CN108875489A - Method for detecting human face, device, system, storage medium and capture machine - Google Patents
Method for detecting human face, device, system, storage medium and capture machine Download PDFInfo
- Publication number
- CN108875489A CN108875489A CN201710915035.1A CN201710915035A CN108875489A CN 108875489 A CN108875489 A CN 108875489A CN 201710915035 A CN201710915035 A CN 201710915035A CN 108875489 A CN108875489 A CN 108875489A
- Authority
- CN
- China
- Prior art keywords
- face
- detection
- human face
- image
- neural network
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
-
- 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
- G06V40/166—Detection; Localisation; Normalisation using acquisition arrangements
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The present invention provides a kind of method for detecting human face, device, system, storage medium and capture machine, the method for detecting human face includes：Face location detection is carried out using image of the multichannel neural network to acquisition and face character detects；And output has the image of face location testing result and face character testing result.Method for detecting human face, device, system, storage medium and capture machine according to an embodiment of the present invention are only with neural network detection face location and its face character, without cascading multiple neural networks, can simple flow, reduce capture it is time-consuming.
Description
Technical field
The present invention relates to human face detection tech field, relates more specifically to a kind of method for detecting human face, device, system and deposit
Storage media and capture machine.
Background technique
Existing capture machine method for detecting human face is to first pass through a detection neural network, is led to again after face is plucked out figure
It crosses one or several attribute neural networks and obtains face character.This method needs to cascade several neural networks, not only process
Complexity, and time-consuming greatly increases.Also, plucking out the picture quality come only by a detection neural network can not obtain
Guarantee.
Summary of the invention
At least one of to solve the above-mentioned problems, the invention proposes a kind of schemes about Face datection, can
To be used for capture machine Face datection, the Face datection of other scenes can be used for.Further it is proposed that about face examine
The scheme of survey can also extend to the detection for any target object, and face need to only be replaced with to other target objects.
The scheme proposed by the present invention about Face datection is briefly described below, more details will be embodied in subsequent combination attached drawing
It is described in mode.
According to an aspect of the present invention, a kind of method for detecting human face is provided, the method for detecting human face includes：Utilize one
Multichannel neural network carries out face location detection to the image of acquisition and face character detects；And output has face location inspection
Survey the image of result and face character testing result.
In one embodiment of the invention, the multichannel neural network includes common sparing and packet partial, described total
With part for pre-processing to the image of the acquisition, the first grouping in the packet partial is for detecting face position
It sets, remaining grouping in the packet partial is respectively used to detect various face characters.
In one embodiment of the invention, the common sparing includes each group shared by the packet partial
Convolutional layer, the packet partial include the convolutional layer respectively used by each grouping of the packet partial.
In one embodiment of the invention, the training of the multichannel neural network is based on multi-model distillating method.
In one embodiment of the invention, the setting of the loss of the multichannel neural network is based on the face location
The weighted average of the loss of detection and the loss of face character detection.
In one embodiment of the invention, the multichannel neural network is realized on programmable gate array at the scene.
In one embodiment of the invention, the face character detection includes at least one of the following：Human face posture
Detection and facial image fuzzy detection.
In one embodiment of the invention, the image of the acquisition is continuous video frame, and the output includes：
A frame or multiple image in the video frame with best detection result are exported to be used for recognition of face.
According to a further aspect of the invention, a kind of human face detection device is provided, the human face detection device includes：Detect mould
Block, for carrying out face location detection and face character detection using image of the multichannel neural network to acquisition；And it is defeated
Module out, for exporting the image with face location testing result and face character testing result.
In one embodiment of the invention, the multichannel neural network includes common sparing and packet partial, described total
With part for pre-processing to the image of the acquisition, the first grouping in the packet partial is for detecting face position
It sets, remaining grouping in the packet partial is respectively used to detect various face characters.
In one embodiment of the invention, the common sparing includes each group shared by the packet partial
Convolutional layer, the packet partial include the convolutional layer respectively used by each grouping of the packet partial.
In one embodiment of the invention, the training of the multichannel neural network is based on multi-model distillating method.
In one embodiment of the invention, the setting of the loss of the multichannel neural network is based on the face location
The weighted average of the loss of detection and the loss of face character detection.
In one embodiment of the invention, the multichannel neural network is realized on programmable gate array at the scene.
In one embodiment of the invention, the face character detection includes at least one of the following：Human face posture
Detection and facial image fuzzy detection.
In one embodiment of the invention, the image of the acquisition is continuous video frame, and the output module
The frame or multiple image for being further used for exporting in the video frame with best detection result are for recognition of face.
Another aspect according to the present invention, provides a kind of face detection system, and the face detection system includes storage dress
It sets and processor, is stored with the computer program run by the processor on the storage device, the computer program exists
Method for detecting human face described in any of the above embodiments is executed when being run by the processor.
According to a further aspect of the present invention, a kind of storage medium is provided, is stored with computer program on the storage medium,
The computer program executes method for detecting human face described in any of the above embodiments at runtime.
According to a further aspect of the present invention, a kind of capture machine is provided, the capture machine includes image collecting device and above-mentioned
Described in any item human face detection devices.
