CN108563675A - Electronic record automatic generation method and device based on target body characteristics - Google Patents
Electronic record automatic generation method and device based on target body characteristics Download PDFInfo
- Publication number
- CN108563675A CN108563675A CN201810166463.3A CN201810166463A CN108563675A CN 108563675 A CN108563675 A CN 108563675A CN 201810166463 A CN201810166463 A CN 201810166463A CN 108563675 A CN108563675 A CN 108563675A
- Authority
- CN
- China
- Prior art keywords
- objective body
- objective
- archive information
- archives
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Human Computer Interaction (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to technical field of data processing, specifically provide a kind of electronic record automatic generation method and device based on target body characteristics, it is intended to the technical issues of how solution quickly and accurately obtains target person archive information.For this purpose, electronic record automatic generation method in the present invention can will belong to the archive information of the objective body of same files classification, the corresponding archival region of current class objective body is stored into the file store built in advance, so that the archive information of each archival region storage is the archive information of same archives kind objective body, it realizes " one grade a kind of ", can be realized when objective body is behaved " one grade of a people ".The attribute information of the target person is searched in dynamic portrait library after determining target person without the mode searched for one by one based on the dynamic portrait library that above-mentioned electronic record automatic generation method is generated, you can obtain whole archive informations of the target person.Meanwhile the device in the present invention is able to carry out and realizes the above method.
Description
Technical field
The present invention relates to technical field of data processing, and in particular to a kind of electronic record based on target body characteristics is automatic
Generation method and device.
Background technology
Dynamic portrait library refers to the information database comprising figure information and its satellite information, wherein satellite information packet
Containing the information such as hair style and clothes color.Information data in current dynamic portrait library is mainly according to acquisition time and locality
Point classification storage, therefore the data information in dynamic portrait library can be scanned in accordance with the following steps:Basis is searched first
Rope target determines search time range and search space range, then according to the time and spatial dimension determined, to dynamic
Corresponding data information is searched for one by one in portrait library, however the data volume of these data informations is often very big, and use is this
The mode searched for one by one certainly will will increase search time, while can also reduce searching accuracy.
Invention content
In order to solve the above problem in the prior art, in order to solve how quickly and accurately to obtain target person
The technical issues of archive information, the present invention provides a kind of electronic record automatic generation methods and dress based on target body characteristics
It sets.
In a first aspect, the electronic record automatic generation method based on target body characteristics in the present invention includes:
Using across camera target tracking algorism, objective body image is obtained;
Acquired objective body image is identified, obtains target body characteristics, and according to the target body characteristics to institute
It states objective body and carries out archives classification;
The archive information that the objective body of same files classification will be belonged to stores into the file store built in advance current class
The corresponding archival region of other objective body.
Further, an optimal technical scheme provided by the invention is:
The target body characteristics include face feature;It " carries out archives to the objective body according to the target body characteristics to return
The step of class ", specifically includes:
According to the face feature, traversal is carried out to stored archive information in the file store built in advance and is searched
Rope obtains the first candidate archive information;
The described first candidate archive information is screened according to preset objective body screening conditions, obtains the second candidate
Archive information;
Obtain the face model of candidate target body corresponding to the described second candidate archive information;
Characteristic matching is carried out to the face feature and the face model, and matching result is weighted
To the first similarity;
First similarity and predetermined threshold value are compared, and determine the shelves of the objective body according to comparison result
Case classification.
Further, an optimal technical scheme provided by the invention is:
The target body characteristics further include hair style/hair decorations feature, garment ornament and/or vehicles feature;
" first similarity and predetermined threshold value are compared, and determine the shelves of the objective body according to comparison result
The step of case classification ", specifically includes:
If fSimi≤Thr_low, then archives kind of the new archives kind as the objective body is built, wherein institute
State fSimiFor the first similarity, the predetermined threshold value includes Low threshold Thr_lowWith high threshold Thr_high, and Thr_low< Thr_high;
If fSimi≥Thr_high, then by current first similarity fSimiArchives kind belonging to corresponding candidate target body is made
For the archives kind of the objective body;
If Thr_low< fSimi< Thr_high, then operations described below is executed:
According to the hair style of the objective body/hair decorations feature, garment ornament and/or vehicles feature construction fisrt feature
Vector, according to the hair style of candidate target body/hair decorations feature, garment ornament and/or vehicles feature construction second feature to
Amount carries out relatedness computation to the first eigenvector and second feature vector, obtains the second similarity;
First similarity and the second similarity are weighted, overall picture similarity is obtained;
Using the archives kind belonging to candidate target body corresponding to the maximum overall picture similarity of similarity value as the target
The archives kind of body.
Further, an optimal technical scheme provided by the invention is:
" according to the face feature, traversal is carried out to stored archive information in the file store built in advance and is searched
Rope obtains the first candidate archive information " the step of specifically include:
Identify whether the head of objective body in the objective body image wears ornaments, and described in adjusting according to recognition result
The weight of face feature;
According to the face feature after the adjustment weight, stored archives in the file store built in advance are believed
Breath is matched, and the first candidate archive information is obtained.
Further, an optimal technical scheme provided by the invention is:
The method further includes:
The file store built in advance is veritified library with preset objective body to be associated with;
Obtain the veritification data that the preset objective body veritifies objective body in library;
Acquired veritification data are matched with the archive information in the file store built in advance, and according to
Matching result determines the correct classification belonging to the objective body.
Further, an optimal technical scheme provided by the invention is:
The method further includes the obtaining step of objective body image and its image recognition training sample, objective body behavior spy
Pre-/alarm step of the analytical procedure of sign, the monitoring step of specific objective body, specific objective body;
The obtaining step of the objective body image and its image recognition training sample includes:According to preset objective body class
Type screens the archive information in the file store built in advance, obtains the corresponding figure of the preset objective body type
Picture and image recognition training sample;
The analytical procedure of the objective body behavioural characteristic includes:The archive information of objective body is counted, and according to system
Meter result generate the action trail of the objective body and obtain the objective body specific region occurrence number, and according to
The statistical result obtains the archive information of other objective bodies in current goal body image, and then is believed according to acquired archives
Breath obtains the behavioural characteristic of other objective bodies;
The monitoring step of the specific objective body includes:It is right based on the archive information in the file store built in advance
The objective body behavioural characteristic of specific objective body is monitored in real time;
Pre-/alarm step of the specific objective body includes:Multiple target body characteristics to an objective body and specific mesh
Multiple features of standard type are compared respectively, and carry out early warning according to comparison result;And/or it obtains and analyzes specific objective body
Action trail so that multiple specific objective bodies are alerted when same period the same area occurs.
