CN108062576A - For the method and apparatus of output data - Google Patents
For the method and apparatus of output data Download PDFInfo
- Publication number
- CN108062576A CN108062576A CN201810010610.8A CN201810010610A CN108062576A CN 108062576 A CN108062576 A CN 108062576A CN 201810010610 A CN201810010610 A CN 201810010610A CN 108062576 A CN108062576 A CN 108062576A
- Authority
- CN
- China
- Prior art keywords
- characteristic
- cluster
- similarity
- data
- clusters
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Abstract
The embodiment of the present application discloses the method and apparatus for output data.One specific embodiment of this method includes:The characteristic in target image is extracted, obtains characteristic set;Based on default first similarity threshold, cluster computing is carried out to the characteristic in characteristic set, obtains at least one characteristic cluster;At least one characteristic of the selection for comparing clusters from least one characteristic cluster;Each characteristic in being clustered at least one characteristic for being used to compare clusters, from it is in characteristic set, be not belonging at least one characteristic compared cluster characteristic in, determine with this feature data clusters it is several classes of between similarity be more than the second similarity threshold and be less than the first similarity threshold characteristic as characteristic to be output;From each characteristic to be output, select the first default quantity characteristic and export.This embodiment improves the efficiency classified to mass data.
Description
Technical field
The invention relates to field of computer technology, and in particular to image identification technical field, more particularly, to
The method and apparatus of output data.
Background technology
With the development of computer technology, the field increasingly light of image identification, and image is identified, common side
Method is to utilize image identification model.When being trained to image identification model, it is necessary to be labeled first to sample data.It is existing
Mask method be that artificial screening and mark are carried out to sample data.
The content of the invention
The embodiment of the present application proposes the method and apparatus for output data.
In a first aspect, the embodiment of the present application provides a kind of method for output data, this method includes:Extract target
Characteristic in image obtains characteristic set;Based on default first similarity threshold, in characteristic set
Characteristic carries out cluster computing, obtains at least one characteristic cluster;It selects to use from least one characteristic cluster
In at least one characteristic cluster compared;Each characteristic in being clustered at least one characteristic for being used to compare
According to cluster, from it is in characteristic set, be not belonging in the characteristic of at least one characteristic compared cluster, really
Similarity is more than the second similarity threshold and is less than the feature of the first similarity threshold between fixed several classes of with this feature data clusters
Data as characteristic to be output, wherein, several classes of similarities are between characteristic feature data and characteristic cluster
Similarity degree;From each characteristic to be output, select the first default quantity characteristic and export.
In some embodiments, at least one characteristic compared is selected from least one characteristic cluster
Cluster, including:It is at least one characteristic phylogenetic group by least one characteristic clustering;For at least one feature
Each characteristic phylogenetic group in data clusters group extracts the second default quantity characteristic from this feature data clusters group
According to cluster;From each characteristic cluster extracted, at least one characteristic compared cluster is determined.
In some embodiments, it is at least one characteristic phylogenetic group by least one characteristic clustering, wraps
It includes:The quantity for the characteristic that feature based data clusters include determines at least one quantity section, wherein, at least one number
The numberical range that amount section is covered includes the characteristic that each characteristic cluster at least one characteristic cluster includes
According to quantity;For each quantity section at least one quantity section, by including the quantity of characteristic be in the number
The characteristic clustering combination in amount section is characterized data clusters group.
In some embodiments, from each characteristic cluster extracted, determine to compare at least one
Characteristic clusters, including:Similarity between the class of the definite each characteristic cluster extracted between any two, wherein, class
Between similarity for the similarity degree between characteristic feature data clusters;Similarity between the class determined is more than default the
The corresponding characteristic Cluster merging of similarity is the characteristic cluster for comparing between the class of three similarity thresholds;By really
Similarity is determined as using less than or equal to third phase like the corresponding characteristic cluster of similarity between the class for spending threshold value between the class made
In the characteristic cluster compared.
In some embodiments, characteristic is vector data;And determine each characteristic extracted cluster
Similarity between class between any two, including:Each characteristic in being clustered for each characteristic extracted clusters,
Determine the average characteristics data for the characteristic that this feature data clusters include, wherein, average characteristics data for each feature to
The vector that the average of the vector element of same position in amount is formed;The each average characteristics data determined are two-by-two
Between similarity;Similarity between the average characteristics data determined is determined as between corresponding characteristic cluster
Class between similarity.
In some embodiments, several classes of similarities determine as follows:Determine that characteristic is gathered with characteristic
The similarity of each characteristic in class;Maximum similarity in each similarity determined is determined as several classes of
Similarity.
In some embodiments, from each characteristic to be output, the default quantity characteristic of selection first is simultaneously defeated
Go out, including:Each characteristic in being clustered at least one characteristic for being used to compare clusters, based on the default 4th
Similarity threshold, characteristic to be output corresponding to this feature data clusters carry out cluster computing, obtain this feature data and gather
The similitude clustering of class;The default quantity characteristic of selection first in the characteristic included from each similitude clustering determined
According to and export.
Second aspect, the embodiment of the present application provide a kind of device for output data, which includes:Extraction is single
Member, the characteristic being configured in extraction target image, obtains characteristic set;Arithmetic element is configured to based on pre-
If the first similarity threshold, cluster computing is carried out to the characteristic in characteristic set, obtains at least one characteristic
According to cluster;Selecting unit is configured to select at least one characteristic compared in clustering from least one characteristic
According to cluster;First determination unit, each characteristic being configured in being clustered at least one characteristic for being used to compare
According to cluster, from it is in characteristic set, be not belonging in the characteristic of at least one characteristic compared cluster, really
Similarity is more than the second similarity threshold and is less than the feature of the first similarity threshold between fixed several classes of with this feature data clusters
Data as characteristic to be output, wherein, several classes of similarities are between characteristic feature data and characteristic cluster
Similarity degree;Output unit is configured to from each characteristic to be output, selects the first default quantity characteristic simultaneously
Output.
In some embodiments, selecting unit, including:Division module is configured to cluster at least one characteristic
It is divided at least one characteristic phylogenetic group;Extraction module is configured at least one characteristic phylogenetic group
Each characteristic phylogenetic group extracts the second default quantity characteristic cluster from this feature data clusters group;Determine mould
Block is configured to from each characteristic cluster extracted, determines at least one characteristic compared cluster.