According to a further aspect of the present invention, a kind of capture machine is provided, the capture machine includes field programmable gate array, institute
It states and realizes method for detecting human face as described in any one of the above embodiments on field programmable gate array.
Method for detecting human face, device, system, storage medium and capture machine according to an embodiment of the present invention are only with a mind
Detect face location and its face character through network, without cascading multiple neural networks, can simple flow, reduce and capture consumption
When.
Detailed description of the invention
The embodiment of the present invention is described in more detail in conjunction with the accompanying drawings, the above and other purposes of the present invention,
Feature and advantage will be apparent.Attached drawing is used to provide to further understand the embodiment of the present invention, and constitutes explanation
A part of book, is used to explain the present invention together with the embodiment of the present invention, is not construed as limiting the invention.In the accompanying drawings,
Identical reference label typically represents same parts or step.
Fig. 1 shows showing for realizing method for detecting human face according to an embodiment of the present invention, device, system and storage medium
The schematic block diagram of example electronic equipment；
Fig. 2 shows the schematic flow charts of method for detecting human face according to an embodiment of the present invention；
Fig. 3 shows the schematic structure for the multichannel neural network that method for detecting human face according to an embodiment of the present invention uses
Figure；
Fig. 4 shows the schematic diagram for pushing away figure strategy of method for detecting human face according to an embodiment of the present invention；
Fig. 5 shows the schematic block diagram of human face detection device according to an embodiment of the present invention；And
Fig. 6 shows the schematic block diagram of face detection system according to an embodiment of the present invention.
Specific embodiment
In order to enable the object, technical solutions and advantages of the present invention become apparent, root is described in detail below with reference to accompanying drawings
According to example embodiments of the present invention.Obviously, described embodiment is only a part of the embodiments of the present invention, rather than this hair
Bright whole embodiments, it should be appreciated that the present invention is not limited by example embodiment described herein.Based on described in the present invention
The embodiment of the present invention, those skilled in the art's obtained all other embodiment in the case where not making the creative labor
It should all fall under the scope of the present invention.
Firstly, describing the method for detecting human face for realizing the embodiment of the present invention, device, system and storage referring to Fig.1
The exemplary electronic device 100 of medium.
As shown in Figure 1, electronic equipment 100 include one or more processors 102, it is one or more storage device 104, defeated
Enter device 106, output device 108 and image collecting device 110, these components pass through bus system 112 and/or other forms
Bindiny mechanism's (not shown) interconnection.It should be noted that the component and structure of electronic equipment 100 shown in FIG. 1 are only exemplary,
And not restrictive, as needed, the electronic equipment also can have other assemblies and structure.
The processor 102 can be central processing unit (CPU) or have data-handling capacity and/or instruction execution
The processing unit of the other forms of ability, and the other components that can control in the electronic equipment 100 are desired to execute
Function.
The storage device 104 may include one or more computer program products, and the computer program product can
To include various forms of computer readable storage mediums, such as volatile memory and/or nonvolatile memory.It is described easy
The property lost memory for example may include random access memory (RAM) and/or cache memory (cache) etc..It is described non-
Volatile memory for example may include read-only memory (ROM), hard disk, flash memory etc..In the computer readable storage medium
On can store one or more computer program instructions, processor 102 can run described program instruction, to realize hereafter institute
The client functionality (realized by processor) in the embodiment of the present invention stated and/or other desired functions.In the meter
Can also store various application programs and various data in calculation machine readable storage medium storing program for executing, for example, the application program use and/or
The various data etc. generated.
The input unit 106 can be the device that user is used to input instruction, and may include keyboard, mouse, wheat
One or more of gram wind and touch screen etc..
The output device 108 can export various information (such as image or sound) to external (such as user), and
It may include one or more of display, loudspeaker etc..
Described image acquisition device 110 can acquire the desired image of user (such as photo, video etc.), and will be adopted
The image of collection is stored in the storage device 104 for the use of other components.Image collecting device 110 can be camera.
Illustratively, the exemplary electronic device for realizing method for detecting human face according to an embodiment of the present invention and device can
To be implemented as capture machine, smart phone, tablet computer etc..
In the following, method for detecting human face 200 according to an embodiment of the present invention will be described with reference to Fig. 2.As shown in Fig. 2, face is examined
Survey method 200 may include steps of：
In step S210, face location detection and face character are carried out using image of the multichannel neural network to acquisition
Detection.
In one embodiment, acquired image for example can be include the image data of a frame image, or can be with
It is the video data for including multiple image.
In an embodiment of the present invention, face location detection and people are carried out merely with image of the neural network to acquisition
Face detection of attribute.There are sequencing, serially implementation face location detection using cascade multiple neural networks with existing
It is compared with face character detection, the method for detecting human face of the embodiment of the present invention is to make face location only with a neural network
Detection and face character detection are concurrently implemented.In addition, if to detect more than one face character (such as human face posture inspection
Survey, fuzzy detection etc.), then between the detection of every kind of face character and they and face location detect between be parallel practice
's.
Due to carrying out face location detection and face character detection only with image of the neural network to acquisition,
It can simplify process, it is time-consuming to be substantially reduced candid photograph.