Further, an optimal technical scheme provided by the invention is:
The archive information of the objective body includes identity information, gender, age, race, macroscopic features, the body of objective body
Part card image, typical representative image and candid photograph image set;The candid photograph image set includes the mesh got by photographic device
Candid photograph image of the standard type in different time and region.
In second aspect, the electronic record automatically generating device based on target body characteristics in the present invention includes:
Objective body image collection module is configured to use across camera target tracking algorism, obtains objective body image;
Objective body archives classifying module is configured to that acquired objective body image is identified, and obtains objective body spy
Sign, and archives classification is carried out to the objective body according to the target body characteristics;
Objective body archive information memory module is configured to that the archive information of the objective body of same files classification will be belonged to, deposits
It stores up to the corresponding archival region of current class objective body in the file store built in advance.
Further, an optimal technical scheme provided by the invention is:
The target body characteristics include face feature;The objective body archives classifying module includes:
First candidate archive information acquisition submodule, is configured to according to the face feature, to the shelves built in advance
Stored archive information carries out traversal search in case library, obtains the first candidate archive information;
Second candidate archive information acquisition submodule, is configured to according to preset objective body screening conditions to described first
Candidate archive information is screened, and the second candidate archive information is obtained;
Face characteristic matching submodule is configured to obtain candidate target body corresponding to the described second candidate archive information
Face model, and characteristic matching is carried out to the face feature and the face model, and meter is weighted to matching result
Calculation obtains the first similarity;
Archives kind determination sub-module is configured to be compared first similarity and predetermined threshold value, and according to than
Relatively result determines the archives kind of the objective body.
Further, an optimal technical scheme provided by the invention is:
The target body characteristics further include hair style/hair decorations feature, garment ornament and/or vehicles feature;The archives
Classification determination sub-module includes:
First archives kind determination unit, is configured in fSimi≤Thr_lowIn the case of, build a new archives class
Archives kind not as the objective body, wherein the fSimiFor the first similarity, the predetermined threshold value includes Low threshold
Thr_lowWith high threshold Thr_hig, h and Thr_low< Thr_high;
Second archives classification determination unit, is configured in fSimi≥Thr_highIn the case of, by current first similarity
fSimiArchives kind of the archives kind as the objective body belonging to corresponding candidate target body;
Third archives classification determination unit, is configured in Thr_low< fSimi< Thr_highIn the case of, execute following behaviour
Make:
According to the hair style of the objective body/hair decorations feature, garment ornament and/or vehicles feature construction fisrt feature
Vector, according to the hair style of candidate target body/hair decorations feature, garment ornament and/or vehicles feature construction second feature to
Amount carries out relatedness computation to the first eigenvector and second feature vector, obtains the second similarity;
First similarity and the second similarity are weighted, overall picture similarity is obtained;
Using the archives kind belonging to candidate target body corresponding to the maximum overall picture similarity of similarity value as the target
The archives kind of body.
Further, an optimal technical scheme provided by the invention is:
The first candidate archive information acquisition submodule includes:
Face feature weight adjustment unit is configured to identify whether the head of objective body in the objective body image wears
Ornaments, and adjust according to recognition result the weight of the face feature;
Face characteristic matching units is configured to, according to the face feature after the adjustment weight, build in advance to described
Stored archive information is matched in file store, obtains the first candidate archive information.
Further, an optimal technical scheme provided by the invention is:
Described device further includes that archives kind veritifies module;The archives kind veritifies module:
File store is associated with submodule, is configured to the file store built in advance and preset objective body veritifying library pass
Connection;
Data acquisition submodule is veritified, is configured to obtain the veritification number that the preset objective body veritifies objective body in library
According to;
Archives kind veritify submodule, be configured to it is described veritification data acquisition submodule acquired in veritification data with
Archive information in the file store built in advance is matched, and is determined belonging to the objective body according to matching result
Correct classification.
Further, an optimal technical scheme provided by the invention is:
Described device further includes objective body image and its image recognition training sample acquisition module, objective body behavioural characteristic
Analysis module, specific objective body monitoring module, specific objective body be pre-/alarm module;
The objective body image and its image recognition training sample acquisition module, are configured to:According to preset objective body class
Type screens the archive information in the file store built in advance, obtains the corresponding figure of the preset objective body type
Picture and image recognition training sample;
The objective body behavioural characteristic analysis module, is configured to:The archive information of objective body is counted, and according to
Statistical result generates the action trail of the objective body and obtains the objective body in the occurrence number of specific region, Yi Jigen
The archive information of other objective bodies in current goal body image is obtained according to the statistical result, and then according to acquired archives
The behavioural characteristic of other objective bodies described in acquisition of information;
The specific objective body monitoring module, is configured to:Based on the archive information in the file store built in advance,
The objective body behavioural characteristic of specific objective body is monitored in real time;
The specific objective body is pre-/alarm module, it is configured to:To multiple target body characteristics of an objective body with it is specific
Multiple features of objective body are compared respectively, and carry out early warning according to comparison result;And/or it obtains and analyzes specific objective
The action trail of body, so that multiple specific objective bodies are alerted when same period the same area occurs.
Storage device in the third aspect, the present invention is stored with a plurality of program, and described program is suitable for being loaded by processor
And it executes to realize the electronic record automatic generation method based on target body characteristics described in above-mentioned technical proposal.
In fourth aspect, the processing unit in the present invention includes:
Processor is adapted for carrying out each program;
Storage device is suitable for storing a plurality of program;
Described program be suitable for loaded by processor and executed with realize described in above-mentioned technical proposal based on target body characteristics
Electronic record automatic generation method.
Compared with the immediate prior art, above-mentioned technical proposal at least has the advantages that:
1, the electronic record automatic generation method based on target body characteristics can will belong to same files classification in the present invention
Objective body archive information, store into the file store built in advance the corresponding archival region of current class objective body so that every
The archive information of a archival region storage is the archive information of same archives kind objective body, realizes " one grade a kind of ", works as target
Body can be realized " one grade of a people " when behaving.Meanwhile the dynamic portrait library generated based on above-mentioned electronic record generation method,
Without the mode searched for one by one, the attribute letter of the target person is searched in dynamic portrait library after determining target person
Breath, such as the number of target person, so that it may to obtain whole archive informations of the target person.