In some embodiments, division module, including:First determination subelement is configured to feature based data clusters
Including characteristic quantity, determine at least one quantity section, wherein, the numberical range that at least one quantity section is covered
The quantity for the characteristic that each characteristic cluster in being clustered comprising at least one characteristic includes;Combine subelement,
Be configured to for each quantity section at least one quantity section, by including the quantity of characteristic be in the quantity
The characteristic clustering combination in section is characterized data clusters group.
In some embodiments, determining module, including:Second determination subelement is configured to definite extracted each
Similarity between the class of a characteristic cluster between any two, wherein, similarity is between characteristic feature data clusters between class
Similarity degree;Merge subelement, similarity is more than default third phase like degree threshold value between being configured to the class that will be determined
The corresponding characteristic Cluster merging of similarity is the characteristic cluster for comparing between class;3rd determination subelement, configuration
Gather for similarity between the class determined to be less than or equal to third phase like the corresponding characteristic of similarity between the class for spending threshold value
Class is determined as the characteristic compared cluster.
In some embodiments, characteristic is vector data;And second determination subelement, further it is configured to:
Each characteristic in being clustered for each characteristic extracted clusters, and determines the spy that this feature data clusters include
The average characteristics data of data are levied, wherein, average characteristics data are the vector element of the same position in each feature vector
The vector that average is formed;The similarity of each average characteristics data determined between any two;By what is determined
Similarity between average characteristics data is determined as similarity between the class between corresponding characteristic cluster.
In some embodiments, further included for the device of output data:Second determination unit is configured to determine feature
The similarity of data and each characteristic in characteristic cluster;3rd determination unit is configured to be determined
Maximum similarity in each similarity is determined as several classes of similarities.
In some embodiments, output unit, including:Computing module is configured at least one for be used to compare
Each characteristic cluster in characteristic cluster, based on default 4th similarity threshold, to this feature data clusters pair
The characteristic to be output answered carries out cluster computing, obtains the similitude clustering of this feature data clusters;Output module is configured to
Selection first is preset quantity characteristic and is exported in the characteristic included from each similitude clustering determined.
The third aspect, the embodiment of the present application provide a kind of electronic equipment, which includes:One or more processing
Device;Storage device, for storing one or more programs;When one or more programs are executed by one or more processors, make
Obtain method of the one or more processors realization as described in any realization method in first aspect.
Fourth aspect, the embodiment of the present application provide a kind of computer readable storage medium, are stored thereon with computer journey
Sequence realizes the method as described in any realization method in first aspect when the computer program is executed by processor.
Method and apparatus provided by the embodiments of the present application for output data, by extracting the characteristic in image,
Characteristic set is obtained, cluster computing then is carried out to characteristic, at least one characteristic cluster is obtained, then from each
Selection is at least one characteristic cluster for comparing in a characteristic cluster, then from it is in characteristic set, do not belong to
In the characteristic of the characteristic cluster for comparing, characteristic to be output is selected, finally from each feature to be output
In data, select the first default quantity characteristic and export, so as to improve the efficiency classified to mass data.
Description of the drawings
By reading the detailed description made to non-limiting example made with reference to the following drawings, the application's is other
Feature, objects and advantages will become more apparent upon:
Fig. 1 is that this application can be applied to exemplary system architecture figures therein;
Fig. 2 is the flow chart according to one embodiment of the method for output data of the application;
Fig. 3 is the schematic diagram according to an application scenarios of the method for output data of the application;
Fig. 4 is the flow chart according to another embodiment of the method for output data of the application;
Fig. 5 is the structure diagram according to one embodiment of the device for output data of the application;
Fig. 6 is adapted for the structure diagram of the computer system of the electronic equipment for realizing the embodiment of the present application.
Specific embodiment
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining related invention rather than the restriction to the invention.It also should be noted that in order to
Convenient for description, illustrated only in attached drawing and invent relevant part with related.
It should be noted that in the case where there is no conflict, the feature in embodiment and embodiment in the application can phase
Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
Fig. 1 shows the implementation of the method for output data that can apply the application or the device for output data
The exemplary system architecture 100 of example.
As shown in Figure 1, system architecture 100 can include terminal device 101,102,103, network 104 and server 105.
Network 104 between terminal device 101,102,103 and server 105 provide communication link medium.Network 104 can be with
Including various connection types, such as wired, wireless communication link or fiber optic cables etc..
User can be interacted with using terminal equipment 101,102,103 by network 104 with server 105, to receive or send out
Send message etc..Various client applications, such as image processing class application, net can be installed on terminal device 101,102,103
The application of page browsing device, the application of shopping class, searching class application, instant messaging tools, mailbox client, social platform software etc..
Terminal device 101,102,103 can be the various electronic equipments with display screen and supported web page browsing, wrap
It includes but is not limited to smart mobile phone, tablet computer, E-book reader, MP3 player (Moving Picture Experts
Group Audio Layer III, dynamic image expert's compression standard audio level 3), MP4 (Moving Picture
Experts Group Audio Layer IV, dynamic image expert's compression standard audio level 4) it is player, on knee portable
Computer and desktop computer etc..
Server 105 can be to provide the server of various services, such as the figure to the upload of terminal device 101,102,103
As the image processing server handled.Image processing server can carry out the data such as the image that receives analyzing etc.
Reason, and handling result (such as testing result of image) is fed back into terminal device.
It should be noted that the method for output data that the embodiment of the present application is provided can be held by server 105
Row, can also be performed by terminal device 101,102,103, correspondingly, server can be arranged at for the device of output data
In 105, it can also be arranged in terminal device 101,102,103.
It should be understood that the number of the terminal device, network and server in Fig. 1 is only schematical.According to realization need
Will, can have any number of terminal device, network and server.It need not be from remotely obtaining in image to be treated
In the case of, above system framework can not include network, and only need terminal device or server.
With continued reference to Fig. 2, the flow of one embodiment of the method for output data according to the application is shown
200.This is used for the method for output data, comprises the following steps:
Step 201, the characteristic in target image is extracted, obtains characteristic set.
In the present embodiment, for method operation electronic equipment (such as the service shown in FIG. 1 thereon of output data
Device or terminal device) characteristic in target image can be extracted.Wherein, target image can be that above-mentioned electronic equipment passes through
Wired connection mode or radio connection are from the image remotely obtained or from the image locally obtained.Target figure
As that can be image in some pre-set image list or some image collection, also, the quantity of target image can be with
It can also be multiple to be one.
Above-mentioned electronic equipment can utilize existing image recognition technology (such as image identification convolutional Neural net
Network), the characteristic in image is extracted, obtains characteristic set.Wherein, characteristic is for some in phenogram picture
The data of the feature in region (such as facial image region or other images of items regions).
