielded face, the feature response of the common easily shielded area of the face is abandoned, the engineering is easier to realize, the calculation time is shorter, the network structure is simpler, and the identification precision of the shielded face is improved.
On the basis of the above embodiments, further, the convolutional neural network model is a response-18 network.
In particular, the event-18 network is a simple and efficient convolutional neural network model. According to the embodiment of the invention, the residual-18 network is used as a convolutional neural network model, so that the accuracy of face recognition is improved.
According to the method for identifying the shielded face based on the feature transformation, provided by the embodiment of the invention, the feature transformation mode of adding the feature mask is adopted for the image of the shielded face, the feature response of the common easily shielded area of the face is abandoned, the engineering is easier to realize, the calculation time is shorter, the network structure is simpler, and the identification precision of the shielded face is improved.
Fig. 2 is a schematic diagram of an occluded face recognition device based on feature transformation according to an embodiment of the present invention, and as shown in fig. 2, an occluded face recognition device based on feature transformation according to an embodiment of the present invention is used for executing the method described in any of the above embodiments, and specifically includes a feature map extraction module 201, a calculation module 202, and a feature acquisition module 203, where:
the feature map extraction module 201 is configured to input a target face image into a preset convolutional neural network model, and obtain a feature map of the target face image from a last convolutional layer of the convolutional neural network model; the calculation module 202 is configured to calculate a fractional-pel product of the feature map and a feature mask obtained in advance; the feature obtaining module 203 is configured to input the pixel product to a feature layer of the convolutional neural network model, and output a feature of the target face image for recognition.
Specifically, the process of performing face recognition by the feature transformation-based occluded face recognition device provided by the embodiment of the present invention includes two stages: the first, experiment stage, the second, application stage.
In the experimental phase, a feature mask is obtained from the experiment.
In the application stage, firstly, the feature map extraction module 201 inputs the target face image into a preset convolutional neural network model, and obtains the feature map of the target face image from the last convolutional layer of the convolutional neural network model.
Before inputting the feature map of the target face image into the feature layer, the computation module 202 computes the fractional product of the feature map of the target face image and the pre-acquired feature mask. The feature mask is acquired in advance at an experimental stage.
Finally, the calculated sub-pixel product is input to a feature layer of the convolutional neural network model as a transformed feature map through the feature acquisition module 203, and the feature used for recognition of the target face image is output so as to be used for recognizing the target face image according to the feature.
The device for recognizing the face to be shielded based on the feature transformation, which is provided by the embodiment of the invention, adopts the feature transformation mode of adding the feature mask to the image of the face to be shielded, abandons the feature response of the common easily-shielded area of the face, is easier to realize engineering, has shorter calculation time and simpler network structure, and improves the recognition precision of the face to be shielded.
On the basis of the above embodiment, further, the apparatus further includes a feature mask acquiring module, configured to acquire the feature mask.
Specifically, the method for recognizing an occluded face based on feature transformation provided by the embodiment of the present invention includes two stages: the first, experiment stage, the second, application stage.
In the experimental stage, the feature mask is obtained through the feature mask obtaining module according to the experiment.
The specific steps for obtaining the feature mask are as follows:
firstly, acquiring a plurality of original sample images in a target data set, aligning the original sample images according to human face characteristic points aiming at each original sample image, adding artificial shielding to the human face image according to the characteristic points, acquiring an image obtained after the artificial shielding is added to the original sample image, taking the original sample image obtained after the artificial shielding is added as a shielding sample image, and enabling one original sample image to correspond to one shielding sample image.
Then, an initial feature mask is obtained according to a pair of the original sample image and the occlusion sample image.
Finally, the average of all the initial feature masks is taken as the final feature mask.
The device for recognizing the face to be shielded based on the feature transformation, which is provided by the embodiment of the invention, adopts the feature transformation mode of adding the feature mask to the image of the face to be shielded, abandons the feature response of the common easily-shielded area of the face, is easier to realize engineering, has shorter calculation time and simpler network structure, and improves the recognition precision of the face to be shielded.
Fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention, and as shown in fig. 3, the electronic device includes: a processor 301, a memory 302, and a bus 303;
the processor 301 and the memory 302 complete communication with each other through the bus 303;
inputting a target face image into a preset convolutional neural network model, and acquiring a feature map of the target face image from the last convolutional layer of the convolutional neural network model;
calculating the pixel product of the feature map and a feature mask acquired in advance;
and inputting the sub-pixel product to a feature layer of the convolutional neural network model, and outputting the feature of the target face image for recognition.