2, the archives that the electronic record automatic generation method based on target body characteristics will can also in advance be built in the present invention
Library and preset objective body are veritified library and be associated withs, such as the target person archive information library that public security organ announces, and based on objective body core
The veritification data tested in library veritify the archive information of objective body in file store, it is ensured that the archive information of objective body is correct
Filing.
3, the electronic record automatic generation method based on target body characteristics can also believe the archives of objective body in the present invention
Cease it is for statistical analysis obtain the behavioural characteristic of objective body, and then depth analysis is carried out to objective body according to behavioural characteristic.Example
Such as, the action trail based on objective body can analyze activity time and zone of action of the objective body in specific region.Based on every
The number energy of a collected objective body of photographic device accesses objective body in particular place, such as station and Internet bar's appearance
Number.Based on the archives kind of other objective bodies in objective body image, it can obtain also occurring which mesh around objective body
Standard type.Archives kind based on fresh target body so that file store operation maintenance personnel can rapidly search for fresh target body.
Description of the drawings
Fig. 1 is a kind of main step of the electronic record automatic generation method based on target body characteristics in the embodiment of the present invention
Rapid schematic diagram;
Fig. 2 is a kind of main knot of the electronic record automatically generating device based on target body characteristics in the embodiment of the present invention
Structure schematic diagram.
Specific implementation mode
The preferred embodiment of the present invention described with reference to the accompanying drawings.It will be apparent to a skilled person that this
A little embodiments are used only for explaining the technical principle of the present invention, it is not intended that limit the scope of the invention.
Refering to attached drawing 1, it is automatic that Fig. 1 illustrates a kind of electronic record based on target body characteristics in the present embodiment
The key step of generation method.As shown in Figure 1, the electronic record automatic generation method based on target body characteristics in the present embodiment
Include the following steps:
Step S101:Using across camera target tracking algorism, objective body image is obtained.
Objective body image is obtained using across camera target tracking algorism in the present embodiment, spatial dimension conjunction can be obtained
Reason, the image or video information of the identical objective body of time sequencing.
Step S102:Acquired objective body image is identified, obtains target body characteristics, and according to objective body spy
Sign carries out archives classification to objective body.
Specifically, target body characteristics may include the face feature of objective body, hair style/hair decorations feature, clothes in the present embodiment
Adorn feature and/or vehicles feature.Wherein, vehicles feature refers to the type of vehicle and face that objective body uses
The features such as color.
Can archives classification be carried out to objective body as steps described below in the present embodiment:
Step S1021:According to face feature, stored archive information in the file store that builds in advance is traversed
Search obtains the first candidate archive information.
The first the rapid of candidate archive information is obtained in the present embodiment to specifically include:First, target in objective body image is identified
Whether the head of body wears ornaments, such as glasses, mask and cap, and the weight of face feature is adjusted according to recognition result.So
Afterwards, according to the face feature after adjustment weight, stored archive information in the file store that builds in advance is matched, is obtained
To the first candidate archive information, i.e., it is weighted matched mode using to face feature in the present embodiment, obtains first and wait
Select archive information.
Step S1022:It is screened according to the candidate archive information of preset objective body screening conditions pair first, obtains the
Two candidate archive informations.
The gender that preset objective body screening conditions may include objective body in the present embodiment is preset sex types,
And/or it is preset ethnic type that the age of objective body, which is in preset age bracket and/or the race of objective body,.
Step S1023:Obtain the face model of candidate target body corresponding to the second candidate archive information.
Step S1024:Characteristic matching is carried out to face feature and face model, and matching result is weighted
Obtain the first similarity.
Specifically, face model is split first, the feature of each face in face model is extracted, then to mesh
The face feature of standard type is matched with the face feature extracted, and finally matching result is weighted to obtain first
Similarity.
Step S1025:First similarity and predetermined threshold value are compared, and objective body is determined according to comparison result
Archives kind.
Predetermined threshold value includes Low threshold T in the present embodimenthr_lowWith high threshold Thr_high, and Thr_low< Thr_high, it is based on
This first similarity fSimiComparison result with predetermined threshold value includes three kinds of results:fSimi≤Thr_low, fSimi≥Thr_high,
Thr_low< fSimi< Thr_high。
Work as fSimi≤Thr_lowWhen show that objective body is the archives kind not stored in a file store, i.e. the objective body institute
It is a new classification to belong to archives kind, can build archives kind of the new archives kind as objective body at this time.
Work as fSimi> Thr_highWhen show the archives class of candidate target body corresponding to objective body and current first similarity
It is not identical, therefore can be using the archives kind of the candidate target body corresponding to current first similarity as the archives of objective body
Classification.
Work as Thr_low< fSimi< Thr_highWhen show also accurately identify objective body in the case of being based only upon face feature
Archives kind, at this time can the hair style based on objective body/hair decorations feature, garment ornament and/or vehicles feature continue to sentence
The archives kind of disconnected objective body, specifically comprises the following steps:
First, according to the hair style of objective body/hair decorations feature, garment ornament and/or vehicles feature construction fisrt feature
Vector, according to the hair style of candidate target body/hair decorations feature, garment ornament and/or vehicles feature construction second feature to
Amount carries out relatedness computation to first eigenvector and second feature vector, obtains the second similarity.
Secondly, to the first similarity fSimiWith the second similarity cSimiIt is weighted, obtains overall picture similarity.This reality
Apply overall picture similarity w in exampleSimiAs shown in following formula (1):
wSimi=α × fSimi+β×cSimi (1)
Each meaning of parameters is in formula (1):α is the first similarity fSimiWeight, β be the second similarity cSimiPower
Weight, and alpha+beta=1.α=0.6 in a preferred embodiment of the present embodiment, β=0.4.
Finally, using the archives kind belonging to candidate target body corresponding to the maximum overall picture similarity of similarity value as mesh
The archives kind of standard type.
Step S103:The archive information that the objective body of same files classification will be belonged to is stored to the file store built in advance
The corresponding archival region of middle current class objective body.
