CN106991405A - A kind of dynamic portrait storehouse method for building up - Google Patents
A kind of dynamic portrait storehouse method for building up Download PDFInfo
- Publication number
- CN106991405A CN106991405A CN201710227880.XA CN201710227880A CN106991405A CN 106991405 A CN106991405 A CN 106991405A CN 201710227880 A CN201710227880 A CN 201710227880A CN 106991405 A CN106991405 A CN 106991405A
- Authority
- CN
- China
- Prior art keywords
- portrait
- picture
- dynamic
- effective
- building
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
- G06V40/165—Detection; Localisation; Normalisation using facial parts and geometric relationships
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
- G06V40/171—Local features and components; Facial parts ; Occluding parts, e.g. glasses; Geometrical relationships
Abstract
The invention discloses a kind of dynamic portrait storehouse method for building up, it is characterised in that this method includes following process step：Step 1：Portrait picture collection is carried out in real time and portrait interception is carried out to the picture of collection according to effective portrait standard of setting, captures out effective portrait picture；Step 2：Effective portrait picture is obtained into portrait picture feature point by dynamic and intelligent portrait algorithm process；Step 3：The portrait picture feature point obtained is preserved in real time forms dynamic portrait storehouse.The present invention is judged effective portrait by capturing effective portrait in collection picture, so as to set up transient target portrait storehouse, can effectively improve discrimination and authentication efficiency.
Description
Technical field
The present invention relates to dynamic portrait storehouse method for building up, more particularly to a kind of dynamic portrait storehouse for effectively improving discrimination
Method for building up.
Background technology
In recent years, security protection industry has started the upsurge of a ripple Identification of Images, and numerous manufacturers are proposed Related product, made one after another
Obtaining Identification of Images becomes the hot spot technology direction in industry.
Although existing portrait recognition technology function is huge, in actual applications, we, which remain unchanged, can run into many problems.Its
In, influence of the portrait image quality to discrimination is higher, and poor image quality, identification is low, and validity feature seldom, is used sometimes
Naked eyes also are difficult to confirm identity.Picture quality is again affected by many factors, such as illumination, posture, expression, portrait size, definition
Deng.Contrasted by picture of the same person under different illumination, it is possible to find out that picture of the same person under different illumination is used
Naked eyes are difficult to distinguish.Thus, current face recognition system can only be carried out in the environment of some are compared with specification, and such as light is equal
Even, portrait needs to face video camera, and guarantor in picture as there is certain pixel wide.But in actual security protection
Monitor in occasion, these restrictive conditions are difficult to meet one by one.At present, it is usually to enter to being certified people during human image collecting
Row static state human image collecting, such processing mode processing time is longer and progress of can only be single is handled, it is impossible to same to many people
Shi Jinhang human image collectings, so that the collection needs in practical application can not be met, influence operating efficiency.
The content of the invention
It is an object of the present invention to provide a kind of dynamic portrait storehouse method for building up, portrait storehouse quickly can be effectively set up, ratio is improved
To efficiency.
Realizing the purpose for this, there is provided a kind of dynamic portrait storehouse method for building up, it is characterised in that this method includes following
Process step：
Step 1：Portrait picture collection is carried out in real time and the picture of collection is carried out according to effective portrait standard of setting
Portrait is intercepted, and captures out effective portrait picture；
Step 2：Effective portrait picture is obtained into portrait picture feature point by dynamic and intelligent portrait algorithm process；
Step 3：The portrait picture feature point obtained is preserved in real time forms dynamic portrait storehouse.
Preferably, in step 2, dynamic and intelligent portrait algorithm includes some sub-steps, and each sub-step is simultaneously to receiving
Effective portrait picture handled and each sub-step while handle different effective portrait pictures respectively.
Preferably, in step 2, while carrying out effective portrait picture feature point extraction, according to effective portrait of setting
Picture is complete degree fraction carries out effective portrait test and judge, if effectively portrait then carries out step 3, if not effectively portrait is then returned
Return step 1.
Preferably, in the dynamic and intelligent portrait algorithm, each sub-step includes input picture step, portrait successively
Detecting step, characteristic extraction step and output result step.
Preferably, the input picture step is to be used as the incoming dynamic and intelligent portrait of parameter to calculate in picture in binary form
In method.
Preferably, the portrait detecting step is the position that portrait is found in the picture after input picture step process
Coordinate is put and then set up, portrait is then positioned by portrait position's coordinate.