It should be noted that the above-mentioned method that characteristic is extracted from image is known in current widely studied and application
Technology, details are not described herein.
Step 202, based on default first similarity threshold, cluster fortune is carried out to the characteristic in characteristic set
It calculates, obtains at least one characteristic cluster.
In the present embodiment, based on the characteristic set obtained in step 201, above-mentioned electronic equipment can utilize existing
Clustering algorithm, based on default first similarity threshold, cluster computing is carried out to the characteristic in characteristic set, is obtained
It is clustered at least one characteristic.Above-mentioned clustering algorithm can be the existing various algorithms for being clustered to data,
Such as single chain algorithm (single linkage), complete chain algorithm (complete linkage), average linkage algorithm
(average linkage) etc..By setting similarity threshold, at least one characteristic can be obtained using above-mentioned clustering algorithm
According to cluster.
Step 203, at least one characteristic of the selection for comparing clusters from least one characteristic cluster.
In the present embodiment, at least one characteristic cluster obtained based on step 202, above-mentioned electronic equipment can be adopted
At least one characteristic of the selection for comparing clusters from least one characteristic cluster in various manners.On for example,
State electronic equipment the default quantity characteristic cluster of random selection can be used as to compare from each characteristic cluster
Characteristic cluster, can also be by each characteristic clustering at least one characteristic Clustering, from each
The characteristic compared cluster is extracted in grouping.
Step 204, each characteristic in being clustered at least one characteristic for being used to compare clusters, from feature
It is in data acquisition system, be not belonging at least one characteristic compared cluster characteristic in, determine with this feature number
It is more than the second similarity threshold and defeated as treating less than the characteristic of the first similarity threshold according to several classes of similarities of cluster
Go out characteristic.
In the present embodiment, based on being selected in step 203 at least one characteristic cluster compared, for
Each characteristic cluster at least one characteristic cluster compared, above-mentioned electronic equipment can be from characteristic data set
It is in conjunction, be not belonging at least one characteristic compared cluster characteristic in, determine with this feature data clusters
Several classes of similarities be more than the second similarity threshold and less than the first similarity threshold characteristic as feature to be output
Data.Wherein, several classes of similarities are for the similarity degree between characteristic feature data and characteristic cluster.Optionally,
Two similarity thresholds could be provided as being less than the first similarity threshold, for example, the first similarity threshold is 0.7, then second is similar
Degree threshold value could be provided as 0.6.In this way, the characteristic at least one characteristic cluster that can compare being not belonging to
In, select characteristic similar in the characteristic in more being clustered with characteristic.In practice, in characteristic cluster
Characteristic can characterize the direct picture of face, then the characteristic to be output determined can characterize face side or
Shielded image.
In some optional realization methods of the present embodiment, several classes of similarities can determine as follows:It is first
First, above-mentioned electronic equipment can determine the similarity of characteristic and each characteristic in characteristic cluster.Then, on
It states electronic equipment and the maximum similarity in each similarity determined is determined as several classes of similarities.
As an example it is supposed that need to calculate characteristic x1 and characteristic cluster y it is several classes of between similarity, wherein, it is special
Levying data clusters y includes tri- characteristics of y1, y2 and y3.X1 and y1, x2 and y2 are calculated respectively, it is similar between x3 and y3
Degree, by the maximum in be calculated three similarities be determined as characteristic x1 and characteristic cluster y it is several classes of between
Similarity.It should be appreciated that the computational methods of above-mentioned similarity can be the existing various algorithms for calculating similarity, such as Pierre
Inferior related coefficient algorithm, Euclidean distance algorithm, cosine similarity algorithm etc..
Optionally, above-mentioned electronic equipment can also determine characteristic and each characteristic in characteristic cluster
Similarity and then calculate the average value of each similarity, the draw value being calculated is determined as several classes of similarities.
Step 205, from each characteristic to be output, select the first default quantity characteristic and export.
In the present embodiment, above-mentioned electronic equipment can select the first default quantity from each characteristic to be output
A characteristic simultaneously exports.
In some optional realization methods of the present embodiment, above-mentioned electronic equipment can be treated from each in accordance with the following steps
In output characteristic data, select the first default quantity characteristic and export:Firstly, at least one spy for comparing
Each characteristic cluster in data clusters is levied, above-mentioned electronic equipment is based on default 4th similarity threshold, to this feature
The corresponding characteristic to be output of data clusters carries out cluster computing, obtains the similitude clustering of this feature data clusters.
Then, above-mentioned electronic equipment selects first to preset from the characteristic that each similitude clustering determined includes
Quantity characteristic.Wherein, above-mentioned electronic equipment can select the first default quantity characteristic, example in various manners
Such as, randomly choose, according to default order (such as the name order of characteristic, storage location order etc.) selection etc..
In practice, after the default quantity characteristic of output first, technical staff can be to the characteristic of output
It is screened and is marked, since the characteristic of output has correspondence with the above-mentioned data clusters for compared with,
When mark, the characteristic cluster that can first compare above-mentioned carries out automatic or manual mark, then by technical staff
The characteristic of output with characteristic cluster corresponding, for compared with is compared, and it is labeled.
With continued reference to Fig. 3, Fig. 3 is a signal according to the application scenarios of the method for output data of the present embodiment
Figure.In the application scenarios of Fig. 3, computer is first identified target image 301, extract in target image on people
The characteristic set 302 of face, wherein, the quantity of target image is multiple.Then, computer is based on default first similarity
Threshold value 0.7 carries out cluster computing to the characteristic in characteristic set, obtains five characteristic clusters 303.Again so
Afterwards, computer selects compare three characteristic clusters 305 in being clustered from five characteristics.Wherein, select
The quantity of characteristic that includes respectively of three characteristics cluster belonging to quantity section it is different.Then, computer is never
Belong in the characteristic 304 of above-mentioned five characteristics cluster, select and characteristic cluster 2, the characteristic in figure
Cluster 3, characteristic cluster 5 corresponding characteristics 306 to be output, wherein, characteristic 1, characteristic 2, characteristic 3
Several classes of similarities with characteristic cluster 2 are more than the second similarity threshold 0.6, characteristic 4, characteristic 5, feature respectively
Several classes of similarities of the data 6 respectively with characteristic cluster 3 are more than the second similarity threshold 0.6, characteristic 7, characteristic
8th, the several classes of similarities of characteristic 9, characteristic 10 respectively with characteristic cluster 5 are more than the second similarity threshold 0.6.Most
Afterwards, computer selects the first default quantity (being 6 in figure) characteristic 307 and is exported from characteristic 306 to be output.
