CN109657557A - A kind of spherical surface image method for detecting human face and system based on deep learning - Google Patents
A kind of spherical surface image method for detecting human face and system based on deep learning Download PDFInfo
- Publication number
- CN109657557A CN109657557A CN201811400420.3A CN201811400420A CN109657557A CN 109657557 A CN109657557 A CN 109657557A CN 201811400420 A CN201811400420 A CN 201811400420A CN 109657557 A CN109657557 A CN 109657557A
- Authority
- CN
- China
- Prior art keywords
- spherical surface
- surface image
- face
- image
- deep learning
- 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
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000013135 deep learning Methods 0.000 title claims abstract description 24
- 238000012549 training Methods 0.000 claims abstract description 21
- 238000001514 detection method Methods 0.000 claims abstract description 19
- 238000013136 deep learning model Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 8
- 238000007689 inspection Methods 0.000 claims 1
- 230000008901 benefit Effects 0.000 abstract description 2
- 230000008569 process Effects 0.000 description 7
- 238000004590 computer program Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000000151 deposition Methods 0.000 description 2
- 230000001815 facial effect Effects 0.000 description 2
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The present invention relates to a kind of spherical surface image method for detecting human face and system based on deep learning, for realizing: it establishes for storing spherical surface image data base;Expansion processing is carried out to the spherical surface image in database, obtains panoramic picture;Face mark is carried out to panoramic picture, using the image for carrying out face mark as the data source of training deep learning model, training managing is executed and obtains Face datection model;Multiple spherical surface images are acquired using photographic device, and successively execute the processing of above-mentioned steps to spherical surface image, further, detection is executed using face detection model, obtains human face region figure.The invention has the benefit that face upsides down the problem of can not being detected in effective solution spherical surface image;Spherical surface image is expanded into after panoramic picture and reuses the better MTCNN deep learning model of robustness, so that the Face datection of spherical surface image edge location is more accurate.
Description
Technical field
The present invention relates to a kind of spherical surface image method for detecting human face and system based on deep learning belongs to computer neck
Domain.
Background technique
It is all usually a ball from the data that camera acquires on the monocular panoramic shooting head apparatus of protection and monitor field
Face image, in order to realize the Face datection function of high performance-price ratio on devices, it will usually the spherical surface acquired using camera
Image carries out Face datection as initial data.For example, realizing the function of Face datection on 360 ° of panorama automobile data recorders of monocular
Energy.
The prior art has the disadvantage in that (1) since the face in spherical surface image may be to upside down, leads to common people
Face detection algorithm can not detect the image upsided down;(2) when facial image is located at the marginal position of spherical surface image, facial image
Distortion is had, causes traditional Face datection algorithm robustness poor;(3) due to the high resolution of spherical surface image, to face
The performance of detection algorithm has higher requirements.
Summary of the invention
The present invention provides a kind of spherical surface image method for detecting human face and system based on deep learning, it is necessary first to by spherical surface
Image is unfolded, and panoramic picture is obtained;Position distortion due to spherical surface image closer to edge is bigger, after deployment
The corresponding position of panoramic picture can have the case where certain stretcher strain, it would therefore be desirable to establish the picture number of oneself
According to library, model training is carried out using the better deep learning algorithm of robustness, obtains the Face datection model based on panoramic picture;
Then Face datection is carried out in practical applications using the face detection model.
Technical solution of the present invention includes a kind of spherical surface image method for detecting human face based on deep learning, and feature exists
In method includes the following steps: A. is established for storing spherical surface image data base;B. the spherical surface image in database is used
Algorithm is handled, and panoramic picture is obtained;C. face mark is carried out to panoramic picture, the image of face mark will be carried out as instruction
Practice the data source of deep learning model, executes training managing and obtain Face datection model;D. multiple balls are acquired using photographic device
Face image, and the processing of step A and B is successively executed to spherical surface image, further, detection is executed using face detection model, is obtained
To human face region figure.
According to the spherical surface image method for detecting human face based on deep learning, wherein the step A further include: described
Database is the database for being stored with a large amount of spherical surface images, wherein a large amount of spherical surface images are as training data source.
According to the spherical surface image method for detecting human face based on deep learning, wherein step B is specifically included: to data
Spherical surface image in library obtains corresponding panoramic picture by deployment algorithm, and the obtained face in spherical surface image is in panoramic picture
In be positive.
According to the spherical surface image method for detecting human face based on deep learning, wherein step C is specifically included: to data
Panoramic picture in library carries out face mark, to data needed for providing trained deep learning model;It is calculated using deep learning
Method is trained labeled data, and wherein deep learning algorithm is MTCNN algorithm, further obtains the panorama sketch based on MTCNN
As Face datection model.