Preferably, the characteristic extraction step is, after the processing of portrait detecting step, to pass through portrait picture and portrait
Position coordinates obtains the feature in the middle of the eyes and mouth and both of the portrait.
Preferably, in step 1, the effective portrait standard set as set the definition of portrait picture, portrait
The setting range of the angle of candid photograph and eyes can be seen.
Preferably, the characteristic point of effective portrait picture and effective portrait picture is locally preserved.
Compared with prior art, its advantage is the present invention：
The present invention gathers effective portrait in picture by crawl, and effective portrait is judged, can be quickly effective
Portrait storehouse is set up, comparison efficiency is improved.It can be changed 1 in traditional images checking by the present invention：1 compares, and enters when realizing checking
Row 1：N is compared, so as to reach the function of comparison and returning result information at high speed.
Brief description of the drawings
Fig. 1 is flow chart of the invention；
Fig. 2 is the block diagram of dynamic and intelligent portrait algorithm in the present invention.
Embodiment
With reference to embodiment, the invention will be further described, but does not constitute any limitation of the invention, any
The modification of the limited number of time made in scope of the invention as claimed, still in scope of the presently claimed invention.
As shown in figure 1, the invention provides a kind of dynamic portrait storehouse method for building up, this method includes following process step：
Step 1：Portrait picture collection is carried out in real time and the picture of collection is carried out according to effective portrait standard of setting
Portrait is intercepted, and captures out effective portrait picture；
Step 2：Effective portrait picture is obtained into portrait picture feature point by dynamic and intelligent portrait algorithm process；
Step 3：The portrait picture feature point obtained is preserved in real time forms dynamic portrait storehouse.
In the present embodiment, dynamic portrait storehouse includes the portrait picture feature point of collected different people, and each individual
The portrait picture feature point of different angles.
In step 2, dynamic and intelligent portrait algorithm includes some sub-steps, the effective people of each sub-step simultaneously to reception
As picture is handled and each sub-step while handle different effective portrait pictures respectively.
In step 2, while carrying out effective portrait picture feature point extraction, according to effective portrait picture is complete of setting
Spend fraction and carry out effective portrait test and judge, if effectively portrait then carries out step 3, if not effectively portrait then return to step 1.
In the present embodiment, the effective portrait picture is complete degree fraction set as facial integrity degree reach 70% and more than.
In addition, effective portrait picture is complete degree fraction of setting also for facial integrity degree reach 50% and above or 80% and more than.
As shown in Fig. 2 in dynamic and intelligent portrait algorithm, each sub-step includes input picture step, portrait detection successively
Step, characteristic extraction step and output result step.It is incoming using picture as parameter in binary form to input picture step
In dynamic and intelligent portrait algorithm.Portrait detecting step is the position that portrait is found in the picture after input picture step process
Coordinate is put and then set up, portrait is then positioned by portrait position's coordinate.Characteristic extraction step is to detect step by portrait
After rapid processing, the feature in the middle of the eyes and mouth and both of the portrait is obtained by portrait picture and portrait position's coordinate.
In step 1, the angle that the effective portrait standard set is captured as the definition of portrait picture that sets, portrait
Setting range and eyes can be seen.
The characteristic point of effective portrait picture and effective portrait picture is locally preserved.
In the present embodiment, portrait position's coordinate is to be built according to the position of portrait eyes in the picture in portrait detecting step
Vertical coordinate.In characteristic extraction step, the feature extracted includes nose for the feature in the middle of the eyes and mouth and both of portrait
With the feature of face, and some characteristic points are obtained, then set up the coordinate of each characteristic point, e.g., nose or a mole on the face,
Then using the mole is as a characteristic point and sets up coordinate.Effective portrait is carried out according to effective portrait standard to be judged as to see double
On the premise of eyes, the angle that definition size and portrait are captured is obtained according to picture pixels, such as the angle of eyes can be being seen
In the range of, effective portrait judgement is carried out to picture.
In the present embodiment, processing equipment is computer, and ground memory cell is the hard disk of computer.
In the present embodiment, dynamic and intelligent portrait algorithm is different from existing static person as algorithm, and static person will as algorithm
Ask picture limpider, light can not be excessive than more uniform, angle.The dynamic and intelligent portrait algorithm of the present invention, which has, focuses on big rule
Mould dynamic Identification of Images, the analysis of large-scale crowd behavioural characteristic and early warning, vehicle attribute analysis are retrieved, and the core such as image retrieval is closed
Key technology.The present invention proposes double-deck isomery deep neural network theoretical frame, and under this framework, research nonlinear images are empty
Between mapping method, by the adaptive fusion of many deep neural networks, same figure will be mapped to across scene, non-homogeneous portrait image
Image space；Research level feature learning method, to extract many portrait areas, the stratification feature with identification；Research is non-
Linear character space mapping method, by the smaller space of difference in Feature Mapping to class inherited bigger, class.In real scene prison
On control video test prove the algorithm of the present invention to changeable illumination, multi-angle, block, obscure, the complex situations such as age range
With good robustness.So that the dynamic and intelligent portrait algorithm of the present invention can greatly reduce to environmental requirement, simultaneously
Can individual machine to multiple portrait algorithm services carry out computing, greatly improve verification efficiency.
The course of work of the present invention：Effective portrait crawl is carried out to being certified the image that people gathers by human image collecting rifle,
And effective portrait picture is preserved to the hard disk of computer；The effective portrait picture preserved on hard disk is by being sent to dynamic
Handled in intelligent portrait algorithm and by some sub-steps, the effective portrait picture of each sub-steps simultaneously to reception
Information processing is carried out, high speed processing is reached, the characteristic point of effective portrait is extracted；Carrying out what effective portrait picture feature point was extracted
Meanwhile, effective portrait judgement is carried out according to the integrity degree fractional value of setting, if integrity degree fraction reaches the 70% of setting, to have
The characteristic point of effective portrait picture is registered and is saved on internal memory by effect portrait, if integrity degree fraction does not reach setting value, no
IMAQ then is carried out to being certified people again for effective portrait, whole processing procedure is repeated, until setting value must be reached
Effective portrait picture, and the characteristic point of effective portrait picture is registered and is saved on internal memory, form dynamic portrait storehouse.
The present invention quickly can effectively set up portrait storehouse, and acquisition rate is high, comparison efficiency is improved, to the anti-interference strong ability of light
By force；And it can be changed 1 in traditional images checking by the present invention：1 is compared, and 1 is carried out when realizing checking：N is compared, so as to reach
The function of high speed returning result information.
It the above is only the preferred embodiment of the present invention, it should be pointed out that for those skilled in the art, do not taking off
On the premise of from structure of the present invention, several modifications and improvements can also be made, these effects implemented all without the influence present invention
And practical applicability.
Claims (9)
1. a kind of dynamic portrait storehouse method for building up, it is characterised in that this method includes following process step：
Step 1：Portrait picture collection is carried out in real time and portrait is carried out to the picture of collection according to effective portrait standard of setting
Interception, captures out effective portrait picture；
Step 2：Effective portrait picture is obtained into portrait picture feature point by dynamic and intelligent portrait algorithm process；
Step 3：The portrait picture feature point obtained is preserved in real time and forms dynamic portrait storehouse.
2. a kind of multi-functional dynamic and intelligent identity card identification verification method according to claim 1, it is characterised in that：
In step 2, dynamic and intelligent portrait algorithm includes some sub-steps, the effective portrait figure of each sub-step simultaneously to reception
Piece is handled and each sub-step while handle different effective portrait pictures respectively.
3. a kind of multi-functional dynamic and intelligent identity card identification verification method according to claim 2, it is characterised in that：
In step 2, while carrying out effective portrait picture feature point extraction, according to effective portrait picture is complete degree of setting point
Number carries out effective portrait test and judges, if effectively portrait then carries out step 3, if not effectively portrait then return to step 1.
4. a kind of dynamic portrait storehouse method for building up according to claim 3, it is characterised in that：In the dynamic and intelligent portrait
In algorithm, each sub-step includes input picture step, portrait detecting step, characteristic extraction step and output result step successively
Suddenly.
5. a kind of dynamic portrait storehouse method for building up according to claim 4, it is characterised in that：It is described input picture step be
In binary form using picture as in the incoming dynamic and intelligent portrait algorithm of parameter.
6. a kind of dynamic portrait storehouse method for building up according to claim 3, it is characterised in that：The portrait detecting step is
The position of portrait is found in the picture after input picture step process and then coordinate is set up, is then sat by portrait position
Mark to position portrait.
7. a kind of dynamic portrait storehouse method for building up according to claim 3, it is characterised in that：The characteristic extraction step is
After the processing of portrait detecting step, the eyes and mouth and its two of the portrait are obtained by portrait picture and portrait position's coordinate
Feature in the middle of person.
8. a kind of dynamic portrait storehouse method for building up according to claim 1, it is characterised in that：
In step 1, the angle that the effective portrait standard set is captured as the definition of portrait picture that sets, portrait
Setting range and eyes can be seen.
9. a kind of dynamic portrait storehouse method for building up according to claim 1, it is characterised in that：Effective portrait picture and
The characteristic point of effective portrait picture is locally preserved.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710227880.XA CN106991405A (en) | 2017-04-10 | 2017-04-10 | A kind of dynamic portrait storehouse method for building up |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710227880.XA CN106991405A (en) | 2017-04-10 | 2017-04-10 | A kind of dynamic portrait storehouse method for building up |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106991405A true CN106991405A (en) | 2017-07-28 |
Family
ID=59415979
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710227880.XA Pending CN106991405A (en) | 2017-04-10 | 2017-04-10 | A kind of dynamic portrait storehouse method for building up |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106991405A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107944424A (en) * | 2017-12-08 | 2018-04-20 | 广东金杭科技有限公司 | Front end human image collecting and Multi-angle human are distributed as comparison method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101347332A (en) * | 2008-08-22 | 2009-01-21 | 深圳先进技术研究院 | Measurement method and equipment of digitized measurement system of human face three-dimensional surface shape |
| US20150154470A1 (en) * | 2013-11-29 | 2015-06-04 | Samsung Techwin Co., Ltd. | Image matching method using feature point matching |
| CN106169067A (en) * | 2016-07-01 | 2016-11-30 | 恒东信息科技无锡有限公司 | A kind of police dynamic human face of high flux gathers comparison method and system |
-
2017
- 2017-04-10 CN CN201710227880.XA patent/CN106991405A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101347332A (en) * | 2008-08-22 | 2009-01-21 | 深圳先进技术研究院 | Measurement method and equipment of digitized measurement system of human face three-dimensional surface shape |
| US20150154470A1 (en) * | 2013-11-29 | 2015-06-04 | Samsung Techwin Co., Ltd. | Image matching method using feature point matching |
| CN106169067A (en) * | 2016-07-01 | 2016-11-30 | 恒东信息科技无锡有限公司 | A kind of police dynamic human face of high flux gathers comparison method and system |
Non-Patent Citations (1)
| Title |
|---|
| 傅泽华: ""基于RGB-D数据的三维人脸建模及标准化"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107944424A (en) * | 2017-12-08 | 2018-04-20 | 广东金杭科技有限公司 | Front end human image collecting and Multi-angle human are distributed as comparison method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104933414B (en) | A kind of living body faces detection method based on WLD-TOP | |
| CN103530599B (en) | The detection method and system of a kind of real human face and picture face | |
| CN101271514B (en) | Image detection method and device for fast object detection and objective output | |
| TW201627917A (en) | Method and device for face in-vivo detection | |
| KR101781358B1 (en) | Personal Identification System And Method By Face Recognition In Digital Image | |
| CN107330390B (en) | People counting method based on image analysis and deep learning | |
| CN105518709A (en) | Method, system and computer program product for identifying human face | |
| CN105989608B (en) | A kind of vision capture method and device towards intelligent robot | |
| CN111241975B (en) | Face recognition detection method and system based on mobile terminal edge calculation | |
| CN107944416A (en) | A kind of method that true man's verification is carried out by video | |
| CN109635686A (en) | In conjunction with two stages pedestrian's searching method of face and appearance | |
| CN107169458A (en) | Data processing method, device and storage medium | |
| CN111091098A (en) | Training method and detection method of detection model and related device | |
| CN106991448A (en) | A kind of portrait compares processing method | |
| CN107103299B (en) | People counting method in monitoring video | |
| WO2020249054A1 (en) | Living body detection method and system for human face by using two long-baseline cameras | |
| CN104537381B (en) | A kind of fuzzy image recognition method based on fuzzy invariant features | |
| CN106791706A (en) | Object lock method and system | |
| CN111582118A (en) | Face recognition method and device | |
| CN103020655A (en) | Remote identity authentication method based on single training sample face recognition | |
| CN112668557A (en) | Method for defending image noise attack in pedestrian re-identification system | |
| CN111723656B (en) | Smog detection method and device based on YOLO v3 and self-optimization | |
| CN106991405A (en) | A kind of dynamic portrait storehouse method for building up | |
| Ma et al. | Multi-perspective dynamic features for cross-database face presentation attack detection | |
| CN111191519B (en) | Living body detection method for user access of mobile power supply device |
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: 20170728 |
|
| RJ01 | Rejection of invention patent application after publication |