The archive information of objective body includes that the identity information of objective body, gender, age, race, appearance are special in the present embodiment
Sign, ID Card Image, typical representative image and candid photograph image set.Wherein, it includes being obtained by photographic device to capture image set
Candid photograph image of the objective body arrived in different time and region.
The present embodiment is based on step S103, and the archive information of the objective body of same files classification is stored entirely in one
Specific archival region so that the archive information of each archival region storage is the archive information of same archives kind objective body, real
It is existing " one grade a kind of ", it can be realized when objective body is behaved " one grade of a people ".Based on the above-mentioned electronics shelves based on target body characteristics
The dynamic portrait library that case automatic generation method is generated, without the mode searched for one by one, after determining target person
The attribute information that the target person is searched in dynamic portrait library, such as the number of target person, so that it may to obtain the target person
Whole archive informations.
Further, in this embodiment the electronic record automatic generation method shown in FIG. 1 based on target body characteristics may be used also
The archive information in file store is veritified with veritifying library based on preset objective body, to ensure that the archives of objective body are sorted out
Correct step, specifically includes:
Step S201:The file store built in advance is veritified library with preset objective body to be associated with.Objective body in the present embodiment
For people when, preset objective body veritifies the target person archive information library that public security organ's announcement may be used in library.
Step S202:Obtain the veritification data that preset objective body veritifies objective body in library.
Step S203:Acquired veritification data are matched with the archive information in the file store built in advance, and
The correct classification belonging to objective body is determined according to matching result.
Specifically, when veritification data are matched with archive information, that is, belong to the data letter of same archives kind objective body
Breath, then do not change the archives kind of current goal body.When veritifying data with archive information mismatch, that is, it is not belonging to same shelves
The archives kind of objective body is then revised as veritifying the archives kind corresponding to data by the data information of case classification objective body.
For example, when objective body is behaved, if veritifying the archive information that data are target person A, currently carried out with veritification data
The archive information of matching judgment is also the archive information of target person A, then need not change the archives class of current goal personnel
Not.When objective body is behaved, if veritifying the archive information that data are target person A, currently matching judgment is carried out with veritification data
Archive information is also the archive information of target person B, then the archives kind of current goal personnel is revised as target person A's
Archives kind, and the archive information of current goal personnel is stored into file store the corresponding archival regions of target person A.
Further, in this embodiment the electronic record automatic generation method shown in FIG. 1 based on target body characteristics also wraps
Include the obtaining step of objective body image and its image recognition training sample, the analytical procedure of objective body behavioural characteristic, specific objective
The monitoring step of body, pre-/alarm step of specific objective body.
Specifically, the obtaining step of objective body image and its image recognition training sample includes in the present embodiment:According to pre-
If objective body type, the archive information in the file store that builds in advance is screened, the preset objective body type is obtained
Corresponding image and image recognition training sample.Objective body image can be collected automatically based on electronic record and carries out image knowledge
Other training sample image.
The analytical procedure of objective body behavioural characteristic includes in the present embodiment:The archive information of objective body is counted, and
Occurrence number of the action trail of objective body with acquisition objective body in specific region is generated according to statistical result, and according to system
The archive information that result obtains other objective bodies in current goal body image is counted, and then is obtained according to acquired archive information
The behavioural characteristic of other objective bodies.Meanwhile after it is a new archives kind to determine the affiliated archives kind of objective body, behavior
Feature can also include the archives kind of fresh target body.Action trail based on objective body can analyze objective body in given zone
The activity time in domain and zone of action.Based on the number energy of the collected objective body of each photographic device, accesses objective body and exist
Particular place, such as station and Internet bar's appearance number.It, can based on the archives kind of other objective bodies in objective body image
Obtain also occurring which objective body around objective body.Archives kind based on fresh target body so that file store operation maintenance personnel
It can rapidly search for fresh target body.For example, when objective body is behaved, the shelves based on other objective bodies in objective body image
Case classification can obtain the same administrative staff of target person.When objective body is behaved, the archives kind archives based on fresh target body
Library operation maintenance personnel can rapidly search for the stranger into specific region.
The monitoring step of specific objective body includes in the present embodiment:Based on the archive information in the file store built in advance,
The objective body behavioural characteristic of specific objective body is monitored in real time.Electronic record is based on can such as to endanger to specific objective body
The real time monitoring of dangerous personage.
Pre-/alarm step of specific objective body includes in the present embodiment:To multiple target body characteristics of an objective body with
Multiple features of specific objective body are compared respectively, and carry out early warning according to comparison result, can be according to multiple aspect ratios
Whether objective body, which is the specific objective body of emphasis monitoring, is judged to result, to carry out early warning according to comparison result.Further
Ground, the present embodiment can also obtain and analyze the action trail of specific objective body, so that multiple specific objective bodies are in same a period of time
Between section the same area alerted when occurring, i.e., ought be labeled as having the specific objective body of same alike result that group occurs in advance
It can be alerted in time when rally activity.
Although each step is described in the way of above-mentioned precedence in above-described embodiment, ability
Field technique personnel are appreciated that realize the effect of the present embodiment, are held not necessarily in such order between different steps
Row (parallel) simultaneously can be executed or be executed with reverse order, these simple variations are all in protection scope of the present invention
Within.
Based on technical concept identical with embodiment of the method, the embodiment of the present invention also provides a kind of based on target body characteristics
Electronic record automatically generating device.Below in conjunction with the accompanying drawings to the electronic record automatically generating device based on target body characteristics
It is specifically described.
Refering to attached drawing 2, it is automatic that Fig. 2 illustrates a kind of electronic record based on target body characteristics in the present embodiment
The primary structure of generating means.As shown in Fig. 2, the electronic record automatically generating device based on target body characteristics in the present embodiment
May include objective body image collection module 11, objective body archives classifying module 12 and objective body archive information memory module 13.
Specifically, objective body image collection module 11 is configurable to use across camera target tracking algorism in the present embodiment, obtains
Take objective body image.Objective body archives classifying module 12 is configurable to the mesh acquired in objective body image collection module 11
Standard type image is identified, and obtains target body characteristics, and carry out archives classification to objective body according to target body characteristics.Objective body
Archive information memory module 13 is configurable to that the archive information of the objective body of same files classification will be belonged to, and stores to advance
The corresponding archival region of current class objective body in the file store of structure.