In one embodiment, the neural network that step S210 is used is properly termed as multichannel neural network, because it can be with
Including multiple channels, certain several channel can be separated after the completion of some convolutional layer operation and is respectively intended to train Face datection and people
The detection of attribute such as face posture, fuzziness.
In one embodiment, above-mentioned multichannel neural network may include common sparing and packet partial.Wherein, described total
With partially including being located in advance by the convolutional layer of each group shared of the packet partial for the image to the acquisition
Reason.The packet partial includes the convolutional layer respectively used by each grouping of the packet partial, and the first grouping therein can
With for detecting face location, remaining grouping may be respectively used for detecting various face characters.It can understand in conjunction with Fig. 3 according to this
The structure for the multichannel neural network that the method for detecting human face of inventive embodiments uses.
Fig. 3 shows the schematic structure for the multichannel neural network that method for detecting human face according to an embodiment of the present invention uses
Figure.As shown in figure 3, multichannel neural network 300 includes common sparing 310 and packet partial 320.Input picture initially enters shared
Part 310 is pre-processed, common sparing 310 include by each group shared of packet partial 320 convolutional layer (covn1,
conv2,……).Packet partial 320 includes the convolutional layer respectively used by each grouping of the packet partial
(convN ...), export respectively face location testing result, attribute 1, attribute 2 ..., attribute N, wherein N is natural number.
As seen from Figure 3, it is based on multichannel neural network 300, makes it possible to carry out face location inspection to the image of input
It surveys and face character detects, avoid complicated cascade, reduce the detection processing time.
In one embodiment, the training of above-mentioned multichannel neural network can be marked using machine mark or manually by people
Face position detection and each road face character, which all mark out, to be come.Illustratively, it can be trained using multi-model distillating method,
Face location can be detected and be marked together with each road face character, trained together, the trained optimal mould of use
Type (face location, posture, fuzziness etc.) is used as training data to generate annotation results jointly.
In one embodiment, the setting of the loss (Loss) of above-mentioned multichannel neural network can be detected based on face location
Loss and face character detection (such as each passerby's face detection of attribute) loss weighted average.
In one embodiment, above-mentioned multichannel neural network can be in central processing unit (CPU), graphics processing unit
(GPU) etc. it is realized in kinds of platform.Illustratively, multichannel neural network used by step S210 can programmable gate at the scene
It is realized on array (FPGA).FPGA implementation has low bit characteristic, GPU, CPU are needed compared with other implementations
Wanting floating-point operation or 32,64 integer arithmetics, FPGA can carry out the training and deployment of more accuracy models, such as 2 bits,
4 bits, 8 bits.Specifically, 2 bit accuracies are lower but fast speed power consumption is lower, 8 bit accuracies are higher but speed compared with
Slow power consumption is higher.Therefore, it can according to need and be configured on different accuracy, so that realizes on FPGA is aforementioned more
Road neural network has the advantages that power is low, fireballing.
Now referring back to Fig. 2, the subsequent step of method for detecting human face 200 according to an embodiment of the present invention is continued to describe.
In step S220, the image with face location testing result and face character testing result is exported.
It is detected based on the step S210 face location detection implemented using a multichannel neural network concurrent and face character,
The image after detection processing can be exported, which has face location testing result and face character testing result simultaneously.
In one embodiment, the image of acquisition handled by step S210 is continuous video frame, this is based on, by step
The processing of rapid S210, can obtain image of the multiframe through detection processing, and every frame image has the identifier ID and the frame of the frame
The face location testing result and face character testing result of image.The face location testing result and face character of every frame image
Testing result may be different.At this point, can only the best frame of output test result or multiple image be used in step S220
Subsequent recognition of face not only can reduce data transmission, can also improve subsequent accuracy of identification.In addition, using
When capture machine carries out above-mentioned Face datection, the best frame of output test result or multiple image can guarantee that capture machine pushes away the matter of figure
Amount.
In one example, the best frame of testing result or multiple image can be understood as in face being effective face
In the case of the optimal frame of human face posture (pose) or multiple image.Illustratively, effective face can be face attribute value mould
Paste the case where angle value has been more than predetermined threshold.Illustratively, human face posture can most preferably refer to face at three of three-dimensional space
(being, for example, less than equal to 10 degree) quadratic sum is minimum in the reasonable scope for angle value (roll, pitch, jaw).
It can understand that step S220's under above situation pushes away figure strategy in conjunction with Fig. 4.A secondary tracking is shown referring to Fig. 4
(track) schematic diagram for pushing away figure strategy in.Track, which refers to, carries out Face datection in one section of continuous video, can
It is continuously detected one section of face.It pushes away figure to refer to, many frame facial images is able to detect that in a track, but are used for
Only best a frame or a few frames therein for subsequent recognition of face, it is therefore desirable to be released for identification from a track
One frame or a few frames.Horizontal axis is frame number in the schematic diagram, and the longitudinal axis is the attribute value (abbreviation values of ambiguity) of fuzziness, and T is fuzziness
Low pass threshold value, indicate only have every frame in face character value values of ambiguity be more than that this threshold value is just denoted as effective face.Every
In the case that secondary track has broken, can release one values of ambiguity effectively (being more than above-mentioned threshold value T) in the case where posture it is optimal
Picture (frame referred to such as the small circle in Fig. 4), to guarantee that capture machine pushes away the quality of figure.