Embodiments of the present invention provide a computer program product comprising a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions that, when executed by a computer, enable the computer to perform the methods provided by the above-mentioned method embodiments, for example, including:
inputting a target face image into a preset convolutional neural network model, and acquiring a feature map of the target face image from the last convolutional layer of the convolutional neural network model;
calculating the pixel product of the feature map and a feature mask acquired in advance;
and inputting the sub-pixel product to a feature layer of the convolutional neural network model, and outputting the feature of the target face image for recognition.
Embodiments of the present invention provide a non-transitory computer-readable storage medium, which stores computer instructions, where the computer instructions cause the computer to perform the methods provided by the above method embodiments, for example, the methods include:
inputting a target face image into a preset convolutional neural network model, and acquiring a feature map of the target face image from the last convolutional layer of the convolutional neural network model;
calculating the pixel product of the feature map and a feature mask acquired in advance;
and inputting the sub-pixel product to a feature layer of the convolutional neural network model, and outputting the feature of the target face image for recognition.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The above-described embodiments of the apparatuses and devices are merely illustrative, where the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (8)
1. An occluded face recognition method based on feature transformation is characterized by comprising the following steps:
inputting a target face image into a preset convolutional neural network model, and acquiring a feature map of the target face image from the last convolutional layer of the convolutional neural network model;
calculating the pixel product of the feature map and a feature mask acquired in advance;
inputting the sub-pixel product to a feature layer of the convolutional neural network model, and outputting the feature of the target face image for recognition;
the specific steps for obtaining the feature mask are as follows:
acquiring a plurality of original sample images in a target data set, adding an image subjected to artificial shielding to the original sample images, taking the original sample images subjected to the artificial shielding as shielding sample images, wherein one original sample image corresponds to one shielding sample image;
acquiring an initial characteristic mask according to a pair of original sample images and a shielding sample image;
the average of all the initial feature masks is taken as the final feature mask.
2. The method according to claim 1, wherein obtaining an initial feature mask from a pair of original sample images and an occlusion sample image comprises:
inputting the target original sample image into the convolutional neural network model aiming at the target original sample image, and acquiring a feature map of the target original sample image from the last convolutional layer of the convolutional neural network model as a first feature map; inputting a target shielding sample image corresponding to the target original sample image into the convolutional neural network model, and acquiring a feature map of the target shielding sample image from the last convolutional layer of the convolutional neural network model as a second feature map;
performing difference on the first characteristic diagram and the second characteristic diagram to obtain a difference value matrix;
and carrying out binarization processing on the elements in the difference value matrix to obtain an initial characteristic mask.
3. The method of claim 1, wherein the target data set is an LFW data set.
4. The method of any one of claims 1-2, wherein the convolutional neural network model is a response-18 network.
5. An occluded face recognition device based on feature transformation, comprising:
the characteristic diagram extraction module is used for inputting a target face image into a preset convolutional neural network model and acquiring a characteristic diagram of the target face image from the last convolutional layer of the convolutional neural network model;
the calculation module is used for calculating the fractional pixel product of the feature map and a feature mask acquired in advance;
the feature acquisition module is used for inputting the sub-pixel product to a feature layer of the convolutional neural network model and outputting the feature of the target face image for identification;
the specific steps for obtaining the feature mask are as follows:
acquiring a plurality of original sample images in a target data set, adding an image subjected to artificial shielding to the original sample images, taking the original sample images subjected to the artificial shielding as shielding sample images, wherein one original sample image corresponds to one shielding sample image;
acquiring an initial characteristic mask according to a pair of original sample images and a shielding sample image;
the average of all the initial feature masks is taken as the final feature mask.
6. The apparatus of claim 5, further comprising a feature mask acquisition module to acquire the feature mask.
7. An electronic device, comprising:
the processor and the memory are communicated with each other through a bus; the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the method of any of claims 1 to 4.