Further, in this embodiment target body characteristics may include face feature, objective body archives shown in Fig. 2 are returned
Generic module 12 may include the first candidate archive information acquisition submodule, the second candidate archive information acquisition submodule, face spy
Levy matched sub-block and archives kind determination sub-module.Specifically, the first candidate archive information acquisition submodule in the present embodiment
It is configurable to according to face feature, traversal search is carried out to stored archive information in the file store that builds in advance, is obtained
To the first candidate archive information.Second candidate archive information acquisition submodule is configurable to be screened according to preset objective body
The candidate archive information of condition pair first screens, and obtains the second candidate archive information.Face characteristic matching submodule can match
It is set to the face model for obtaining candidate target body corresponding to the second candidate archive information, and to face feature and face model
Characteristic matching is carried out, and matching result is weighted to obtain the first similarity.Archives kind determination sub-module can match
It is set to and the first similarity and predetermined threshold value is compared, and determine the archives kind of objective body according to comparison result.
Further, in this embodiment target body characteristics can also include hair style/hair decorations feature, garment ornament and/or friendship
Logical tool characteristics, and archives kind determination sub-module may include that the first archives kind determination unit, the second archives kind are true
Order member and third archives classification determination unit.Specifically, the first archives kind determination unit is configurable in the present embodiment
In fSimi≤Thr_lowIn the case of, archives kind of the one new archives kind of structure as objective body, wherein fSimiIt is first
Similarity, predetermined threshold value include Low threshold Thr_lowWith high threshold Thr_high, and Thr_low< Thr_high.Second archives kind determines
Unit is configurable in fSimi≥Thr_highIn the case of, by current first similarity fSimiCorresponding candidate target body institute
Archives kind of the archives kind of category as objective body.Third archives classification determination unit is configurable in Thr_low< fSimi
< Thr_highIn the case of, execute operations described below:First, according to the hair style of objective body/hair decorations feature, garment ornament and/or friendship
Logical tool characteristics build first eigenvector, according to the hair style of candidate target body/hair decorations feature, garment ornament and/or traffic work
Have feature construction second feature vector, relatedness computation is carried out to first eigenvector and second feature vector, obtains second
Similarity.Secondly, the first similarity and the second similarity are weighted, overall picture similarity is obtained.It finally, will be similar
Archives kind of the archives kind as objective body belonging to candidate target body corresponding to the maximum overall picture similarity of angle value.
Further, in this embodiment the first candidate archive information acquisition submodule may include face feature weight tune
Whole unit and face characteristic matching unit.Specifically, face feature weight adjustment unit is configurable to identify in the present embodiment
Whether the head of objective body wears ornaments in objective body image, and the weight of face feature is adjusted according to recognition result.Face are special
Sign matching unit is configurable to according to the face feature after adjustment weight, to stored shelves in the file store that builds in advance
Case information is matched, and the first candidate archive information is obtained.
Further, in this embodiment the electronic record automatically generating device shown in Fig. 2 based on target body characteristics may be used also
To include that archives kind veritifies module, it may include that file store is associated with submodule, veritification data obtain which, which veritifies module,
Submodule and archives kind is taken to veritify submodule.Specifically, be configurable to will be pre- for file store association submodule in the present embodiment
The file store first built is veritified library with preset objective body and is associated with.Veritification data acquisition submodule is configurable to obtain default
Objective body veritify library in objective body veritification data.Archives kind is veritified submodule and is configurable to veritifying data acquisition
Veritification data acquired in submodule are matched with the archive information in the file store built in advance, and according to matching result
Determine the correct classification belonging to objective body.
Further, in this embodiment the electronic record automatically generating device shown in Fig. 2 based on target body characteristics may be used also
To include objective body image and its image recognition training sample acquisition module, objective body behavioural characteristic analysis module, specific objective
Body monitoring module, specific objective body be pre-/alarm module..
Specifically, objective body image and its image recognition training sample acquisition module are configurable in the present embodiment:Root
According to preset objective body type, the archive information in the file store that builds in advance is screened, the preset objective body is obtained
The corresponding image of type and image recognition training sample.
Objective body behavioural characteristic analysis module is configurable in the present embodiment:It unites to the archive information of objective body
Meter, and the action trail of objective body is generated according to statistical result and obtains objective body in the occurrence number of specific region, Yi Jigen
Result obtains the archive information of other objective bodies in current goal body image according to statistics, and then according to acquired archive information
Obtain the behavioural characteristic of other objective bodies.
Specific objective body monitoring module is configurable in the present embodiment:Based on the archives in the file store built in advance
Information monitors the objective body behavioural characteristic of specific objective body in real time.
Specific objective body is pre- in the present embodiment/and alarm module is configurable to:It is special to multiple objective bodies of an objective body
Sign is compared respectively with multiple features of specific objective body, and carries out early warning according to comparison result;And/or it obtains and analyzes
The action trail of specific objective body, so that multiple specific objective bodies are alerted when same period the same area occurs.
It will be understood by those skilled in the art that the above-mentioned electronic record automatically generating device based on target body characteristics also wraps
Include some other known features, such as processor, controller, memory etc., wherein memory is including but not limited to deposited at random
Reservoir, flash memory, read-only memory, programmable read only memory, volatile memory, nonvolatile memory, serial storage
Device, parallel storage or register etc., processor include but not limited to CPLD/FPGA, DSP, arm processor, MIPS processing
Device etc., in order to unnecessarily obscure embodiment of the disclosure, these well known structures are not shown in FIG. 2.
It should be understood that the quantity of the modules in Fig. 2 is only schematical.According to actual needs, each module can be with
With arbitrary quantity.
Based on the above-mentioned electronic record automatic generation method embodiment based on target body characteristics, the embodiment of the present invention also carries
A kind of storage device is supplied, which is stored with a plurality of program, and these programs are suitable for being loaded and being executed by processor
To realize the electronic record automatic generation method based on target body characteristics described in above method embodiment.
Further, the electronic record automatic generation method embodiment based on target body characteristics based on above-mentioned, the present invention are real
It applies example and additionally provides a kind of processing unit, which includes processor and storage device, wherein processor may be adapted to
Each program is executed, storage device may be adapted to store a plurality of program, and these programs may be adapted to by processor load simultaneously
It executes to realize the electronic record automatic generation method based on target body characteristics described in above method embodiment.