Based on above description, method for detecting human face according to an embodiment of the present invention detects people only with a neural network
Face position and its face character, without cascading multiple neural networks, can simple flow, reduce and capture time-consuming, guarantee that face is scratched
Plot quality.
Method for detecting human face according to an embodiment of the present invention is described above exemplarily.Illustratively, according to the present invention
The method for detecting human face of embodiment can with memory and processor unit or system in realize.
In addition, method for detecting human face processing speed according to an embodiment of the present invention is fast, it is deployed to capture machine with can be convenient
On, it is deployed in the mobile devices such as smart phone, tablet computer, personal computer with also can be convenient.Alternatively, according to this hair
The method for detecting human face of bright embodiment can also be deployed in server end (or cloud).Alternatively, according to an embodiment of the present invention
Method for detecting human face can also be deployed at server end (or cloud) and personal terminal with being distributed.
The human face detection device of another aspect of the present invention offer is described below with reference to Fig. 5.Fig. 5 shows real according to the present invention
Apply the schematic block diagram of the human face detection device 500 of example.
As shown in figure 5, human face detection device 500 according to an embodiment of the present invention includes detection module 510 and output module
520.The modules can execute each step/function of the method for detecting human face above in conjunction with Fig. 2 description respectively.Below
Only the major function of each module of human face detection device 500 is described, and omits the detail content having been described above.
Detection module 510 is used to carry out face location detection and face using image of the multichannel neural network to acquisition
Detection of attribute.Output module 520 is used to export the image with face location testing result and face character testing result.Detection
Module 510 and output module 520 can be deposited in 102 Running storage device 104 of processor in electronic equipment as shown in Figure 1
The program instruction of storage is realized.
In one embodiment, acquired image for example can be include the image data of a frame image, or can be with
It is the video data for including multiple image.
In an embodiment of the present invention, detection module 510 carries out face merely with image of the neural network to acquisition
Position detection and face character detection.With it is existing using cascade multiple neural networks have sequencing, serially implement
Face location detection is compared with face character detection, and the detection module 510 of the human face detection device of the embodiment of the present invention is only to adopt
With a neural network face location detection and face character detection are concurrently implemented.In addition, if to detect more than one
Kind of face character (such as human face posture detection, fuzzy detection etc.), then between every kind of face character detects and they and face
It is parallel practice between position detection.
Since detection module 510 carries out face location detection and face category only with image of the neural network to acquisition
Property detection, therefore can simplify process, it is time-consuming to be substantially reduced candid photograph.
In one embodiment, the neural network that detection module 510 uses is properly termed as multichannel neural network, because it can
To include multiple channels, can be separated after the completion of some convolutional layer operation certain several channel be respectively intended to train Face datection and
The detection of attribute such as human face posture, fuzziness.
In one embodiment, above-mentioned multichannel neural network may include common sparing and packet partial.Wherein, described total
With partially including being located in advance by the convolutional layer of each group shared of the packet partial for the image to the acquisition
Reason.The packet partial includes the convolutional layer respectively used by each grouping of the packet partial, and the first grouping therein can
With for detecting face location, remaining grouping may be respectively used for detecting various face characters.It can combine with reference to Fig. 3 and close above
The exemplary structure for the multichannel neural network that method for detecting human face according to an embodiment of the present invention uses is understood in the description of Fig. 3,
For sake of simplicity, details are not described herein again.
In one embodiment, the training of above-mentioned multichannel neural network can be marked using machine mark or manually by people
Face position detection and each road face character, which all mark out, to be come.Illustratively, it can be trained using multi-model distillating method,
Face location can be detected and be marked together with each road face character, trained together, the trained optimal mould of use
Type (face location, posture, fuzziness etc.) is used as training data to generate annotation results jointly.
In one embodiment, the setting of the loss (Loss) of above-mentioned multichannel neural network can be detected based on face location
Loss and face character detection (such as each passerby's face detection of attribute) loss weighted average.
In one embodiment, above-mentioned multichannel neural network can be in central processing unit (CPU), graphics processing unit
(GPU) etc. it is realized in kinds of platform.Illustratively, multichannel neural network used by detection module 510 can be compiled at the scene
It is realized in journey gate array (FPGA).FPGA implementation has low bit characteristic, GPU, CPU compared with other implementations
Floating-point operation or 32,64 integer arithmetics are all obtained, FPGA can carry out the training and deployment of more accuracy models, such as 2 ratios
Spy, 4 bits, 8 bits.Specifically, 2 bit accuracies are lower but fast speed power consumption is lower, and 8 bit accuracies are higher but speed
Slower power consumption is higher.Therefore, it can according to need and be configured on different accuracy, so that realizes on FPGA is aforementioned
Multichannel neural network has the advantages that power is low, fireballing.
In one embodiment, the image of acquisition handled by detection module 510 is continuous video frame, is based on this, warp
The processing of detection module 510 is crossed, multiple image can be obtained, identifier ID and the frame image of every frame image with the frame
Face location testing result and face character testing result.The face location testing result and face character of every frame image detect knot
Fruit may be different.At this point, output module 520 only the best frame of output test result or multiple image can be used for subsequent people
Face identification.
In one example, the best frame of testing result or multiple image can be understood as in face being effective face
In the case of the optimal frame of human face posture or multiple image.Illustratively, effective face can be face attribute value values of ambiguity
The case where being more than predetermined threshold.Illustratively, human face posture can most preferably refer to face in three angle values of three-dimensional space
(roll, pitch, jaw) (being, for example, less than equal to 10 degree) quadratic sum is minimum in the reasonable scope.
Being referred to Fig. 4 combines what the description previously for Fig. 4 understood output module 520 under above situation to push away figure strategy,
For sake of simplicity, details are not described herein again.It is above-mentioned to push away figure strategy not only and reduce data transmission, it is smart that subsequent identification can also be improved
Degree.In addition, the certifiable quality for pushing away figure of a frame or multiple image that output test result is best.
Based on above description, human face detection device according to an embodiment of the present invention detects people only with a neural network
Face position and its face character, without cascading multiple neural networks, can simple flow, reduce and capture time-consuming, guarantee that face is scratched
Plot quality.
Fig. 6 shows the schematic block diagram of face detection system 600 according to an embodiment of the present invention.Face detection system
600 include storage device 610 and processor 620.
Wherein, the storage of storage device 610 is for realizing the corresponding step in method for detecting human face according to an embodiment of the present invention
Rapid program code.Program code of the processor 620 for being stored in Running storage device 610, it is real according to the present invention to execute
The corresponding steps of the method for detecting human face of example are applied, and for realizing the phase in human face detection device according to an embodiment of the present invention
Answer module.
In one embodiment, when said program code is run by processor 620 face detection system 600 is executed
Following steps：Face location detection is carried out using image of the multichannel neural network to acquisition and face character detects；And
Export the image with face location testing result and face character testing result.
In one embodiment, the multichannel neural network includes common sparing and packet partial, and the common sparing is used
It being pre-processed in the image of the acquisition, the first grouping in the packet partial is for detecting face location, and described point
Remaining grouping in group part is respectively used to detect various face characters.
In one embodiment, the common sparing include by the convolutional layer of each group shared of the packet partial,
The packet partial includes the convolutional layer respectively used by each grouping of the packet partial.
In one embodiment, the training of the multichannel neural network is based on multi-model distillating method.
In one embodiment, the setting of the loss of the multichannel neural network is the damage based on face location detection
Lose the weighted average with the loss of face character detection.
In one embodiment, the multichannel neural network is realized on programmable gate array at the scene.
In one embodiment, the face character detection includes at least one of the following：Human face posture detection and
Facial image fuzzy detection.
In one embodiment, the image of the acquisition is continuous video frame, and the output includes：Described in output
A frame or multiple image in video frame with best detection result is to be used for recognition of face.
In addition, according to embodiments of the present invention, additionally providing a kind of storage medium, storing program on said storage
Instruction, when described program instruction is run by computer or processor for executing the method for detecting human face of the embodiment of the present invention
Corresponding steps, and for realizing the corresponding module in human face detection device according to an embodiment of the present invention.The storage medium
It such as may include the storage card of smart phone, the storage unit of tablet computer, the hard disk of personal computer, read-only memory
(ROM), Erasable Programmable Read Only Memory EPROM (EPROM), portable compact disc read-only memory (CD-ROM), USB storage,
Or any combination of above-mentioned storage medium.The computer readable storage medium can be one or more computer-readable deposit
Any combination of storage media, such as a computer readable storage medium include the figure using a multichannel neural network to acquisition
Computer-readable program code as carrying out face location detection and face character detection, another computer-readable storage medium
Matter includes the computer-readable program code that output has the image of face location testing result and face character testing result.
In one embodiment, the computer program instructions may be implemented real according to the present invention when being run by computer
Each functional module of the human face detection device of example is applied, and/or Face datection according to an embodiment of the present invention can be executed
Method.
In one embodiment, the computer program instructions make computer or place when being run by computer or processor
It manages device and executes following steps：Face location detection is carried out using image of the multichannel neural network to acquisition and face character is examined
It surveys；And output has the image of face location testing result and face character testing result.
In one embodiment, the multichannel neural network includes common sparing and packet partial, and the common sparing is used
It being pre-processed in the image of the acquisition, the first grouping in the packet partial is for detecting face location, and described point
Remaining grouping in group part is respectively used to detect various face characters.
In one embodiment, the common sparing include by the convolutional layer of each group shared of the packet partial,
The packet partial includes the convolutional layer respectively used by each grouping of the packet partial.
In one embodiment, the training of the multichannel neural network is based on multi-model distillating method.
In one embodiment, the setting of the loss of the multichannel neural network is the damage based on face location detection
Lose the weighted average with the loss of face character detection.
In one embodiment, the multichannel neural network is realized on programmable gate array at the scene.
In one embodiment, the face character detection includes at least one of the following：Human face posture detection and
Facial image fuzzy detection.
In one embodiment, the image of the acquisition is continuous video frame, and the output includes：Described in output
A frame or multiple image in video frame with best detection result is to be used for recognition of face.
Each module in human face detection device according to an embodiment of the present invention can pass through people according to an embodiment of the present invention
The processor computer program instructions that store in memory of operation of the electronic equipment of face detection realize, or can be in root
The computer instruction stored in computer readable storage medium according to the computer program product of the embodiment of the present invention is by computer
It is realized when operation.
In addition, according to embodiments of the present invention, additionally providing a kind of capture machine, which may include image collecting device
And human face detection device.Wherein, which can be used for acquiring candid photograph image, which can be to figure
The image obtained as acquisition device carries out face location detection and face character detection.Wherein, which can adopt
It is realized, or can also be examined using previously in conjunction with face described in Fig. 6 with previously in conjunction with human face detection device 500 described in Fig. 5
Examining system 600 is realized.Those of ordinary skill in the art can understand according to the present invention in conjunction with the description previously for Fig. 5 or Fig. 6
The operation of human face detection device included by the capture machine of embodiment and the operation of the capture machine, for sake of simplicity, herein no longer
It repeats.
In addition, according to embodiments of the present invention, additionally providing a kind of capture machine, which may include FPGA, the FPGA
On method for detecting human face according to an embodiment of the present invention above-mentioned, such as method for detecting human face 200 may be implemented.This field is common
Technical staff can understand capture machine according to an embodiment of the present invention structurally and operationally in conjunction with the description previously for Fig. 2, be
Succinct, details are not described herein again.
Method for detecting human face, device, system, storage medium and capture machine according to an embodiment of the present invention are only with a mind
Detect face location and its face character through network, without cascading multiple neural networks, can simple flow, reduce and capture consumption
When, guarantee that face scratches plot quality.
Although describing example embodiment by reference to attached drawing here, it should be understood that above example embodiment are only exemplary
, and be not intended to limit the scope of the invention to this.Those of ordinary skill in the art can carry out various changes wherein
And modification, it is made without departing from the scope of the present invention and spiritual.All such changes and modifications are intended to be included in appended claims
Within required the scope of the present invention.
Those of ordinary skill in the art may be aware that list described in conjunction with the examples disclosed in the embodiments of the present disclosure
Member and algorithm steps can be realized with the combination of electronic hardware or computer software and electronic hardware.These functions are actually
It is implemented in hardware or software, the specific application and design constraint depending on technical solution.Professional technician
Each specific application can be used different methods to achieve the described function, but this realization is it is not considered that exceed
The scope of the present invention.
In several embodiments provided herein, it should be understood that disclosed device and method can pass through it
Its mode is realized.For example, apparatus embodiments described above are merely indicative, for example, the division of the unit, only
Only a kind of logical function partition, there may be another division manner in actual implementation, such as multiple units or components can be tied
Another equipment is closed or is desirably integrated into, or some features can be ignored or not executed.
In the instructions provided here, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be practiced without these specific details.In some instances, well known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this specification.
Similarly, it should be understood that in order to simplify the present invention and help to understand one or more of the various inventive aspects,
To in the description of exemplary embodiment of the present invention, each feature of the invention be grouped together into sometimes single embodiment, figure,
Or in descriptions thereof.However, the method for the invention should not be construed to reflect following intention：It is i.e. claimed
The present invention claims features more more than feature expressly recited in each claim.More precisely, such as corresponding power
As sharp claim reflects, inventive point is that the spy of all features less than some disclosed single embodiment can be used
Sign is to solve corresponding technical problem.Therefore, it then follows thus claims of specific embodiment are expressly incorporated in this specific
Embodiment, wherein each, the claims themselves are regarded as separate embodiments of the invention.
It will be understood to those skilled in the art that any combination pair can be used other than mutually exclusive between feature
All features disclosed in this specification (including adjoint claim, abstract and attached drawing) and so disclosed any method
Or all process or units of equipment are combined.Unless expressly stated otherwise, this specification (is wanted including adjoint right
Ask, make a summary and attached drawing) disclosed in each feature can be replaced with an alternative feature that provides the same, equivalent, or similar purpose.
In addition, it will be appreciated by those of skill in the art that although some embodiments described herein include other embodiments
In included certain features rather than other feature, but the combination of the feature of different embodiments mean it is of the invention
Within the scope of and form different embodiments.For example, in detail in the claims, embodiment claimed it is one of any
Can in any combination mode come using.
Various component embodiments of the invention can be implemented in hardware, or to run on one or more processors
Software module realize, or be implemented in a combination thereof.It will be understood by those of skill in the art that can be used in practice
Microprocessor or digital signal processor (DSP) realize some or all of some modules according to an embodiment of the present invention
Function.The present invention is also implemented as some or all program of device (examples for executing method as described herein
Such as, computer program and computer program product).It is such to realize that program of the invention can store in computer-readable medium
On, or may be in the form of one or more signals.Such signal can be downloaded from an internet website to obtain, or
Person is provided on the carrier signal, or is provided in any other form.
It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and ability
Field technique personnel can be designed alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol between parentheses should not be configured to limitations on claims.Word "comprising" does not exclude the presence of not
Element or step listed in the claims.Word "a" or "an" located in front of the element does not exclude the presence of multiple such
Element.The present invention can be by means of including the hardware of several different elements and being come by means of properly programmed computer real
It is existing.In the unit claims listing several devices, several in these devices can be through the same hardware branch
To embody.The use of word first, second, and third does not indicate any sequence.These words can be explained and be run after fame
Claim.
The above description is merely a specific embodiment or to the explanation of specific embodiment, protection of the invention
Range is not limited thereto, and anyone skilled in the art in the technical scope disclosed by the present invention, can be easily
Expect change or replacement, should be covered by the protection scope of the present invention.Protection scope of the present invention should be with claim
Subject to protection scope.
Claims (20)
1. a kind of method for detecting human face, which is characterized in that the method for detecting human face includes：
Face location detection is carried out using image of the multichannel neural network to acquisition and face character detects；And
Export the image with face location testing result and face character testing result.
2. method for detecting human face according to claim 1, which is characterized in that the multichannel neural network includes common sparing
And packet partial, the common sparing is for pre-processing the image of the acquisition, first in the packet partial point
For detecting face location, remaining grouping in the packet partial is respectively used to detect various face characters group.
3. method for detecting human face according to claim 2, which is characterized in that the common sparing includes by the grouping portion
The convolutional layer of each group shared divided, the packet partial includes the volume respectively used by each grouping of the packet partial
Lamination.
4. method for detecting human face according to claim 1, which is characterized in that the training of the multichannel neural network is to be based on
Multi-model distillating method.
5. method for detecting human face according to claim 1, which is characterized in that the setting of the loss of the multichannel neural network
It is the weighted average of loss and the loss of face character detection based on face location detection.
6. method for detecting human face according to claim 1, which is characterized in that the multichannel neural network may be programmed at the scene
It is realized in gate array.
7. method for detecting human face according to claim 1, which is characterized in that the face character detection includes in following
At least one of：Human face posture detection and facial image fuzzy detection.
8. method for detecting human face according to claim 1, which is characterized in that the image of the acquisition is continuous video
Frame, and the output includes：A frame or multiple image in the video frame with best detection result are exported to be used for people
Face identification.
9. a kind of human face detection device, which is characterized in that the human face detection device includes：
Detection module, for carrying out face location detection and face character inspection using image of the multichannel neural network to acquisition
It surveys；And
Output module, for exporting the image with face location testing result and face character testing result.
10. human face detection device according to claim 9, which is characterized in that the multichannel neural network includes common portion
Divide and packet partial, the common sparing are used to pre-process the image of the acquisition, first in the packet partial
For detecting face location, remaining grouping in the packet partial is respectively used to detect various face characters for grouping.
11. human face detection device according to claim 10, which is characterized in that the common sparing includes by the grouping
The convolutional layer of partial each group shared, the packet partial include respectively being used by each grouping of the packet partial
Convolutional layer.
12. human face detection device according to claim 9, which is characterized in that the training of the multichannel neural network is base
In multi-model distillating method.
13. human face detection device according to claim 9, which is characterized in that the loss of the multichannel neural network is set
Setting is the weighted average lost with the loss of face character detection based on face location detection.
14. human face detection device according to claim 9, which is characterized in that the multichannel neural network can be compiled at the scene
It is realized in journey gate array.
15. human face detection device according to claim 9, which is characterized in that the face character detection includes in following
At least one of：Human face posture detection and facial image fuzzy detection.
16. human face detection device according to claim 9, which is characterized in that the image of the acquisition is continuous video
Frame, and the output module is further used for exporting a frame or multiple image in the video frame with best detection result
To be used for recognition of face.
17. a kind of face detection system, which is characterized in that the face detection system includes storage device and processor, described
The computer program run by the processor is stored on storage device, the computer program is run by the processor
Method for detecting human face of the Shi Zhihang as described in any one of claim 1-8.
18. a kind of storage medium, which is characterized in that be stored with computer program, the computer program on the storage medium
The method for detecting human face as described in any one of claim 1-8 is executed at runtime.
19. a kind of capture machine, which is characterized in that the capture machine includes any in image collecting device and claim 9-16
Human face detection device described in.
20. a kind of capture machine, which is characterized in that the capture machine includes field programmable gate array, the field programmable gate
The method for detecting human face as described in any one of claim 1-8 is realized on array.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710915035.1A CN108875489A (en) | 2017-09-30 | 2017-09-30 | Method for detecting human face, device, system, storage medium and capture machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710915035.1A CN108875489A (en) | 2017-09-30 | 2017-09-30 | Method for detecting human face, device, system, storage medium and capture machine |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108875489A true CN108875489A (en) | 2018-11-23 |
Family
ID=64325858
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710915035.1A Pending CN108875489A (en) | 2017-09-30 | 2017-09-30 | Method for detecting human face, device, system, storage medium and capture machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108875489A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109462724A (en) * | 2018-12-24 | 2019-03-12 | 天津天地伟业机器人技术有限公司 | A kind of high-speed ball-forming machine for face snap |
| CN109584358A (en) * | 2018-11-28 | 2019-04-05 | 深圳市商汤科技有限公司 | A kind of three-dimensional facial reconstruction method and device, equipment and storage medium |
| CN111339818A (en) * | 2019-12-18 | 2020-06-26 | 中国人民解放军第四军医大学 | Face multi-attribute recognition system |
| CN111414812A (en) * | 2020-03-03 | 2020-07-14 | 平安科技（深圳）有限公司 | Human body attribute identification method, system, computer device and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105404877A (en) * | 2015-12-08 | 2016-03-16 | 商汤集团有限公司 | Human face attribute prediction method and apparatus based on deep study and multi-task study |
| CN106503669A (en) * | 2016-11-02 | 2017-03-15 | 重庆中科云丛科技有限公司 | A kind of based on the training of multitask deep learning network, recognition methods and system |
| US20170147905A1 (en) * | 2015-11-25 | 2017-05-25 | Baidu Usa Llc | Systems and methods for end-to-end object detection |
| CN106815566A (en) * | 2016-12-29 | 2017-06-09 | 天津中科智能识别产业技术研究院有限公司 | A kind of face retrieval method based on multitask convolutional neural networks |
-
2017
- 2017-09-30 CN CN201710915035.1A patent/CN108875489A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170147905A1 (en) * | 2015-11-25 | 2017-05-25 | Baidu Usa Llc | Systems and methods for end-to-end object detection |
| CN105404877A (en) * | 2015-12-08 | 2016-03-16 | 商汤集团有限公司 | Human face attribute prediction method and apparatus based on deep study and multi-task study |
| CN106503669A (en) * | 2016-11-02 | 2017-03-15 | 重庆中科云丛科技有限公司 | A kind of based on the training of multitask deep learning network, recognition methods and system |
| CN106815566A (en) * | 2016-12-29 | 2017-06-09 | 天津中科智能识别产业技术研究院有限公司 | A kind of face retrieval method based on multitask convolutional neural networks |
Non-Patent Citations (5)
| Title |
|---|
| GEOFFREY H.等: "Distilling the Knowledge in a Neural Network", 《ARXIV》 * |
| GERALD COOK: "《移动机器人导航、控制与遥感》", 31 October 2015, 北京：国防工业出版社 * |
| RAJEEV R.等: "An All-In-One Convolutional Neural Network for Face Analysis", 《2017 12TH IEEE INTERNATIONAL CONFERENCE ON AUTOMATIC FACE & GESTURE RECOGNITION (FG 2017)》 * |
| 严严 等: "基于深度学习的人脸分析研究进展", 《基于深度学习的人脸分析研究进展》 * |
| 刘成才: "《五邑大学优秀毕业论文集（2007届本科）》", 30 April 2008, 广州：华南理工大学出版社 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109584358A (en) * | 2018-11-28 | 2019-04-05 | 深圳市商汤科技有限公司 | A kind of three-dimensional facial reconstruction method and device, equipment and storage medium |
| CN109462724A (en) * | 2018-12-24 | 2019-03-12 | 天津天地伟业机器人技术有限公司 | A kind of high-speed ball-forming machine for face snap |
| CN111339818A (en) * | 2019-12-18 | 2020-06-26 | 中国人民解放军第四军医大学 | Face multi-attribute recognition system |
| CN111339818B (en) * | 2019-12-18 | 2023-06-06 | 中国人民解放军第四军医大学 | Face multi-attribute recognition system |
| CN111414812A (en) * | 2020-03-03 | 2020-07-14 | 平安科技（深圳）有限公司 | Human body attribute identification method, system, computer device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108197532B (en) | The method, apparatus and computer installation of recognition of face | |
| CN106203305B (en) | Face living body detection method and device | |
| CN108876833A (en) | Image processing method, image processing apparatus and computer readable storage medium | |
| CN108875522A (en) | Face cluster methods, devices and systems and storage medium | |
| CN108875486A (en) | Recongnition of objects method, apparatus, system and computer-readable medium | |
| CN108875489A (en) | Method for detecting human face, device, system, storage medium and capture machine | |
| CN108875732A (en) | Model training and example dividing method, device and system and storage medium | |
| CN111738243B (en) | Method, device and equipment for selecting face image and storage medium | |
| CN110008808A (en) | Panorama dividing method, device and system and storage medium | |
| CN108875540A (en) | Image processing method, device and system and storage medium | |
| CN108875546A (en) | Face auth method, system and storage medium | |
| CN108875452A (en) | Face identification method, device, system and computer-readable medium | |
| CN108875523A (en) | Human synovial point detecting method, device, system and storage medium | |
| CN109146830A (en) | For generating the method, apparatus, system and storage medium of training data | |
| CN108229331A (en) | Face false-proof detection method and system, electronic equipment, program and medium | |
| CN108875521A (en) | Method for detecting human face, device, system and storage medium | |
| CN107590473B (en) | Human face living body detection method, medium and related device | |
| CN108875481A (en) | Method, apparatus, system and storage medium for pedestrian detection | |
| CN108875525A (en) | Behavior prediction method, apparatus, system and storage medium | |
| CN108876804A (en) | It scratches as model training and image are scratched as methods, devices and systems and storage medium | |
| CN108875487A (en) | Pedestrian is identified the training of network again and is identified again based on its pedestrian | |
| CN109522790A (en) | Human body attribute recognition approach, device, storage medium and electronic equipment | |
| CN108875516A (en) | Face identification method, device, system, storage medium and electronic equipment | |
| CN106845352A (en) | Pedestrian detection method and device | |
| CN108875517A (en) | Method for processing video frequency, device and system and storage medium |
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: 20181123 |
|
| RJ01 | Rejection of invention patent application after publication |