8. A non-transitory computer-readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910006884.4A CN109840477B (en) | 2019-01-04 | 2019-01-04 | Method and device for recognizing shielded face based on feature transformation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910006884.4A CN109840477B (en) | 2019-01-04 | 2019-01-04 | Method and device for recognizing shielded face based on feature transformation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109840477A CN109840477A (en) | 2019-06-04 |
| CN109840477B true CN109840477B (en) | 2020-11-24 |
Family
ID=66883669
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910006884.4A Active CN109840477B (en) | 2019-01-04 | 2019-01-04 | Method and device for recognizing shielded face based on feature transformation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109840477B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110287930B (en) * | 2019-07-01 | 2021-08-20 | 厦门美图之家科技有限公司 | Wrinkle classification model training method and device |
| CN110647993A (en) * | 2019-09-23 | 2020-01-03 | 南方科技大学 | Infrared sensor mask manufacturing method, device and system and storage medium |
| CN110909595B (en) * | 2019-10-12 | 2023-04-18 | 平安科技（深圳）有限公司 | Facial motion recognition model training method and facial motion recognition method |
| CN111461959B (en) * | 2020-02-17 | 2023-04-25 | 浙江大学 | Face emotion synthesis method and device |
| CN111414879B (en) * | 2020-03-26 | 2023-06-09 | 抖音视界有限公司 | Face shielding degree identification method and device, electronic equipment and readable storage medium |
| CN111476200B (en) * | 2020-04-27 | 2022-04-19 | 华东师范大学 | Face de-identification generation method based on generation of confrontation network |
| CN111680597B (en) * | 2020-05-29 | 2023-09-01 | 北京百度网讯科技有限公司 | Face recognition model processing method, device, equipment and storage medium |
| CN111898413A (en) * | 2020-06-16 | 2020-11-06 | 深圳市雄帝科技股份有限公司 | Face recognition method, face recognition device, electronic equipment and medium |
| CN111931862B (en) * | 2020-09-11 | 2021-07-23 | 杭州追猎科技有限公司 | Method and system for detecting illegal posted advertisements and electronic equipment |
| CN112149605B (en) * | 2020-09-30 | 2023-04-18 | 济南博观智能科技有限公司 | Face recognition method, device, equipment and storage medium |
| CN112101302B (en) * | 2020-11-05 | 2021-04-27 | 杭州追猎科技有限公司 | Illegal poster detection method and system and electronic equipment |
| CN112633144A (en) * | 2020-12-21 | 2021-04-09 | 平安科技（深圳）有限公司 | Face occlusion detection method, system, device and storage medium |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105095856B (en) * | 2015-06-26 | 2019-03-22 | 上海交通大学 | Face identification method is blocked based on mask |
| CN106203331B (en) * | 2016-07-08 | 2019-05-17 | 苏州平江历史街区保护整治有限责任公司 | A kind of crowd density evaluation method based on convolutional neural networks |
| CN106372595A (en) * | 2016-08-31 | 2017-02-01 | 重庆大学 | Shielded face identification method and device |
| CN108038435B (en) * | 2017-12-04 | 2022-01-04 | 中山大学 | Feature extraction and target tracking method based on convolutional neural network |
| CN108509915B (en) * | 2018-04-03 | 2021-10-26 | 百度在线网络技术（北京）有限公司 | Method and device for generating face recognition model |
-
2019
- 2019-01-04 CN CN201910006884.4A patent/CN109840477B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109840477A (en) | 2019-06-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109840477B (en) | Method and device for recognizing shielded face based on feature transformation | |
| US11120254B2 (en) | Methods and apparatuses for determining hand three-dimensional data | |
| CN108122234B (en) | Convolutional neural network training and video processing method and device and electronic equipment | |
| US11380017B2 (en) | Dual-view angle image calibration method and apparatus, storage medium and electronic device | |
| CN114186632B (en) | Method, device, equipment and storage medium for training key point detection model | |
| CN108229301B (en) | Eyelid line detection method and device and electronic equipment | |
| CN111340077B (en) | Attention mechanism-based disparity map acquisition method and device | |
| CN109948441B (en) | Model training method, image processing method, device, electronic equipment and computer readable storage medium | |
| CN110969046B (en) | Face recognition method, face recognition device and computer-readable storage medium | |
| CN111860398A (en) | Remote sensing image target detection method and system and terminal equipment | |
| CN112200056B (en) | Face living body detection method and device, electronic equipment and storage medium | |
| CN107578053B (en) | Contour extraction method and device, computer device and readable storage medium | |
| CN114758145A (en) | Image desensitization method and device, electronic equipment and storage medium | |
| CN113378864B (en) | Method, device and equipment for determining anchor frame parameters and readable storage medium | |
| CN113255700B (en) | Image feature map processing method and device, storage medium and terminal | |
| CN112990046B (en) | Differential information acquisition method, related device and computer program product | |
| CN110287943B (en) | Image object recognition method and device, electronic equipment and storage medium | |
| CN112288748B (en) | Semantic segmentation network training and image semantic segmentation method and device | |
| CN113344200A (en) | Method for training separable convolutional network, road side equipment and cloud control platform | |
| CN113724176A (en) | Multi-camera motion capture seamless connection method, device, terminal and medium | |
| CN109816709B (en) | Monocular camera-based depth estimation method, device and equipment | |
| CN113052827B (en) | Crowd counting method and system based on multi-branch expansion convolutional neural network | |
| CN114119609B (en) | Method, device and equipment for detecting image stain concentration and storage medium | |
| CN111091593B (en) | Image processing method, device, electronic equipment and storage medium | |
| CN110969651B (en) | 3D depth of field estimation method and device and terminal equipment |
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 |