It will be understood by those skilled in the art that can adaptively be changed to the module in device in embodiment
And they are arranged in the one or more devices different from the embodiment.It can be the module or unit in embodiment
It is combined into a module or unit, and multiple submodule or subelement can be divided into addition.In addition to such feature
And/or except at least some of process or unit exclude each other, may be used any combinations to this specification (including
Adjoint claim, abstract and attached drawing) disclosed in all features and so disclosed any method or equipment
All processes or unit are combined.Unless expressly stated otherwise, this specification (including adjoint claim, abstract and attached
Figure) disclosed in each feature can be replaced by providing the alternative features of identical, 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 means in the present invention
Within the scope of and form different embodiments.For example, in claims of the present invention, implementation claimed
The one of arbitrary of example mode can use in any combination.
The all parts embodiment of the present invention can be with hardware realization, or to transport on one or more processors
Capable software module is realized, or is realized with combination thereof.It will be understood by those of skill in the art that can be in practice
It is realized using microprocessor or digital signal processor (DSP) in server according to the ... of the embodiment of the present invention, client
The some or all functions of some or all components.The present invention is also implemented as executing side as described herein
Some or all equipment or program of device (for example, PC programs and PC program products) of method.Such this hair of realization
Bright program can be stored on PC readable mediums, or can be with the form of one or more signal.Such signal
It can download and obtain from internet website, either provide on carrier signal or provide in any other forms.
It should be noted that the present invention will be described rather than limits the invention for above-described embodiment, and this
Field technology personnel can design alternative embodiment without departing from the scope of the appended claims.In claim
In, any reference mark between bracket should not be configured to limitations on claims.Word " comprising " is not excluded for depositing
In element or step not listed in the claims.Word "a" or "an" before element does not exclude the presence of multiple
Such element.The present invention can by means of include several different elements hardware and by means of properly programmed PC come
It realizes.In the unit claims listing several devices, several in these devices can be by the same hardware
It embodies.The use of word first, second, and third does not indicate that any sequence.These words can be construed to
Title.
So far, it has been combined preferred embodiment shown in the drawings and describes technical scheme of the present invention, still, ability
Field technique personnel are it is easily understood that protection scope of the present invention is expressly not limited to these specific implementation modes.Without departing from
Under the premise of the principle of the present invention, those skilled in the art can make the relevant technologies feature equivalent change or replacement, this
Technical solution after a little changes or replacement is fallen within protection scope of the present invention.
Claims (15)
1. a kind of electronic record automatic generation method based on target body characteristics, which is characterized in that the method includes:
Using across camera target tracking algorism, objective body image is obtained;
Acquired objective body image is identified, obtains target body characteristics, and according to the target body characteristics to the mesh
Standard type carries out archives classification;
The archive information that the objective body of same files classification will be belonged to stores into the file store built in advance current class target
The corresponding archival region of body.
2. the electronic record automatic generation method according to claim 1 based on target body characteristics, which is characterized in that described
Target body characteristics include face feature;The step of " carrying out archives classification to the objective body according to the target body characteristics ", is specific
Including:
According to the face feature, traversal search is carried out to stored archive information in the file store built in advance, is obtained
To the first candidate archive information;
The described first candidate archive information is screened according to preset objective body screening conditions, obtains the second candidate archives letter
Breath;
Obtain the face model of candidate target body corresponding to the described second candidate archive information;
Characteristic matching is carried out to the face feature and the face model, and matching result is weighted to obtain first
Similarity;
First similarity and predetermined threshold value are compared, and determine the archives class of the objective body according to comparison result
Not.
3. the electronic record automatic generation method according to claim 2 based on target body characteristics, which is characterized in that described
Target body characteristics further include hair style/hair decorations feature, garment ornament and/or vehicles feature;
" first similarity and predetermined threshold value are compared, and determine the archives class of the objective body according to comparison result
Not " the step of, specifically includes:
If fSimi≤Thr_low, then archives kind of the new archives kind as the objective body is built, wherein the fSimi
For the first similarity, the predetermined threshold value includes Low threshold Thr_lowWith high threshold Thr_high, and Thr_low< Thr_high;
If fSimi≥Thr_high, then by current first similarity fSimiArchives kind belonging to corresponding candidate target body is as institute
State the archives kind of objective body;
If Thr_low< fSimi< Thr_high, then operations described below is executed:
According to the hair style of the objective body/hair decorations feature, garment ornament and/or vehicles feature construction first eigenvector,
According to the hair style of candidate target body/hair decorations feature, garment ornament and/or vehicles feature construction second feature vector, to institute
It states first eigenvector and second feature vector carries out relatedness computation, obtain the second similarity;
First similarity and the second similarity are weighted, overall picture similarity is obtained;
Using the archives kind belonging to candidate target body corresponding to the maximum overall picture similarity of similarity value as the objective body
Archives kind.
4. the electronic record automatic generation method according to claim 2 based on target body characteristics, which is characterized in that " root
According to the face feature, traversal search is carried out to stored archive information in the file store built in advance, obtains first
The step of candidate archive information ", specifically includes:
It identifies whether the head of objective body in the objective body image wears ornaments, and the face spy is adjusted according to recognition result
The weight of sign;
According to the face feature after the adjustment weight, stored archive information in the file store built in advance is carried out
Matching obtains the first candidate archive information.
5. according to electronic record automatic generation method of the claim 1-4 any one of them based on target body characteristics, feature
It is, the method further includes:
The file store built in advance is veritified library with preset objective body to be associated with;
Obtain the veritification data that the preset objective body veritifies objective body in library;
Acquired veritification data are matched with the archive information in the file store built in advance, and are tied according to matching
Fruit determines the correct classification belonging to the objective body.