According to the spherical surface image method for detecting human face based on deep learning, wherein step D is specifically included: using single
Mesh camera collects multiple spherical surface images;Spherical surface image is expanded into panoramic picture by deployment algorithm;Use what is trained
Panoramic picture Face datection model based on MTCNN detects current panorama image;By what is detected based on panoramic picture
Human face region is mapped to the corresponding position of spherical surface image, to obtain the region of face in spherical surface image.
Technical solution of the present invention further includes a kind of spherical surface image people based on deep learning for above-mentioned any means
Face detection system, which is characterized in that the system includes: image storage module, for establishing for storing spherical surface image data base;
Image spread module obtains panoramic picture for being handled using algorithm the spherical surface image in database;Image trains mould
Block will carry out the image of face mark as the data of training deep learning model for carrying out face mark to panoramic picture
Source executes training managing and obtains Face datection model;Image detection module, for using photographic device to acquire multiple spherical diagrams
Picture, and spherical surface image is successively handled using memory module and image spread module, further, use described image training
The Face datection model that module obtains executes detection, obtains human face region figure.
The invention has the benefit that spherical surface image is expanded into panoramic picture by panoramic expansion algorithm, then pass through
The panoramic picture carries out Face datection, and face upsides down the problem of can not being detected in effective solution spherical surface image;Spherical surface
Position distortion of the image closer to edge is bigger, for the face in spherical surface image edge location, traditional method for detecting human face
Detection accuracy is low, expands into after panoramic picture spherical surface image reuse the better MTCNN deep learning mould of robustness herein
Type, so that the Face datection of spherical surface image edge location is more accurate.
Detailed description of the invention
Fig. 1 show overview flow chart according to the method for the present invention.
Fig. 2 show overall system block diagram according to the present invention;
The Face datection model that Fig. 3 show embodiment according to the present invention obtains flow chart;
Fig. 4 show the Face datection flow chart of embodiment according to the present invention.
Specific embodiment
It is carried out below with reference to technical effect of the embodiment and attached drawing to design of the invention, specific structure and generation clear
Chu, complete description, to be completely understood by the purpose of the present invention, scheme and effect.
It should be noted that unless otherwise specified, when a certain feature referred to as " fixation ", " connection " are in another feature,
It can directly fix, be connected to another feature, and can also fix, be connected to another feature indirectly.In addition, this
The descriptions such as the upper and lower, left and right used in open are only the mutual alignment pass relative to each component part of the disclosure in attached drawing
For system.The "an" of used singular, " described " and "the" are also intended to including most forms in the disclosure, are removed
Non- context clearly expresses other meaning.In addition, unless otherwise defined, all technical and scientific terms used herein
It is identical as the normally understood meaning of those skilled in the art.Term used in the description is intended merely to describe herein
Specific embodiment is not intended to be limiting of the invention.Term as used herein "and/or" includes one or more relevant
The arbitrary combination of listed item.
It will be appreciated that though various elements, but this may be described using term first, second, third, etc. in the disclosure
A little elements should not necessarily be limited by these terms.These terms are only used to for same type of element being distinguished from each other out.For example, not departing from
In the case where disclosure range, first element can also be referred to as second element, and similarly, second element can also be referred to as
One element.The use of provided in this article any and all example or exemplary language (" such as ", " such as ") is intended merely to more
Illustrate the embodiment of the present invention well, and unless the context requires otherwise, otherwise the scope of the present invention will not be applied and be limited.
Fig. 1 show overview flow chart according to the method for the present invention.The process includes the following steps: that A. is established for depositing
Store up spherical surface image data base;B. the spherical surface image in database is handled using algorithm, obtains panoramic picture;C. to panorama
Image carries out face mark, using the image for carrying out face mark as the data source of training deep learning model, executes at training
Reason obtains Face datection model;D. multiple spherical surface images are acquired using photographic device, and to spherical surface image successively execute step A and
The processing of B further executes detection using face detection model, obtains human face region figure.
Fig. 2 show overall system block diagram according to the present invention.The system includes: image storage module, is used for establishing
In storage spherical surface image data base;Image spread module obtains complete for carrying out expansion processing to the spherical surface image in database
Scape image;Image training module, it is for carrying out face mark to panoramic picture, the image for carrying out face mark is deep as training
The data source of learning model is spent, training managing is executed and obtains Face datection model;Image detection module, for using photographic device
Multiple spherical surface images are acquired, and spherical surface image is successively handled using memory module and image spread module, further, are made
The Face datection model obtained with described image training module executes detection, obtains human face region figure.
The Face datection model that Fig. 3 show embodiment according to the present invention obtains flow chart.Its process such as S31-S35 institute
Show, comprising:
S31 establishes the spherical surface image data base an of big data quantity;
S32 obtains corresponding panoramic picture by deployment algorithm to the spherical surface image in database, in this way in spherical surface image
The face upsided down be in panoramic picture just;
S33 carries out face mark to the panoramic picture in database, to number needed for providing trained deep learning model
According to;
S34 is trained labeled data using deep learning algorithm, calculates herein using the better MTCNN of efficiency
Method;
S35 obtains the panoramic picture Face datection model based on MTCNN.