The method that above-described embodiment of the application provides, by extracting the characteristic in image, obtains characteristic data set
It closes, cluster computing then is carried out to characteristic, obtain at least one characteristic cluster, then clustered from each characteristic
It is middle selection for compare at least one characteristic cluster, then from it is in characteristic set, be not belonging to the spy compared
In the characteristic for levying data clusters, characteristic to be output is selected, finally from each characteristic to be output, selects first
Default quantity characteristic simultaneously exports, so as to improve the efficiency classified to mass data.
With further reference to Fig. 4, it illustrates the flows 400 of another embodiment of the method for output data.The use
In the flow 400 of the method for output data, comprise the following steps:
Step 401, the characteristic in target image is extracted, obtains characteristic set.
In the present embodiment, the step 201 in step 401 embodiment corresponding with Fig. 2 is basically identical, and which is not described herein again.
Step 402, based on default first similarity threshold, cluster fortune is carried out to the characteristic in characteristic set
It calculates, obtains at least one characteristic cluster.
In the present embodiment, the step 202 in step 402 embodiment corresponding with Fig. 2 is basically identical, and which is not described herein again.
Step 403, it is at least one characteristic phylogenetic group by least one characteristic clustering.
In the present embodiment, at least one characteristic cluster obtained based on step 402, for the method for output data
At least one characteristic clustering can be at least one characteristic phylogenetic group by the electronic equipment of operation thereon.Its
In, above-mentioned electronic equipment may be employed various modes and cluster each characteristic clustering at least one characteristic
Group, for example, can by comprising the quantity of characteristic be in the characteristic clustering combination in same quantity section into a spy
Sign data clusters group or the default quantity characteristic cluster of random selection are combined, and obtain at least one characteristic
Phylogenetic group.
In some optional realization methods of the present embodiment, following steps may be employed in above-mentioned electronic equipment will at least one
A characteristic clustering is at least one characteristic phylogenetic group:
First, the quantity for the characteristic that above-mentioned electronic equipment feature based data clusters include, determines at least one number
Measure section.Wherein, the numberical range that at least one quantity section is covered includes each spy at least one characteristic cluster
The quantity for the characteristic that sign data clusters include.For example, in each characteristic cluster, certain characteristic clusters the spy included
It is most to levy the quantity n of data, and n is m digits, wherein n and m are natural numbers, then quantity section can be divided into:[1,9]、
[10,99]、[100,999]、…、[10m,n]。
Then, for each quantity section at least one quantity section, above-mentioned electronic equipment can by including spy
The characteristic clustering combination that the quantity of sign data is in the quantity section is characterized data clusters group.Feature e.g., including
The characteristic cluster that the quantity of data is in section [1,9] has 100, then by this 100 characteristic clustering combinations into one
A characteristic phylogenetic group;Including characteristic quantity be in section [10,99] characteristic cluster have 200, then
By this 200 characteristic clustering combinations into another characteristic phylogenetic group, and so on.
Step 404, for each characteristic phylogenetic group at least one characteristic phylogenetic group, from this feature data
The default quantity characteristic cluster of extraction second in phylogenetic group.
In the present embodiment, for each characteristic phylogenetic group at least one characteristic phylogenetic group, above-mentioned electricity
Sub- equipment may be employed various modes (such as it is random, by comprising the modes such as the order of quantity of characteristic) from the spy
Levy the default quantity characteristic cluster of extraction second in data clusters group.
Step 405, from each characteristic cluster extracted, at least one characteristic compared is determined
Cluster.
In the present embodiment, various modes (such as random) may be employed from each spy extracted in above-mentioned electronic equipment
It levies in data clusters, determines at least one characteristic compared cluster.
In some optional realization methods of the present embodiment, following steps may be employed from being extracted in above-mentioned electronic equipment
In each characteristic cluster gone out, at least one characteristic compared cluster is determined:First, above-mentioned electronic equipment is true
Similarity between the class of the fixed each characteristic cluster extracted between any two.Wherein, similarity is used for characteristic feature between class
Similarity degree between data clusters.Then, similarity between the class determined is more than the default 3rd by above-mentioned electronic equipment
The corresponding characteristic Cluster merging of similarity is the characteristic cluster for comparing between the class of similarity threshold.Finally, on
It states electronic equipment and similarity between the class determined is less than or equal to third phase like the corresponding feature of similarity between the class for spending threshold value
Data clusters are determined as the characteristic compared cluster.
In some optional realization methods of the present embodiment, characteristic can be vector data.Above-mentioned electronic equipment
Following steps may be employed and determine similarity between the class of each characteristic extracted cluster between any two:
It is clustered firstly, for each characteristic in each characteristic cluster extracted, above-mentioned electronic equipment
Determine the average characteristics data for the characteristic that this feature data clusters include.Wherein, average characteristics data for each feature to
The vector that the average of the vector element of same position in amount is formed.Illustratively, it is assumed that characteristic cluster A includes vectorWherein,Then characteristic
According to the average characteristics data of the cluster A characteristics includedFor
Then, above-mentioned electronic equipment determines the similarity of each average characteristics data determined between any two.
Finally, the similarity between the average characteristics data determined is determined as corresponding feature by above-mentioned electronic equipment
Similarity between class between data clusters.
Step 406, each characteristic in being clustered at least one characteristic for being used to compare clusters, from feature
It is in data acquisition system, be not belonging at least one characteristic compared cluster characteristic in, determine with this feature number
It is more than the second similarity threshold and defeated as treating less than the characteristic of the first similarity threshold according to several classes of similarities of cluster
Go out characteristic.
In the present embodiment, the step 204 in step 406 embodiment corresponding with Fig. 2 is basically identical, and which is not described herein again.
Step 407, from each characteristic to be output, select the first default quantity characteristic and export.
In the present embodiment, the step 205 in step 407 embodiment corresponding with Fig. 2 is basically identical, and which is not described herein again.
Figure 4, it is seen that compared with the corresponding embodiments of Fig. 2, the method for output data in the present embodiment
Flow 400 highlight from least one characteristic cluster in selection for compare at least one characteristic cluster step
Suddenly, so that selection cluster the characteristic included for the characteristic that compares being capable of more fully characteristic feature data set
It closes, the characteristic included for the characteristic cluster compared for making selection is more representative, improves to largely counting
According to the accuracy and efficiency classified.
With further reference to Fig. 5, as the realization to method shown in above-mentioned each figure, this application provides one kind for exporting number
According to device one embodiment, the device embodiment is corresponding with embodiment of the method shown in Fig. 2, which can specifically answer
For in various electronic equipments.
As shown in figure 5, the present embodiment includes for the device 500 of output data:Extraction unit 501 is configured to carry
The characteristic in target image is taken, obtains characteristic set;Arithmetic element 502 is configured to based on default first phase
Like degree threshold value, cluster computing is carried out to the characteristic in characteristic set, obtains at least one characteristic cluster;Selection
Unit 503 is configured to select at least one characteristic compared cluster in clustering from least one characteristic;The
One determination unit 504, each characteristic being configured in being clustered at least one characteristic for being used to compare cluster,
From it is in characteristic set, be not belonging at least one characteristic compared cluster characteristic in, determine and should
Several classes of similarities of characteristic cluster are more than the second similarity threshold and the characteristic less than the first similarity threshold is made
For characteristic to be output, wherein, several classes of similarities are for the similar journey between characteristic feature data and characteristic cluster
Degree;Output unit 505 is configured to from each characteristic to be output, and the default quantity characteristic of selection first is simultaneously defeated
Go out.
In the present embodiment, extraction unit 501 can extract the characteristic in target image.Wherein, target image can
Be above-mentioned electronic equipment by wired connection mode or radio connection from the image remotely obtained or from this
The image that ground obtains.Target image can be the image in some pre-set image list or some image collection, also,
The quantity of target image is either one or more.
In the present embodiment, the characteristic set obtained based on extraction unit 501, above-mentioned arithmetic element 502 can profits
With existing clustering algorithm, based on default first similarity threshold, the characteristic in characteristic set is clustered
Computing obtains at least one characteristic cluster.
In the present embodiment, at least one characteristic cluster obtained based on arithmetic element 502, above-mentioned selecting unit
503, which may be employed the various modes at least one characteristic of selection for comparing from least one characteristic cluster, gathers
Class.Make for example, above-mentioned electronic equipment can randomly choose default quantity characteristic cluster from each characteristic cluster
To be used for the characteristic compared cluster, each characteristic clustering can also be divided at least one characteristic cluster
Group extracts the characteristic compared cluster from each grouping.
In the present embodiment, clustered based on what selecting unit 503 selected at least one characteristic for comparing, for
For each characteristic cluster at least one characteristic cluster for comparing, above-mentioned first determination unit 504 can be from
It is in characteristic set, be not belonging at least one characteristic compared cluster characteristic in, determine with the spy
Several classes of similarities for levying data clusters are more than the second similarity threshold and less than the characteristic conduct of the first similarity threshold
Characteristic to be output.Wherein, several classes of similarities are for the similarity degree between characteristic feature data and characteristic cluster.
In the present embodiment, above-mentioned output unit 505 can select the first present count from each characteristic to be output
Amount characteristic simultaneously exports.
In some optional realization methods of the present embodiment, selecting unit 503 can include:Division module is (in figure not
Show), it is at least one characteristic phylogenetic group to be configured at least one characteristic clustering;Extraction module (figure
Not shown in), it is configured to for each characteristic phylogenetic group at least one characteristic phylogenetic group, from this feature number
According to the default quantity characteristic cluster of extraction second in phylogenetic group;Determining module (not shown) is configured to from being carried
In each characteristic cluster taken out, at least one characteristic compared cluster is determined.
In some optional realization methods of the present embodiment, division module can include:First determination subelement is (in figure
It is not shown), the quantity for the characteristic that feature based data clusters include is configured to, determines at least one quantity section,
In, the numberical range that at least one quantity section is covered includes each characteristic cluster at least one characteristic cluster
Including characteristic quantity;Subelement (not shown) is combined, is configured at least one quantity section
Each quantity section, by including the quantity of characteristic be in the characteristic clustering combination in the quantity section and be characterized data
Phylogenetic group.
In some optional realization methods of the present embodiment, determining module, including:Second determination subelement is (in figure not
Show), similarity between the class of definite extracted each characteristic cluster between any two is configured to, wherein, phase between class
Like degree for the similarity degree between characteristic feature data clusters;Merge subelement (not shown), be configured to by really
Similarity is like the corresponding characteristic Cluster merging of similarity between the class for spending threshold value more than default third phase between the class made
For the characteristic cluster compared;3rd determination subelement (not shown) is configured to phase between the class that will be determined
It is less than or equal to third phase like degree and is determined as the feature compared like the corresponding characteristic cluster of similarity between the class for spending threshold value
Data clusters.
In some optional realization methods of the present embodiment, characteristic is vector data;And second determine that son is single
Member can be further configured to:Each characteristic in being clustered for each characteristic extracted clusters, and determines
The average characteristics data for the characteristic that this feature data clusters include, wherein, average characteristics data are in each feature vector
Same position vector element the vector that is formed of average;The each average characteristics data determined are between any two
Similarity;Similarity between the average characteristics data determined is determined as the class between corresponding characteristic cluster
Between similarity.
In some optional realization methods of the present embodiment, it can also include for the device 500 of output data:Second
Determination unit (not shown) is configured to determine that characteristic is similar to each characteristic in characteristic cluster
Degree;3rd determination unit (not shown), the maximum similarity being configured in each similarity that will be determined are true
It is set to several classes of similarities.
In some optional realization methods of the present embodiment, output unit 505 can include:Computing module, configuration are used
Each characteristic cluster at least one characteristic cluster for being used to compare, based on default 4th similarity
Threshold value, characteristic to be output corresponding to this feature data clusters carry out cluster computing, obtain the phase of this feature data clusters
Like cluster;It is default to be configured to the selection first from the characteristic that each similitude clustering determined includes for output module
Quantity characteristic simultaneously exports.
The device that above-described embodiment of the application provides extracts the characteristic in image by extraction unit, obtains spy
Data acquisition system is levied, then arithmetic element 502 carries out cluster computing to characteristic, obtains at least one characteristic cluster, with
For at least one characteristic cluster compared, first determines selecting unit 503 for selection from each characteristic cluster afterwards
Unit 504 again from it is in characteristic set, be not belonging to the characteristic compared cluster characteristic in, selection treat it is defeated
Go out characteristic, last output unit 505 selects the first default quantity characteristic simultaneously from each characteristic to be output
Output, so as to improve the efficiency classified to mass data.
Below with reference to Fig. 6, it illustrates suitable for being used for realizing the computer system 600 of the electronic equipment of the embodiment of the present application
Structure diagram.Electronic equipment shown in Fig. 6 is only an example, to the function of the embodiment of the present application and should not use model
Shroud carrys out any restrictions.
As shown in fig. 6, computer system 600 includes central processing unit (CPU) 601, it can be read-only according to being stored in
Program in memory (ROM) 602 or be loaded into program in random access storage device (RAM) 603 from storage part 608 and
Perform various appropriate actions and processing.In RAM 603, also it is stored with system 600 and operates required various programs and data.
CPU 601, ROM 602 and RAM 603 are connected with each other by bus 604.Input/output (I/O) interface 605 is also connected to always
Line 604.
I/O interfaces 605 are connected to lower component:Importation 606 including keyboard, mouse etc.;It is penetrated including such as cathode
The output par, c 607 of spool (CRT), liquid crystal display (LCD) etc. and loud speaker etc.;Storage part 608 including hard disk etc.;
And the communications portion 609 of the network interface card including LAN card, modem etc..Communications portion 609 via such as because
The network of spy's net performs communication process.Driver 610 is also according to needing to be connected to I/O interfaces 605.Detachable media 611, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on driver 610, as needed in order to read from it
Computer program be mounted into as needed storage part 608.
Particularly, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart description
Software program.For example, embodiment of the disclosure includes a kind of computer program product, including being carried on computer-readable medium
On computer program, which includes for the program code of the method shown in execution flow chart.In such reality
It applies in example, which can be downloaded and installed from network by communications portion 609 and/or from detachable media
611 are mounted.When the computer program is performed by central processing unit (CPU) 601, perform what is limited in the present processes
Above-mentioned function.It should be noted that computer-readable medium described herein can be computer-readable signal media or
Computer readable storage medium either the two any combination.Computer readable storage medium for example can be --- but
It is not limited to --- electricity, magnetic, optical, electromagnetic, system, device or the device of infrared ray or semiconductor or arbitrary above combination.
The more specific example of computer readable storage medium can include but is not limited to:Electrical connection with one or more conducting wires,
Portable computer diskette, hard disk, random access storage device (RAM), read-only memory (ROM), erasable type may be programmed read-only deposit
Reservoir (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-ROM), light storage device, magnetic memory
Part or above-mentioned any appropriate combination.In this application, computer readable storage medium can any be included or store
The tangible medium of program, the program can be commanded the either device use or in connection of execution system, device.And
In the application, computer-readable signal media can include the data letter propagated in a base band or as a carrier wave part
Number, wherein carrying computer-readable program code.Diversified forms may be employed in the data-signal of this propagation, including but not
It is limited to electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be computer
Any computer-readable medium beyond readable storage medium storing program for executing, the computer-readable medium can send, propagate or transmit use
In by instruction execution system, device either device use or program in connection.It is included on computer-readable medium
Program code any appropriate medium can be used to transmit, include but not limited to:Wirelessly, electric wire, optical cable, RF etc., Huo Zheshang
Any appropriate combination stated.
It can be with one or more programming languages or its calculating for combining to write to perform the operation of the application
Machine program code, described program design language include object oriented program language-such as Java, Smalltalk, C+
+, further include conventional procedural programming language-such as " C " language or similar programming language.Program code can
Fully to perform on the user computer, partly perform, performed as an independent software package on the user computer,
Part performs or performs on a remote computer or server completely on the remote computer on the user computer for part.
In the situation for being related to remote computer, remote computer can pass through the network of any kind --- including LAN (LAN)
Or wide area network (WAN)-be connected to subscriber computer or, it may be connected to outer computer (such as utilizes Internet service
Provider passes through Internet connection).
Flow chart and block diagram in attached drawing, it is illustrated that according to the system of the various embodiments of the application, method and computer journey
Architectural framework in the cards, function and the operation of sequence product.In this regard, each box in flow chart or block diagram can generation
The part of one module of table, program segment or code, the part of the module, program segment or code include one or more use
In the executable instruction of logic function as defined in realization.It should also be noted that it is marked at some as in the realization replaced in box
The function of note can also be occurred with being different from the order marked in attached drawing.For example, two boxes succeedingly represented are actually
It can perform substantially in parallel, they can also be performed in the opposite order sometimes, this is depending on involved function.Also to note
Meaning, the combination of each box in block diagram and/or flow chart and the box in block diagram and/or flow chart can be with holding
The dedicated hardware based system of functions or operations as defined in row is realized or can use specialized hardware and computer instruction
Combination realize.
Being described in unit involved in the embodiment of the present application can be realized by way of software, can also be by hard
The mode of part is realized.Described unit can also be set in the processor, for example, can be described as:A kind of processor bag
Include extraction unit, arithmetic element, selecting unit, the first determination unit and output unit.Wherein, the title of these units is at certain
In the case of do not form restriction to the unit in itself, for example, extraction unit is also described as " in extraction target image
Characteristic obtains the unit of characteristic set ".
As on the other hand, present invention also provides a kind of computer-readable medium, which can be
Included in electronic equipment described in above-described embodiment;Can also be individualism, and without be incorporated the electronic equipment in.
Above computer readable medium carries one or more program, when said one or multiple programs are held by the electronic equipment
During row so that the electronic equipment:The characteristic in target image is extracted, obtains characteristic set;Based on default first
Similarity threshold carries out cluster computing to the characteristic in characteristic set, obtains at least one characteristic cluster;From
At least one characteristic of the selection for comparing clusters at least one characteristic cluster;For being used for compare at least one
Each characteristic cluster in a characteristic cluster, from it is in characteristic set, be not belonging to compare at least one
In the characteristic of a characteristic cluster, determine with this feature data clusters it is several classes of between similarity more than the second similarity threshold
It is worth and is less than the characteristic of the first similarity threshold as characteristic to be output;From each characteristic to be output, choosing
It selects the first default quantity characteristic and exports.
The preferred embodiment and the explanation to institute's application technology principle that above description is only the application.People in the art
Member should be appreciated that invention scope involved in the application, however it is not limited to the technology that the particular combination of above-mentioned technical characteristic forms
Scheme, while should also cover in the case where not departing from foregoing invention design, it is carried out by above-mentioned technical characteristic or its equivalent feature
The other technical solutions for being combined and being formed.Such as features described above has similar work(with (but not limited to) disclosed herein
The technical solution that the technical characteristic of energy is replaced mutually and formed.
Claims (16)
1. a kind of method for output data, including:
The characteristic in target image is extracted, obtains characteristic set;
Based on default first similarity threshold, cluster computing is carried out to the characteristic in the characteristic set, is obtained
At least one characteristic cluster;
At least one characteristic of the selection for comparing clusters from least one characteristic cluster;
Each characteristic during at least one characteristic for described for comparing clusters clusters, from the characteristic
It is in set, be not belonging in the characteristic at least one characteristic cluster compared, determine and this feature number
It is more than the second similarity threshold and defeated as treating less than the characteristic of the first similarity threshold according to several classes of similarities of cluster
Go out characteristic, wherein, several classes of similarities are for the similarity degree between characteristic feature data and characteristic cluster;
From each characteristic to be output, select the first default quantity characteristic and export.
2. according to the method described in claim 1, wherein, the selection from least one characteristic cluster for than
Compared at least one characteristic cluster, including:
It is at least one characteristic phylogenetic group by least one characteristic clustering;
For each characteristic phylogenetic group at least one characteristic phylogenetic group, from this feature data clusters group
The default quantity characteristic cluster of extraction second;
From each characteristic cluster extracted, at least one characteristic compared cluster is determined.
3. according to the method described in claim 2, wherein, described by least one characteristic clustering is at least one
A characteristic phylogenetic group, including:
The quantity for the characteristic that feature based data clusters include determines at least one quantity section, wherein, described at least one
Each characteristic cluster that the numberical range that a quantity section is covered is included at least one characteristic cluster includes
Characteristic quantity;
For each quantity section at least one quantity section, by including the quantity of characteristic be in the quantity
The characteristic clustering combination in section is characterized data clusters group.
4. it according to the method described in claim 2, wherein, in the cluster from each characteristic extracted, determines to use
It is clustered at least one characteristic compared, including:
Similarity between the class of the definite each characteristic cluster extracted between any two, wherein, similarity is used for table between class
Levy the similarity degree between characteristic cluster;
Similarity between the class determined is more than default third phase like the corresponding characteristic of similarity between the class for spending threshold value
Cluster merging is the characteristic cluster for comparing;
Similarity between the class determined is less than or equal to the third phase like the corresponding characteristic of similarity between the class for spending threshold value
It is determined as the characteristic compared cluster according to cluster.
5. according to the method described in claim 4, wherein, characteristic is vector data;And
The similarity between determining the class of each characteristic extracted cluster between any two, including:
Each characteristic in being clustered for each characteristic extracted clusters, and determines that this feature data clusters include
Characteristic average characteristics data, wherein, average characteristics data be each feature vector in same position element vector
The vector that the average of element is formed;
The similarity of each average characteristics data determined between any two;
Similarity between the average characteristics data determined is determined as phase between the class between corresponding characteristic cluster
Like degree.
6. according to the method described in claim 1, wherein, several classes of similarities determine as follows:
Determine the similarity of characteristic and each characteristic in characteristic cluster;
Maximum similarity in each similarity determined is determined as several classes of similarities.
7. according to the method described in one of claim 1-6, wherein, it is described from each characteristic to be output, select first
Default quantity characteristic simultaneously exports, including:
Each characteristic during at least one characteristic for described for comparing clusters clusters, based on the default 4th
Similarity threshold, characteristic to be output corresponding to this feature data clusters carry out cluster computing, obtain this feature data and gather
The similitude clustering of class;
Selection first is preset quantity characteristic and is exported in the characteristic included from each similitude clustering determined.
8. a kind of device for output data, including:
Extraction unit, the characteristic being configured in extraction target image, obtains characteristic set;
Arithmetic element is configured to based on default first similarity threshold, to the characteristic in the characteristic set
Cluster computing is carried out, obtains at least one characteristic cluster;
Selecting unit is configured to select at least one characteristic compared from least one characteristic cluster
According to cluster;
First determination unit, each characteristic being configured to during at least one characteristic for described for comparing clusters
According to cluster, from it is in the characteristic set, be not belonging to it is described for the feature of at least one characteristic cluster compared
In data, determine with this feature data clusters it is several classes of between similarity be more than the second similarity threshold and be less than the first similarity threshold
The characteristic of value as characteristic to be output, wherein, several classes of similarities are gathered for characteristic feature data with characteristic
Similarity degree between class;
Output unit is configured to from each characteristic to be output, is selected the first default quantity characteristic and is exported.
9. device according to claim 8, wherein, the selecting unit, including:
Division module, it is at least one characteristic phylogenetic group to be configured at least one characteristic clustering;
Extraction module is configured to for each characteristic phylogenetic group at least one characteristic phylogenetic group, from
The default quantity characteristic cluster of extraction second in this feature data clusters group;
Determining module is configured to from each characteristic cluster extracted, determines at least one spy compared
Levy data clusters.
10. device according to claim 9, wherein, the division module, including:
First determination subelement is configured to the quantity for the characteristic that feature based data clusters include, and determines at least one
Quantity section, wherein, the numberical range that at least one quantity section is covered includes at least one characteristic cluster
In the quantity of characteristic that includes of each characteristic cluster;
Combine subelement, be configured to for each quantity section at least one quantity section, by including feature
The characteristic clustering combination that the quantity of data is in the quantity section is characterized data clusters group.
11. device according to claim 9, wherein, the determining module, including:
Second determination subelement is configured to similar between the class of definite extracted each characteristic cluster between any two
Degree, wherein, similarity is for the similarity degree between characteristic feature data clusters between class;
Merge subelement, similarity is more than default third phase like phase between the class for spending threshold value between being configured to the class that will be determined
It is the characteristic cluster for comparing like corresponding characteristic Cluster merging is spent;
3rd determination subelement, similarity is less than or equal to the third phase like degree threshold value between being configured to the class that will be determined
The corresponding characteristic cluster of similarity is determined as the characteristic compared cluster between class.
12. according to the devices described in claim 11, wherein, characteristic is vector data;And
Second determination subelement, is further configured to:
Each characteristic in being clustered for each characteristic extracted clusters, and determines that this feature data clusters include
Characteristic average characteristics data, wherein, average characteristics data be each feature vector in same position element vector
The vector that the average of element is formed;
The similarity of each average characteristics data determined between any two;
Similarity between the average characteristics data determined is determined as phase between the class between corresponding characteristic cluster
Like degree.
13. device according to claim 8, wherein, described device further includes:
Second determination unit is configured to determine the similarity of characteristic and each characteristic in characteristic cluster;
3rd determination unit, the maximum similarity being configured in each similarity that will be determined are determined as several classes of phases
Like degree.
14. according to the device described in one of claim 8-13, wherein, the output unit, including:
Computing module, each characteristic being configured to during at least one characteristic for described for comparing clusters are gathered
Class, based on default 4th similarity threshold, characteristic to be output corresponding to this feature data clusters carries out cluster computing,
Obtain the similitude clustering of this feature data clusters;
Output module is configured to the default quantity of the selection first from the characteristic that each similitude clustering determined includes
A characteristic simultaneously exports.
15. a kind of electronic equipment, including:
One or more processors;
Storage device, for storing one or more programs,
When one or more of programs are performed by one or more of processors so that one or more of processors are real
The now method as described in any in claim 1-7.
16. a kind of computer readable storage medium, is stored thereon with computer program, wherein, when which is executed by processor
Realize the method as described in any in claim 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810010610.8A CN108062576B (en) | 2018-01-05 | 2018-01-05 | Method and apparatus for output data |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810010610.8A CN108062576B (en) | 2018-01-05 | 2018-01-05 | Method and apparatus for output data |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108062576A true CN108062576A (en) | 2018-05-22 |
CN108062576B CN108062576B (en) | 2019-05-03 |
Family
ID=62141291
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810010610.8A Active CN108062576B (en) | 2018-01-05 | 2018-01-05 | Method and apparatus for output data |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108062576B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210535A (en) * | 2019-05-21 | 2019-09-06 | 北京市商汤科技开发有限公司 | Neural network training method and device and image processing method and device |
CN110245580A (en) * | 2019-05-24 | 2019-09-17 | 北京百度网讯科技有限公司 | A kind of method, apparatus of detection image, equipment and computer storage medium |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101523412A (en) * | 2006-10-11 | 2009-09-02 | 惠普开发有限公司 | Face-based image clustering |
CN102930296A (en) * | 2012-11-01 | 2013-02-13 | 长沙纳特微视网络科技有限公司 | Image identifying method and device |
CN104391879A (en) * | 2014-10-31 | 2015-03-04 | 小米科技有限责任公司 | Method and device for hierarchical clustering |
CN105469108A (en) * | 2015-11-17 | 2016-04-06 | 深圳先进技术研究院 | Clustering method, clustering system, clustering result evaluation method and clustering result evaluation system based on biological data |
CN105654039A (en) * | 2015-12-24 | 2016-06-08 | 小米科技有限责任公司 | Image processing method and device |
CN106228188A (en) * | 2016-07-22 | 2016-12-14 | 北京市商汤科技开发有限公司 | Clustering method, device and electronic equipment |
CN106446797A (en) * | 2016-08-31 | 2017-02-22 | 腾讯科技(深圳)有限公司 | Image clustering method and device |
CN106503656A (en) * | 2016-10-24 | 2017-03-15 | 厦门美图之家科技有限公司 | A kind of image classification method, device and computing device |
CN106844748A (en) * | 2017-02-16 | 2017-06-13 | 湖北文理学院 | Text Clustering Method, device and electronic equipment |
CN107016004A (en) * | 2016-01-28 | 2017-08-04 | 百度在线网络技术(北京)有限公司 | Image processing method and device |
-
2018
- 2018-01-05 CN CN201810010610.8A patent/CN108062576B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101523412A (en) * | 2006-10-11 | 2009-09-02 | 惠普开发有限公司 | Face-based image clustering |
CN102930296A (en) * | 2012-11-01 | 2013-02-13 | 长沙纳特微视网络科技有限公司 | Image identifying method and device |
CN104391879A (en) * | 2014-10-31 | 2015-03-04 | 小米科技有限责任公司 | Method and device for hierarchical clustering |
CN105469108A (en) * | 2015-11-17 | 2016-04-06 | 深圳先进技术研究院 | Clustering method, clustering system, clustering result evaluation method and clustering result evaluation system based on biological data |
CN105654039A (en) * | 2015-12-24 | 2016-06-08 | 小米科技有限责任公司 | Image processing method and device |
CN107016004A (en) * | 2016-01-28 | 2017-08-04 | 百度在线网络技术(北京)有限公司 | Image processing method and device |
CN106228188A (en) * | 2016-07-22 | 2016-12-14 | 北京市商汤科技开发有限公司 | Clustering method, device and electronic equipment |
CN106446797A (en) * | 2016-08-31 | 2017-02-22 | 腾讯科技(深圳)有限公司 | Image clustering method and device |
CN106503656A (en) * | 2016-10-24 | 2017-03-15 | 厦门美图之家科技有限公司 | A kind of image classification method, device and computing device |
CN106844748A (en) * | 2017-02-16 | 2017-06-13 | 湖北文理学院 | Text Clustering Method, device and electronic equipment |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110210535A (en) * | 2019-05-21 | 2019-09-06 | 北京市商汤科技开发有限公司 | Neural network training method and device and image processing method and device |
CN110245580A (en) * | 2019-05-24 | 2019-09-17 | 北京百度网讯科技有限公司 | A kind of method, apparatus of detection image, equipment and computer storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN108062576B (en) | 2019-05-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108229419A (en) | For clustering the method and apparatus of image | |
CN108446387A (en) | Method and apparatus for updating face registration library | |
CN107944481A (en) | Method and apparatus for generating information | |
CN108509915A (en) | The generation method and device of human face recognition model | |
CN107908789A (en) | Method and apparatus for generating information | |
CN108171191B (en) | Method and apparatus for detecting face | |
CN109376267A (en) | Method and apparatus for generating model | |
CN108764319A (en) | A kind of sample classification method and apparatus | |
CN107958247A (en) | Method and apparatus for facial image identification | |
CN109063653A (en) | Image processing method and device | |
CN108494778A (en) | Identity identifying method and device | |
CN108256591A (en) | For the method and apparatus of output information | |
CN108230346A (en) | For dividing the method and apparatus of image semantic feature, electronic equipment | |
CN109145828A (en) | Method and apparatus for generating video classification detection model | |
CN109993179A (en) | The method and apparatus that a kind of pair of data are clustered | |
CN107220652A (en) | Method and apparatus for handling picture | |
CN108776692A (en) | Method and apparatus for handling information | |
CN109087377A (en) | Method and apparatus for handling image | |
CN108960110A (en) | Method and apparatus for generating information | |
CN109711508A (en) | Image processing method and device | |
CN107729928A (en) | Information acquisition method and device | |
CN108062416B (en) | Method and apparatus for generating label on map | |
CN110209658A (en) | Data cleaning method and device | |
CN108182457A (en) | For generating the method and apparatus of information | |
CN108509994A (en) | character image clustering 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 |