6. according to electronic record automatic generation method of the claim 1-4 any one of them based on target body characteristics, feature
It is, the method further includes the obtaining step of objective body image and its image recognition training sample, objective body behavioural characteristic
Pre-/alarm step of analytical procedure, the monitoring step of specific objective body, specific objective body;
The obtaining step of the objective body image and its image recognition training sample includes:It is right according to preset objective body type
Archive information in the file store built in advance is screened, and the corresponding image of the preset objective body type and figure are obtained
As recognition training sample;
The analytical procedure of the objective body behavioural characteristic includes:The archive information of objective body is counted, and is tied according to statistics
Fruit generates occurrence number of the action trail of the objective body with the acquisition objective body in specific region, and according to the system
The archive information that result obtains other objective bodies in current goal body image is counted, and then institute is obtained according to acquired archive information
State the behavioural characteristic of other objective bodies;
The monitoring step of the specific objective body includes:Based on the archive information in the file store built in advance, to specific
The objective body behavioural characteristic of objective body is monitored in real time;
Pre-/alarm step of the specific objective body includes:Multiple target body characteristics to specific objective body and preset
Multiple features are compared respectively, and carry out early warning according to comparison result;And/or obtain and analyze the behavior of specific objective body
Track, so that multiple specific objective bodies are alerted when same period the same area occurs.
7. according to electronic record automatic generation method of the claim 1-4 any one of them based on target body characteristics, feature
It is,
The archive information of the objective body includes identity information, gender, age, race, macroscopic features, the identity card figure of objective body
Picture, typical representative image and candid photograph image set;The candid photograph image set includes being existed by the objective body that photographic device is got
Different time and the candid photograph image in region.
8. a kind of electronic record automatically generating device based on target body characteristics, which is characterized in that described device includes:
Objective body image collection module is configured to use across camera target tracking algorism, obtains objective body image;
Objective body archives classifying module is configured to that acquired objective body image is identified, obtains target body characteristics, and root
Archives classification is carried out to the objective body according to the target body characteristics;
Objective body archive information memory module is configured to that the archive information of the objective body of same files classification will be belonged to, store to
The corresponding archival region of current class objective body in the file store built in advance.
9. the electronic record automatically generating device according to claim 8 based on target body characteristics, which is characterized in that described
Target body characteristics include face feature;The objective body archives classifying module includes:
First candidate archive information acquisition submodule, is configured to according to the face feature, to the file store built in advance
In stored archive information carry out traversal search, obtain the first candidate archive information;
Second candidate archive information acquisition submodule is configured to according to preset objective body screening conditions to the described first candidate shelves
Case information is screened, and the second candidate archive information is obtained;
Face characteristic matching submodule is configured to obtain the face mould of candidate target body corresponding to the described second candidate archive information
Type, and characteristic matching is carried out to the face feature and the face model, and matching result is weighted to obtain
First similarity;
Archives kind determination sub-module is configured to be compared first similarity and predetermined threshold value, and is tied according to comparing
Fruit determines the archives kind of the objective body.
10. the electronic record automatically generating device according to claim 9 based on target body characteristics, which is characterized in that institute
It further includes hair style/hair decorations feature, garment ornament and/or vehicles feature to state target body characteristics;The archives kind determines son
Module includes:
First archives kind determination unit, is configured in fSimi≤Thr_lowIn the case of, build a new archives kind conduct
The archives kind of the objective body, wherein the fSimiFor the first similarity, the predetermined threshold value includes Low threshold Thr_lowWith
High threshold Thr_high, and Thr_low< Thr_high;
Second archives classification determination unit, is configured in fSimi≥Thr_highIn the case of, by current first similarity fSimiInstitute is right
Answer archives kind of the archives kind as the objective body belonging to candidate target body;
Third archives classification determination unit, is configured in Thr_low< fSimi< Thr_highIn the case of, execute operations described below:
According to the hair style of the objective body/hair decorations feature, garment ornament and/or vehicles feature construction first eigenvector,
According to the hair style of candidate target body/hair decorations feature, garment ornament and/or vehicles feature construction second feature vector, to institute
It states first eigenvector and second feature vector carries out relatedness computation, obtain the second similarity;
First similarity and the second similarity are weighted, overall picture similarity is obtained;
Using the archives kind belonging to candidate target body corresponding to the maximum overall picture similarity of similarity value as the objective body
Archives kind.
11. the electronic record automatically generating device according to claim 8 based on target body characteristics, which is characterized in that institute
Stating the first candidate archive information acquisition submodule includes:
Face feature weight adjustment unit is configured to identify whether the head of objective body in the objective body image wears ornaments,
And the weight of the face feature is adjusted according to recognition result;
Face characteristic matching unit is configured to according to the face feature after the adjustment weight, to the archives built in advance
Stored archive information is matched in library, obtains the first candidate archive information.
12. special according to electronic record automatically generating device of the claim 8-11 any one of them based on target body characteristics
Sign is that described device further includes that archives kind veritifies module;The archives kind veritifies module:
File store is associated with submodule, is configured to the file store built in advance being associated with preset objective body veritification library;
Data acquisition submodule is veritified, is configured to obtain the veritification data that the preset objective body veritifies objective body in library;
Archives kind veritify submodule, be configured to it is described veritification data acquisition submodule acquired in veritification data with it is described pre-
Archive information in the file store first built is matched, and the correct class belonging to the objective body is determined according to matching result
Not.
13. special according to electronic record automatically generating device of the claim 8-11 any one of them based on target body characteristics
Sign is that described device further includes objective body image and its image recognition training sample acquisition module, objective body behavioural characteristic point
Analyse module, specific objective body monitoring module, specific objective body it is pre-/alarm module;
The objective body image and its image recognition training sample acquisition module, are configured to:It is right according to preset objective body type
Archive information in the file store built in advance is screened, and the corresponding image of the preset objective body type and figure are obtained
As recognition training sample;
The objective body behavioural characteristic analysis module, is configured to:The archive information of objective body is counted, and is tied according to statistics
Fruit generates occurrence number of the action trail of the objective body with the acquisition objective body in specific region, and according to the system
The archive information that result obtains other objective bodies in current goal body image is counted, and then institute is obtained according to acquired archive information
State the behavioural characteristic of other objective bodies;
The specific objective body monitoring module, is configured to:Based on the archive information in the file store built in advance, to specific
The objective body behavioural characteristic of objective body is monitored in real time;
The specific objective body is pre-/alarm module, it is configured to:To the multiple target body characteristics and specific objective body of an objective body
Multiple features be compared respectively, and according to comparison result carry out early warning;And/or obtain and analyze the row of specific objective body
For track, so that multiple specific objective bodies are alerted when same period the same area occurs.
14. a kind of storage device, wherein being stored with a plurality of program, which is characterized in that described program is suitable for by processor load simultaneously
It executes to realize electronic record automatic generation method of the claim 1-7 any one of them based on target body characteristics.