Fig. 4 show the Face datection flow chart of embodiment according to the present invention.Its process is as shown in S41-S45, comprising:
S41 collects spherical surface image by monocular cam first;
Spherical surface image is expanded into panoramic picture by deployment algorithm by S42;
S43 carries out current panorama image using the above-mentioned trained panoramic picture Face datection model based on MTCNN
Detection;
S44 will be mapped to the corresponding position of spherical surface image based on the human face region that panoramic picture detects;
S45 obtains the region of face in spherical surface image.
It should be appreciated that the embodiment of the present invention can be by computer hardware, the combination of hardware and software or by depositing
The computer instruction in non-transitory computer-readable memory is stored up to be effected or carried out.Standard volume can be used in the method
Journey technology-includes that the non-transitory computer-readable storage media configured with computer program is realized in computer program,
In configured in this way storage medium computer is operated in a manner of specific and is predefined --- according in a particular embodiment
The method and attached drawing of description.Each program can with the programming language of level process or object-oriented come realize with department of computer science
System communication.However, if desired, the program can be realized with compilation or machine language.Under any circumstance, which can be volume
The language translated or explained.In addition, the program can be run on the specific integrated circuit of programming for this purpose.
In addition, the operation of process described herein can be performed in any suitable order, unless herein in addition instruction or
Otherwise significantly with contradicted by context.Process described herein (or modification and/or combination thereof) can be held being configured with
It executes, and is can be used as jointly on the one or more processors under the control of one or more computer systems of row instruction
The code (for example, executable instruction, one or more computer program or one or more application) of execution, by hardware or its group
It closes to realize.The computer program includes the multiple instruction that can be performed by one or more processors.
Further, the method can be realized in being operably coupled to suitable any kind of computing platform, wrap
Include but be not limited to PC, mini-computer, main frame, work station, network or distributed computing environment, individual or integrated
Computer platform or communicated with charged particle tool or other imaging devices etc..Each aspect of the present invention can be to deposit
The machine readable code on non-transitory storage medium or equipment is stored up to realize no matter be moveable or be integrated to calculating
Platform, such as hard disk, optical reading and/or write-in storage medium, RAM, ROM, so that it can be read by programmable calculator, when
Storage medium or equipment can be used for configuration and operation computer to execute process described herein when being read by computer.This
Outside, machine readable code, or part thereof can be transmitted by wired or wireless network.When such media include combining microprocessor
Or other data processors realize steps described above instruction or program when, invention as described herein including these and other not
The non-transitory computer-readable storage media of same type.When methods and techniques according to the present invention programming, the present invention
It further include computer itself.
Computer program can be applied to input data to execute function as described herein, to convert input data with life
At storing to the output data of nonvolatile memory.Output information can also be applied to one or more output equipments as shown
Device.In the preferred embodiment of the invention, the data of conversion indicate physics and tangible object, including the object generated on display
Reason and the particular visual of physical objects are described.
The above, only presently preferred embodiments of the present invention, the invention is not limited to above embodiment, as long as
It reaches technical effect of the invention with identical means, all within the spirits and principles of the present invention, any modification for being made,
Equivalent replacement, improvement etc., should be included within the scope of the present invention.Its technical solution within the scope of the present invention
And/or embodiment can have a variety of different modifications and variations.
Claims (6)
1. a kind of spherical surface image method for detecting human face based on deep learning, which is characterized in that method includes the following steps:
A. it establishes for storing spherical surface image data base;
B. expansion processing is carried out to the spherical surface image in database, obtains panoramic picture;
C. face mark is carried out to panoramic picture, the image of face mark will be carried out as the data of training deep learning model
Source executes training managing and obtains Face datection model;
D. multiple spherical surface images are acquired using photographic device, and successively executes the processing of step A and B to spherical surface image, further,
Detection is executed using face detection model, obtains human face region figure.
2. the spherical surface image method for detecting human face according to claim 1 based on deep learning, which is characterized in that the step
Rapid A further include: the database is the database for being stored with a large amount of spherical surface images, wherein a large amount of spherical surface images are as training data
Source.
3. the spherical surface image method for detecting human face according to claim 1 based on deep learning, which is characterized in that the step
Rapid B is specifically included:
Corresponding panoramic picture is obtained by deployment algorithm to the spherical surface image in database, the obtained face in spherical surface image
It is positive in panoramic picture.
4. the spherical surface image method for detecting human face according to claim 1 based on deep learning, which is characterized in that the step
Rapid C is specifically included:
Face mark is carried out to the panoramic picture in database, to data needed for providing trained deep learning model;
Labeled data is trained using deep learning algorithm, wherein deep learning algorithm is MTCNN algorithm, is further obtained
Panoramic picture Face datection model based on MTCNN.
5. the spherical surface image method for detecting human face according to claim 1 or 4 based on deep learning, which is characterized in that institute
Step D is stated to specifically include:
Multiple spherical surface images are collected using monocular cam;
Spherical surface image is expanded into panoramic picture by deployment algorithm;
Current panorama image is detected using the panoramic picture Face datection model based on MTCNN trained;
It will be mapped to the corresponding position of spherical surface image based on the human face region that panoramic picture detects, to obtain in spherical surface image
The region of face.
6. a kind of for executing the spherical surface image Face datection system based on deep learning of the claim 1-5 any means
System, which is characterized in that the system includes:
Image storage module, for establishing for storing spherical surface image data base;
Image spread module obtains panoramic picture for carrying out expansion processing to the spherical surface image in database;
Image training module will carry out the image of face mark as training depth for carrying out face mark to panoramic picture
The data source of learning model executes training managing and obtains Face datection model;
Image detection module, for using photographic device to acquire multiple spherical surface images, and to spherical surface image successively using storage mould
Block and image spread module are handled, and further, the Face datection model obtained using described image training module executes inspection
It surveys, obtains human face region figure.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811400420.3A CN109657557A (en) | 2018-11-22 | 2018-11-22 | A kind of spherical surface image method for detecting human face and system based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811400420.3A CN109657557A (en) | 2018-11-22 | 2018-11-22 | A kind of spherical surface image method for detecting human face and system based on deep learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109657557A true CN109657557A (en) | 2019-04-19 |
Family
ID=66110998
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811400420.3A Pending CN109657557A (en) | 2018-11-22 | 2018-11-22 | A kind of spherical surface image method for detecting human face and system based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109657557A (en) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107358157A (en) * | 2017-06-07 | 2017-11-17 | 阿里巴巴集团控股有限公司 | A kind of human face in-vivo detection method, device and electronic equipment |
CN107992844A (en) * | 2017-12-14 | 2018-05-04 | 合肥寰景信息技术有限公司 | Face identification system and method based on deep learning |
-
2018
- 2018-11-22 CN CN201811400420.3A patent/CN109657557A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107358157A (en) * | 2017-06-07 | 2017-11-17 | 阿里巴巴集团控股有限公司 | A kind of human face in-vivo detection method, device and electronic equipment |
CN107992844A (en) * | 2017-12-14 | 2018-05-04 | 合肥寰景信息技术有限公司 | Face identification system and method based on deep learning |
Non-Patent Citations (1)
Title |
---|
徐佳等: "基于鱼眼摄像头的一种人脸识别技术", 《信息通信》 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110866953B (en) | Map construction method and device, and positioning method and device | |
US9898647B2 (en) | Systems and methods for detecting, identifying and tracking objects and events over time | |
US10334168B2 (en) | Threshold determination in a RANSAC algorithm | |
CN108875524B (en) | Sight estimation method, device, system and storage medium | |
CN105229703B (en) | System and method for generating threedimensional model using the position data of sensing | |
WO2018228218A1 (en) | Identification method, computing device, and storage medium | |
JP5899472B2 (en) | Person attribute estimation system and learning data generation apparatus | |
CN109416744A (en) | Improved camera calibration system, target and process | |
CN109815843A (en) | Object detection method and Related product | |
CN109711472B (en) | Training data generation method and device | |
CN109873979A (en) | Camera-based static image difference comparison method and device | |
JP2013105258A (en) | Information presentation apparatus, method therefor, and program therefor | |
CN105765628A (en) | Depth map generation | |
CN110969045B (en) | Behavior detection method and device, electronic equipment and storage medium | |
CN114241338A (en) | Building measuring method, device, equipment and storage medium based on image recognition | |
CN115298693A (en) | Data generation method, learning method, and estimation method | |
KR20200078258A (en) | Apparatus for fitting of articulated model and method for the same | |
JP6875646B2 (en) | Image processing device and image processing program | |
CN109657557A (en) | A kind of spherical surface image method for detecting human face and system based on deep learning | |
CN115880448A (en) | Three-dimensional measurement method, device and equipment based on binocular imaging and storage medium | |
CN113190013A (en) | Method and device for controlling terminal movement | |
US9842402B1 (en) | Detecting foreground regions in panoramic video frames | |
US9824455B1 (en) | Detecting foreground regions in video frames | |
JP2019113959A (en) | Muzzle pattern collation system, muzzle pattern collation method and muzzle pattern collation program | |
CN109389645A (en) | Camera method for self-calibrating, system, camera, robot and cloud server |
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: 20190419 |
|
RJ01 | Rejection of invention patent application after publication |