15. a kind of processing unit, including:
Processor is adapted for carrying out each program;
Storage device is suitable for storing a plurality of program;
It is characterized in that, described program is suitable for being loaded by processor and being executed to realize claim 1-7 any one of them bases
In the electronic record automatic generation method of target body characteristics.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810166463.3A CN108563675B (en) | 2018-02-28 | 2018-02-28 | Electronic file automatic generation method and device based on target body characteristics |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810166463.3A CN108563675B (en) | 2018-02-28 | 2018-02-28 | Electronic file automatic generation method and device based on target body characteristics |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108563675A true CN108563675A (en) | 2018-09-21 |
CN108563675B CN108563675B (en) | 2021-03-30 |
Family
ID=63531423
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810166463.3A Active CN108563675B (en) | 2018-02-28 | 2018-02-28 | Electronic file automatic generation method and device based on target body characteristics |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108563675B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110008362A (en) * | 2019-03-05 | 2019-07-12 | 青岛海信网络科技股份有限公司 | A kind of case classifying method and device |
CN110163137A (en) * | 2019-05-13 | 2019-08-23 | 深圳市商汤科技有限公司 | A kind of image processing method, device and storage medium |
CN110390031A (en) * | 2019-06-28 | 2019-10-29 | 深圳市商汤科技有限公司 | Information processing method and device, vision facilities and storage medium |
CN111479225A (en) * | 2020-03-17 | 2020-07-31 | 赵军 | Medical intelligent body temperature electronic chest card and method based on Internet of things |
CN113408557A (en) * | 2020-03-17 | 2021-09-17 | 深圳云天励飞技术有限公司 | File merging method and device and electronic equipment |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014180093A1 (en) * | 2013-05-10 | 2014-11-13 | Tencent Technology (Shenzhen) Company Limited | Systems and methods for facial property identification |
CN105404860A (en) * | 2015-11-13 | 2016-03-16 | 北京旷视科技有限公司 | Method and device for managing information of lost person |
CN106776619A (en) * | 2015-11-20 | 2017-05-31 | 百度在线网络技术(北京)有限公司 | Method and apparatus for determining the attribute information of destination object |
CN107590154A (en) * | 2016-07-08 | 2018-01-16 | 阿里巴巴集团控股有限公司 | Object similarity decision method and device based on image recognition |
-
2018
- 2018-02-28 CN CN201810166463.3A patent/CN108563675B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2014180093A1 (en) * | 2013-05-10 | 2014-11-13 | Tencent Technology (Shenzhen) Company Limited | Systems and methods for facial property identification |
CN105404860A (en) * | 2015-11-13 | 2016-03-16 | 北京旷视科技有限公司 | Method and device for managing information of lost person |
CN106776619A (en) * | 2015-11-20 | 2017-05-31 | 百度在线网络技术(北京)有限公司 | Method and apparatus for determining the attribute information of destination object |
CN107590154A (en) * | 2016-07-08 | 2018-01-16 | 阿里巴巴集团控股有限公司 | Object similarity decision method and device based on image recognition |
Non-Patent Citations (1)
Title |
---|
杨洪臣,李苑: "《视频侦查技术》", 31 May 2015, 《中国人民公安大学出版社》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110008362A (en) * | 2019-03-05 | 2019-07-12 | 青岛海信网络科技股份有限公司 | A kind of case classifying method and device |
CN110163137A (en) * | 2019-05-13 | 2019-08-23 | 深圳市商汤科技有限公司 | A kind of image processing method, device and storage medium |
CN110390031A (en) * | 2019-06-28 | 2019-10-29 | 深圳市商汤科技有限公司 | Information processing method and device, vision facilities and storage medium |
CN111479225A (en) * | 2020-03-17 | 2020-07-31 | 赵军 | Medical intelligent body temperature electronic chest card and method based on Internet of things |
CN113408557A (en) * | 2020-03-17 | 2021-09-17 | 深圳云天励飞技术有限公司 | File merging method and device and electronic equipment |
CN113408557B (en) * | 2020-03-17 | 2023-10-13 | 深圳云天励飞技术有限公司 | File merging method and device and electronic equipment |
Also Published As
Publication number | Publication date |
---|---|
CN108563675B (en) | 2021-03-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108563675A (en) | Electronic record automatic generation method and device based on target body characteristics | |
CN102110122B (en) | Method and device for establishing sample picture index table, method and device for filtering pictures and method and device for searching pictures | |
CN108228696A (en) | Research on face image retrieval and system, filming apparatus, computer storage media | |
CN108090508A (en) | A kind of classification based training method, apparatus and storage medium | |
DE112016005006T5 (en) | AUTOMATIC VIDEO EXECUTIVE SUMMARY | |
US9466012B2 (en) | Phantom image classification | |
CN108629413A (en) | Neural network model training, trading activity Risk Identification Method and device | |
CN109583297A (en) | Retina OCT volume data identification method and device | |
US10896319B2 (en) | Person identification device | |
CN114445785A (en) | Internet of things-based litchi insect pest monitoring and early warning method and system and storage medium | |
CN106559634A (en) | For the date storage method and device of traffic block port video monitoring | |
CN107133629B (en) | Picture classification method and device and mobile terminal | |
CN114187036B (en) | Internet advertisement intelligent recommendation management system based on behavior characteristic recognition | |
CN106874951A (en) | A kind of passenger's attention rate ranking method and device | |
JP2022547248A (en) | Scalable architecture for automatic generation of content delivery images | |
CN109766470A (en) | Image search method, device and processing equipment | |
CN109063984A (en) | Risk passenger method, apparatus, computer equipment and storage medium | |
US20220076279A1 (en) | Marketing engine based on traits and characteristics of prospective consumers | |
US20220083587A1 (en) | Systems and methods for organizing an image gallery | |
US20130230219A1 (en) | Systems and methods for efficient comparative non-spatial image data analysis | |
US11151598B2 (en) | Scoring image engagement in digital media | |
US20230153332A1 (en) | Method for verifying and/or correcting geographical map data | |
CN114359783A (en) | Abnormal event detection method, device and equipment | |
CN106803106A (en) | SAR image automatic classification system and method | |
CN112819527A (en) | User grouping processing method and 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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |