CN107423757A - clustering processing method and device - Google Patents

clustering processing method and device Download PDF

Info

Publication number
CN107423757A
CN107423757A CN201710573089.4A CN201710573089A CN107423757A CN 107423757 A CN107423757 A CN 107423757A CN 201710573089 A CN201710573089 A CN 201710573089A CN 107423757 A CN107423757 A CN 107423757A
Authority
CN
China
Prior art keywords
similarity
probability
class
elements
cluster
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
Application number
CN201710573089.4A
Other languages
Chinese (zh)
Other versions
CN107423757B (en
Inventor
陈志军
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Xiaomi Mobile Software Co Ltd
Original Assignee
Beijing Xiaomi Mobile Software Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing Xiaomi Mobile Software Co Ltd filed Critical Beijing Xiaomi Mobile Software Co Ltd
Priority to CN201710573089.4A priority Critical patent/CN107423757B/en
Publication of CN107423757A publication Critical patent/CN107423757A/en
Application granted granted Critical
Publication of CN107423757B publication Critical patent/CN107423757B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • 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

Abstract

The disclosure provides a kind of clustering processing method and device, and this method includes：Each element and each similarity of element between any two in the second class in the first kind are obtained respectively, obtain M*N similarity, wherein, the first kind includes N number of element including M element, the second class；According to the similarity from big to small, it is determined that the default K similarities；Using cluster probability between two elements corresponding to each similarity in the K similarity, it is of a sort analogy probability to calculate the confidence value between the first kind and second class and the first kind and second class；When the confidence value and the analogy probability are satisfied by preparatory condition, the first kind and second class are merged.In this new clustering method, it is contemplated that the cluster probability between different elements, and then inhomogeneous confidence value and analogy probability are calculated, substantially increase the degree of accuracy of cluster.

Description

Clustering processing method and device
Technical field
This disclosure relates to the information processing technology, more particularly to a kind of clustering processing method and device.
Background technology
Cluster refers to the process of for the set of physics or abstract object to be divided into the multiple classes being made up of similar object.By gathering The cluster that class is generated is the set of one group of data object, and these objects are similar each other to the object in same cluster, with other clusters In object it is different.
In various fields, the application to cluster is all very extensive, such as recognition of face, Products Show etc..It is but existing poly- Class mode is mainly bad using one-to-many, man-to-man mode, Clustering Effect.
The content of the invention
The disclosure provides a kind of clustering processing method and device, for improving Clustering Effect.
According to the first aspect of the embodiment of the present disclosure, there is provided a kind of clustering processing method, including：
Each element and each similarity of element between any two in the second class in the first kind are obtained respectively, and it is individual to obtain M*N Similarity, wherein, the first kind includes M element, the second class includes N number of element, and M, N are the integer more than 0；
According to the similarity from big to small, it is determined that presetting the K similarities, wherein, K is the integer more than 0；
Using cluster probability between two elements corresponding to each similarity in the K similarity, described first is calculated Confidence value and the first kind and second class between class and second class are of a sort analogy probability, its In, the cluster probability is used for the probability for indicating that two elements are identical element；
When the confidence value and the analogy probability are satisfied by preparatory condition, by the first kind and second class Merge.
Alternatively, cluster probability between two elements using corresponding to each similarity in the K similarity, meter It is of a sort class to calculate the confidence value between the first kind and second class and the first kind and second class Before probability, in addition to：
Pending element set is obtained, the cluster calculated in the pending element set between any two element is general Rate, the pending element set include：All elements in all elements and second class in the first kind.
Alternatively, the cluster probability calculated in the pending element set between any two element, including：
Count the similarity between any two element in the pending element set；
According to the similarity between any two element, the cluster probability between any two element is determined.
Alternatively, cluster probability between two elements using corresponding to each similarity in the K similarity, meter The confidence value between the first kind and second class is calculated, including：
Using formulaCalculate the confidence value D between the first kind and second classAB, its In, piCluster probability between two elements in K similarity of expression corresponding to i-th of similarity, i are more than 0 and are less than or equal to K。
Alternatively, cluster probability between two elements using corresponding to each similarity in the K similarity, meter It is of a sort analogy probability to calculate the first kind and second class, including：
Using formulaIt is of a sort analogy probability P to calculate the first kind and second classAB, wherein, piCluster probability between two elements in K similarity of expression corresponding to i-th of similarity, i are more than 0 and are less than or equal to K.
According to the second aspect of the embodiment of the present disclosure, there is provided a kind of clustering processing device, including：
Acquisition module, it is configured to obtain each element in the first kind and each element is between any two in the second class Similarity, M*N similarity is obtained, wherein, the first kind includes M element, the second class includes N number of element, and M, N are more than 0 Integer；
Determining module, it is configured as according to the similarity from big to small, it is determined that the default K similarities, wherein, K For the integer more than 0；
Processing module, it is configured as clustering using between two elements corresponding to each similarity in the K similarity Probability, it is same to calculate confidence value between the first kind and second class and the first kind and second class A kind of analogy probability, wherein, the cluster probability is used for the probability for indicating that two elements are identical element；
Cluster module, it is configured as when the confidence value and the analogy probability are satisfied by preparatory condition, by described in The first kind and second class merge.
Alternatively, described device also includes：
Probability acquisition module, it is configured as obtaining pending element set, calculates any in the pending element set Cluster probability between two elements, the pending element set include：All elements and described second in the first kind All elements in class.
Alternatively, the probability acquisition module, including：
Statistic submodule, it is configured as counting the similarity in the pending element set between any two element；
Determination sub-module, it is configured as, according to the similarity between any two element, determining any two element Between cluster probability.
Alternatively, the processing module, it is configured as using formulaCalculate the first kind and institute State the confidence value D between the second classAB, wherein, piBetween two elements in K similarity of expression corresponding to i-th of similarity Cluster probability, i are more than 0 and are less than or equal to K.
Alternatively, the processing module, it is configured as using formulaCalculate the first kind and described second Class is of a sort analogy probability PAB, wherein, piRepresent to gather between two elements in K similarity corresponding to i-th of similarity Class probability, i are more than 0 and are less than or equal to K.
According to the third aspect of the embodiment of the present disclosure, there is provided a kind of clustering processing device, including：
Processor；
For storing the memory of executable instruction；
Wherein, the processor is configured as：
Each element and each similarity of element between any two in the second class in the first kind are obtained respectively, and it is individual to obtain M*N Similarity, wherein, the first kind includes M element, the second class includes N number of element, and M, N are the integer more than 0；
According to the similarity from big to small, it is determined that presetting the K similarities, wherein, K is the integer more than 0；
Using cluster probability between two elements corresponding to each similarity in the K similarity, described first is calculated Confidence value and the first kind and second class between class and second class are of a sort analogy probability, its In, the cluster probability is used for the probability for indicating that two elements are identical element；
When the confidence value and the analogy probability are satisfied by preparatory condition, by the first kind and second class Merge.
The technical scheme provided by this disclosed embodiment can include the following benefits：Obtain respectively each in the first kind The each similarity of element between any two, obtains M*N similarity in element and the second class, according to above-mentioned similarity from greatly to It is small, it is determined that presetting K similarity, using cluster probability between two elements corresponding to each similarity in K similarity, calculate Confidence value and the first kind and the second class between the above-mentioned first kind and the second class are of a sort analogy probability, Jin Er When above-mentioned confidence value and analogy probability are satisfied by preparatory condition, the first kind and the second class are merged.This new clustering method In, it is contemplated that the cluster probability between different elements, and then inhomogeneous confidence value and analogy probability are calculated, greatly improve The degree of accuracy of cluster.
It should be appreciated that the general description and following detailed description of the above are only exemplary and explanatory, not The disclosure can be limited.
Brief description of the drawings
, below will be to embodiment or existing in order to illustrate more clearly of the embodiment of the present disclosure or technical scheme of the prior art There is the required accompanying drawing used in technology description to be briefly described, it should be apparent that, drawings in the following description are this public affairs Some embodiments opened, for those of ordinary skill in the art, without having to pay creative labor, can be with Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is a kind of flow chart of clustering processing method according to an exemplary embodiment；
Fig. 2 is a kind of flow chart of clustering processing method according to another exemplary embodiment；
Fig. 3 is a kind of structural representation of clustering processing device according to an exemplary embodiment；
Fig. 4 is a kind of structural representation of clustering processing device according to further example embodiment；
Fig. 5 is a kind of structural representation of clustering processing device according to another exemplary embodiment；
Fig. 6 is a kind of structural representation of clustering processing device according to another exemplary embodiment；
Fig. 7 is a kind of structural representation of clustering processing device according to another exemplary embodiment.
Embodiment
Here exemplary embodiment will be illustrated in detail, its example is illustrated in the accompanying drawings.Following description is related to During accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represent same or analogous key element.Following exemplary embodiment Described in embodiment do not represent all embodiments consistent with the disclosure.On the contrary, they be only with it is such as appended The example of the consistent apparatus and method of some aspects be described in detail in claims, the disclosure.
Term " first ", " second ", " the 3rd " in the specification and claims of the disclosure etc. are to be used to distinguish class As object, without for describing specific order or precedence.It should be appreciated that the data so used are in appropriate situation Under can exchange, so as to embodiment of the disclosure described herein for example can with except illustrate or describe herein those with Outer order is implemented.In addition, term " comprising " and " having " and their any deformation, it is intended that cover non-exclusive bag Contain, for example, containing the process of series of steps or unit, method, system, product or equipment is not necessarily limited to what is clearly listed Those steps or unit, but may include not listing clearly or intrinsic for these processes, method, product or equipment Other steps or unit." * " in the specification and claims of the disclosure represents multiplication sign.
In the embodiment of the present disclosure, introduced during clustering processing between element for identical element probability and put Reliability etc., so that the accuracy of cluster is effectively ensured.
Fig. 1 is a kind of flow chart of clustering processing method according to an exemplary embodiment.As shown in figure 1, the party Method includes：
In step S101, between any two similar of each element and each element in the second class in the first kind is obtained respectively Degree, obtains M*N similarity.
The embodiment of the present disclosure illustrates by taking two classes as an example, and during specific implementation, multiple classes can be two-by-two with reference to the first kind Performed with the second class.It should be noted that any two class in multiple classes that the above-mentioned first kind and the second class cluster for needs.Just It is then similar between calculating each two element, it is necessary to which all elements of polymerization can element be each a class under beginning state Degree, specifically can represent similarity with distance, bigger apart from smaller similarity, and the element that distance is less than to predetermined threshold value gathers one Individual class, form multiple classes.
Wherein, the first kind includes M element, the second class includes N number of element, and M, N are the integer more than 0.
Element in the disclosure can refer to the things of any required cluster, such as：Face, fruit etc., this is not restricted.
In step s 102, according to above-mentioned similarity from big to small, it is determined that presetting K similarity.
Wherein, K is the integer more than 0.K is preset value, i.e., is preset according to the sequential selection of above-mentioned similarity from big to small The similarity of number.Specifically, this K Similarity value can be added similarity set E.All similarities in K similarity Corresponding element all adds set V.
Set V and set E can also be constantly updated in whole cluster process, after sequencing of similarity, be opened from big to small Begin, it is assumed that first pair of similarity highest between elements A and element B, determine whether elements A and element B have at least one gathering In V, if so, carrying out down the judgement of a pair of elements；If not provided, just elements A and element B are added in set V.
In step s 103, using cluster probability between two elements corresponding to each similarity in K similarity, calculate Confidence value and the first kind and the second class between the above-mentioned first kind and the second class are of a sort analogy probability.
Wherein, cluster probability is used for the probability for indicating that two elements are identical element.
After selecting K similarity, see these similarities are the similarity between which element respectively.Such as some is similar Degree is the similarity between elements A and element B, then the cluster probability of elements A and element B is obtained, the like, obtain K Cluster probability between two elements corresponding to each similarity in similarity, and then calculate the confidence level between the first kind and the second class Value and the first kind and the second class are of a sort analogy probability.
Alternatively, K can be equal to M, can represent " probable value of the first kind to the second class " and " the second class to the first kind Probable value " is identical.
It should be noted that confidence value is information entropy, for representing that two elements are the confidence levels of identity element. In information source, consideration is not the uncertainty of a certain individual element generation, but to consider that this information source is likely to occur feelings The average uncertainty of condition.The average uncertainty of information source should be the probabilistic assembly average of single symbol, and this is flat Average can be described as information entropy.
In step S104, when above-mentioned confidence value and analogy probability are satisfied by preparatory condition, by the first kind and second Class merges.
In the present embodiment, between any two similar of each element and each element in the second class in the first kind is obtained respectively Degree, obtains M*N similarity, according to above-mentioned similarity from big to small, it is determined that K similarity is preset, using every in K similarity Cluster probability between two elements corresponding to individual similarity, calculate confidence value between the above-mentioned first kind and the second class and The first kind and the second class are of a sort analogy probability, and then are satisfied by preparatory condition in above-mentioned confidence value and analogy probability When, the first kind and the second class are merged.In this new clustering method, it is contemplated that the cluster probability between different elements, and then Inhomogeneous confidence value and analogy probability are calculated, substantially increases the degree of accuracy of cluster.
Cluster probability in above-mentioned pending element set between any two element can obtain from presetting database, Acquisition can be calculated temporarily.
Alternatively, using cluster probability between two elements corresponding to each similarity in K similarity, above-mentioned the is calculated Before confidence value and the first kind and the second class between a kind of and the second class are of a sort analogy probability, it can gather Substantial amounts of element samples, the cluster probability between any two element is calculated, and be stored in presetting database.
Specifically, can include：Obtain pending element set, calculate in pending element set any two element it Between cluster probability.It during specific calculating, can two-by-two be combined, the cluster probability between two elements be calculated, until pending In element set, all elements all calculated cluster probability between any two.
Above-mentioned pending element set, can include substantial amounts of element samples, these element samples include identical element, Different elements.Wherein, pending element set includes all elements in all elements, the second class in the first kind, but does not limit to In this.
Fig. 2 is a kind of flow chart of clustering processing method according to another exemplary embodiment.
Alternatively, as shown in Fig. 2 calculating in pending element set cluster probability between any two element, can wrap Include：
In step s 201, the similarity between any two element in pending element set is counted.
In step S202, according to the similarity between any two element, the cluster between any two element is determined Probability.
Specifically, when combination of two calculates the similarity between two elements, any similarity algorithm, this public affairs can be used It is not limited in opening.For example, if element is all face, face recognition algorithms can be used, extract the feature of each face, then Calculate the similarity between two faces.
And then by the statistics to similarity, calculate the probability that two elements are identity elements.Such as waveform can be used Statistics, that is, generate similarity curve, similarity is higher, and two elements are that the probability of identity element is bigger, i.e. cluster probability value is got over It is high.
The value of similarity can be embodied by distance value in statistic processes, and the distance of two elements is nearer, identify similarity It is higher.It can be seen that distance is nearer on curve map, cluster probability is higher.
Specifically, using cluster probability between two elements corresponding to each similarity in K similarity, the first kind is calculated And the second confidence value between class, can be specifically：Using formulaCalculate the first kind and described Confidence value D between second classAB。
Wherein, piRepresent cluster probability between two elements in K similarity corresponding to i-th of similarity, i more than 0 and Less than or equal to K.
-log2piIt is the uncertainty of identical element to represent two corresponding to i-th of similarity elements.
DABThe element in element and the second class in the smaller expression first kind is the uncertain smaller of identical element.
Similarly, using cluster probability between two elements corresponding to each similarity in K similarity, the first kind is calculated And the second analogy probability between class, Ke Yishi：Using formulaIt is same to calculate the first kind and second class A kind of analogy probability PAB, wherein, piCluster probability between two elements in K similarity of expression corresponding to i-th of similarity, I is more than 0 and is less than or equal to K.
On the basis of above-described embodiment, the threshold value of confidence level and the threshold value of analogy probability can be set respectively.In confidence When angle value and analogy probability are satisfied by preparatory condition, the first kind and second class are merged, Ke Yishi：In confidence value When being more than the second predetermined threshold value more than the first predetermined threshold value and analogy probability, the above-mentioned first kind and the second class are merged.
Fig. 3 is a kind of structural representation of clustering processing device according to an exemplary embodiment.The disclosure is implemented Example provides a kind of clustering processing device, can be integrated in terminal, or terminal.Here terminal can refer to computer, clothes The equipment such as business device, this is not restricted.As shown in figure 3, the device includes：Acquisition module 301, determining module 302, processing module 303 and cluster module 304, wherein：
Acquisition module 301, be configured to obtain in the first kind each element with the second class each element two-by-two it Between similarity, obtain M*N similarity, wherein, the first kind, which includes M element, the second class, includes N number of element, and M, N are big In 0 integer.
Determining module 302, it is configured as according to the similarity from big to small, it is determined that the default K similarities, its In, K is the integer more than 0.
Processing module 303, it is configured as gathering using between two elements corresponding to each similarity in the K similarity Class probability, calculates the confidence value between the first kind and second class and the first kind and second class is Of a sort analogy probability, wherein, the cluster probability is used for the probability for indicating that two elements are identical element.
Cluster module 304, it is configured as when the confidence value and the analogy probability are satisfied by preparatory condition, by institute State the first kind and second class merges.
In the clustering processing device that the present embodiment provides, obtain respectively in the first kind each element with it is each first in the second class The similarity of element between any two, obtains M*N similarity, according to above-mentioned similarity from big to small, it is determined that K similarity is preset, Using cluster probability between two elements corresponding to each similarity in K similarity, calculate the above-mentioned first kind and the second class it Between confidence value and the first kind and the second class be of a sort analogy probability, it is and then general in above-mentioned confidence value and analogy When rate is satisfied by preparatory condition, the first kind and the second class are merged.In this new clustering method, it is contemplated that between different elements Cluster probability, and then calculate inhomogeneous confidence value and analogy probability, substantially increase the degree of accuracy of cluster.
Fig. 4 is a kind of structural representation of clustering processing device according to further example embodiment.Such as Fig. 4 institutes Show, on the basis of Fig. 3, the device can also include：Probability acquisition module 401.
Probability acquisition module 401, it is configured as obtaining pending element set, calculates in the pending element set and appoint Cluster probability between two elements of meaning, the pending element set include：All elements and described in the first kind All elements in two classes.
Fig. 5 is a kind of structural representation of clustering processing device according to another exemplary embodiment.Alternatively, such as Shown in Fig. 5, on the basis of Fig. 4, probability acquisition module 401, it can include：Statistic submodule 501 and determination sub-module 502, Wherein：
Statistic submodule 501, it is configured as counting similar between any two element in the pending element set Degree.
Determination sub-module 502, it is configured as according to the similarity between any two element, determines any two member Cluster probability between element.
Further, processing module 303, can be specifically configured to use formulaDescribed in calculating Confidence value D between the first kind and second classAB, wherein, piRepresent in K similarity corresponding to i-th of similarity Cluster probability between two elements, i are more than 0 and are less than or equal to K.
Alternatively, processing module 303, can be specifically configured to use formulaCalculate the first kind and Second class is of a sort analogy probability PAB, wherein, piRepresent two in K similarity corresponding to i-th of similarity Cluster probability between element, i are more than 0 and are less than or equal to K.
Fig. 6 is a kind of structural representation of clustering processing device according to another exemplary embodiment.At the cluster Device is managed, can be integrated in terminal, or terminal.Here terminal can refer to the equipment such as computer, server, herein not It is restricted.
As shown in fig. 6, the device includes：Processor 601 and the memory 602 for storing executable instruction.Wherein, locate Reason device 601 and memory 602 couple.
Processor 601 is configured as：
Each element and each similarity of element between any two in the second class in the first kind are obtained respectively, and it is individual to obtain M*N Similarity, wherein, the first kind includes M element, the second class includes N number of element, and M, N are the integer more than 0；
According to the similarity from big to small, it is determined that presetting the K similarities, wherein, K is the integer more than 0；
Using cluster probability between two elements corresponding to each similarity in the K similarity, described first is calculated Confidence value and the first kind and second class between class and second class are of a sort analogy probability, its In, the cluster probability is used for the probability for indicating that two elements are identical element；
When the confidence value and the analogy probability are satisfied by preparatory condition, by the first kind and second class Merge.
In summary, in the clustering processing device that the present embodiment provides, each element and second in the first kind is obtained respectively The each similarity of element between any two, obtains M*N similarity in class, according to above-mentioned similarity from big to small, it is determined that default K Individual similarity, using cluster probability between two elements corresponding to each similarity in K similarity, calculate the above-mentioned first kind and Confidence value and the first kind and the second class between second class are of a sort analogy probability, and then in above-mentioned confidence value When being satisfied by preparatory condition with analogy probability, the first kind and the second class are merged.In this new clustering method, it is contemplated that different Cluster probability between element, and then inhomogeneous confidence value and analogy probability are calculated, substantially increase the accurate of cluster Degree.
Fig. 7 is a kind of structural representation of clustering processing device according to another exemplary embodiment.
Reference picture 7, clustering processing device 700 can include following one or more assemblies：Processing component 702, memory 704, electric power assembly 706, multimedia groupware 708, audio-frequency assembly 710, input/output (input/output, referred to as：I/O) connect Mouth 712, sensor cluster 714, and communication component 716.
Processing component 702 generally controls the integrated operation of clustering processing device 700, is such as communicated with display, data, camera The operation that operation and record operation are associated.Processing component 702 can carry out execute instruction including one or more processors 720, To complete all or part of step of above-mentioned method.In addition, processing component 702 can include one or more modules, it is easy to Interaction between processing component 702 and other assemblies.For example, processing component 702 can include multi-media module, to facilitate more matchmakers Interaction between body component 708 and processing component 702.
Memory 704 is configured as storing various types of data to support the operation in clustering processing device 700.These The example of data includes the instruction of any application program or method for being operated on clustering processing device 700, contacts number According to, telephone book data, message, picture, video etc..Memory 704 can be by any kind of volatibility or non-volatile memories Equipment or combinations thereof are realized, such as static RAM (Static Random Access Memory, abbreviation： SRAM), Electrically Erasable Read Only Memory (Electrically Erasable Programmable Read-Only Memory, referred to as：EEPROM), Erasable Programmable Read Only Memory EPROM (Erasable Programmable Read Only Memory, referred to as：EPROM), programmable read only memory (Programmable Red-Only Memory, abbreviation：PROM), Read-only storage (Read-Only Memory, referred to as：ROM), magnetic memory, flash memory, disk or CD.
Electric power assembly 706 provides electric power for the various assemblies of clustering processing device 700.Electric power assembly 706 can include electricity Management system, one or more power supplys, and other generate with for clustering processing device 700, manage and to distribute electric power associated Component.
Multimedia groupware 708 is included in one output interface of offer between the clustering processing device 700 and user Screen.In certain embodiments, screen can include liquid crystal display (Liquid Crystal Display, abbreviation：LCD) and Touch panel (Touch Panel, referred to as：TP).If screen includes touch panel, screen may be implemented as touch-screen, with Receive the input signal from user.Touch panel includes one or more touch sensors with sensing touch, slip and touch Gesture on panel.The touch sensor can the not only border of sensing touch or sliding action, but also detect with it is described Touch or the duration and pressure of slide correlation.In certain embodiments, multimedia groupware 708 includes one and preposition taken the photograph As head and/or rear camera.It is preposition during such as screening-mode or video mode when clustering processing device 700 is in operator scheme Camera and/or rear camera can receive the multi-medium data of outside.Each front camera and rear camera can be with It is a fixed optical lens system or there is focusing and optical zoom capabilities.
Audio-frequency assembly 710 is configured as output and/or input audio signal.For example, audio-frequency assembly 710 includes a Mike Wind (Microphone, referred to as：MIC), when clustering processing device 700 is in operator scheme, such as call model, logging mode and language During sound recognition mode, microphone is configured as receiving external audio signal.The audio signal received can be further stored Sent in memory 704 or via communication component 716.In certain embodiments, audio-frequency assembly 710 also includes a loudspeaker, For exports audio signal.
I/O interfaces 712 provide interface between processing component 702 and peripheral interface module, and above-mentioned peripheral interface module can To be keyboard, click wheel, button etc..These buttons may include but be not limited to：Home button, volume button, start button and lock Determine button.
Sensor cluster 714 includes one or more sensors, for providing various aspects for clustering processing device 700 State estimation.For example, sensor cluster 714 can detect opening/closed mode of clustering processing device 700, the phase of component To positioning, such as the display and keypad that the component is clustering processing device 700, sensor cluster 714 can also detect The position of 700 1 components of clustering processing device 700 or clustering processing device changes, and user contacts with clustering processing device 700 Existence or non-existence, the orientation of clustering processing device 700 or acceleration/deceleration and the temperature change of clustering processing device 700.Sensing Device assembly 714 can include proximity transducer, be configured to detect depositing for neighbouring object in no any physical contact .Sensor cluster 714 can also include optical sensor, such as complementary metal oxide semiconductor (Complementary Metal Oxide Semiconductor, referred to as：CMOS) or charge coupled cell (Charge-coupled Device, referred to as：CCD) Photosensitive imaging element, for being used in imaging applications.In certain embodiments, the sensor cluster 714 can also include adding Velocity sensor, gyro sensor, Magnetic Sensor, pressure sensor or temperature sensor.
Communication component 716 is configured to facilitate wired or wireless way between clustering processing device 700 and other equipment Communication.Clustering processing device 700 can access the wireless network based on communication standard, such as Wi-Fi, 2G or 3G, or their group Close.In one exemplary embodiment, communication component 716 receives the broadcast from external broadcasting management system via broadcast channel Signal or broadcast related information.In one exemplary embodiment, the communication component 716 also includes near-field communication (Near Field Communication, referred to as：NFC) module, to promote junction service.For example, radio frequency identification can be based in NFC module (Radio Frequency Identification, referred to as：RFID) technology, Infrared Data Association (Infrared Data Association, referred to as：IrDA) technology, and ultra wide band (Ultra Wideband, referred to as：UWB) technology, bluetooth (Bluetooth, referred to as：BT) technology and other technologies are realized.
In the exemplary embodiment, clustering processing device 700 can be by one or more application specific integrated circuits (Application Specific Integrated Circuit, referred to as：ASIC), digital signal processor (Digital Signal Processor, referred to as：DSP), digital signal processing appts (Digital Signal Processing Device, Referred to as：DSPD), PLD (Programmable Logic Device, abbreviation：PLD), field-programmable gate array Row (Field Programmable Gate Array, referred to as：FPGA), controller, microcontroller, microprocessor or other electricity Subcomponent is realized, for performing the above method.
In the exemplary embodiment, a kind of non-transitorycomputer readable storage medium including instructing, example are additionally provided Such as include the memory 704 of instruction, above-mentioned instruction can be performed by the processor 720 of clustering processing device 700 to complete above-mentioned side Method.For example, the non-transitorycomputer readable storage medium can be ROM, random access memory (Random Access Memory, referred to as：RAM), read-only optical disc (Compact Disc Read-Only Memory, abbreviation：CD-ROM), tape, soft Disk and optical data storage devices etc..
A kind of non-transitorycomputer readable storage medium, when the instruction in the storage medium is by clustering processing device During 700 computing device so that clustering processing device 700 is able to carry out the above method.
Those skilled in the art will readily occur to the disclosure its after considering specification and putting into practice invention disclosed herein Its embodiment.The application is intended to any modification, purposes or the adaptations of the disclosure, these modifications, purposes or Person's adaptations follow the general principle of the disclosure and including the undocumented common knowledges in the art of the disclosure Or conventional techniques.Description and embodiments are considered only as exemplary, and the true scope of the disclosure and spirit are by following Claims are pointed out.
It should be appreciated that the precision architecture that the disclosure is not limited to be described above and is shown in the drawings, and And various modifications and changes can be being carried out without departing from the scope.The scope of the present disclosure is only limited by appended claims System.

Claims (11)

  1. A kind of 1. clustering processing method, it is characterised in that including：
    Each element and each similarity of element between any two in the second class in the first kind are obtained respectively, and it is individual similar to obtain M*N Degree, wherein, the first kind includes M element, the second class includes N number of element, and M, N are the integer more than 0；
    According to the similarity from big to small, it is determined that presetting the K similarities, wherein, K is the integer more than 0；
    Using cluster probability between two elements corresponding to each similarity in the K similarity, calculate the first kind and Confidence value and the first kind and second class between second class are of a sort analogy probability, wherein, institute State cluster probability be used for indicate two elements for identical element probability；
    When the confidence value and the analogy probability are satisfied by preparatory condition, the first kind and second class are closed And.
  2. 2. according to the method for claim 1, it is characterised in that described using each similarity institute in the K similarity Cluster probability between corresponding two elements, calculate the confidence value between the first kind and second class and described Before a kind of and described second class is of a sort analogy probability, in addition to：
    Pending element set is obtained, calculates the cluster probability between any two element, institute in the pending element set Stating pending element set includes：All elements in all elements and second class in the first kind.
  3. 3. according to the method for claim 2, it is characterised in that described to calculate any two in the pending element set Cluster probability between element, including：
    Count the similarity between any two element in the pending element set；
    According to the similarity between any two element, the cluster probability between any two element is determined.
  4. 4. according to the method in claim 2 or 3, it is characterised in that described using each similarity in the K similarity Cluster probability between two corresponding elements, the confidence value between the first kind and second class is calculated, including：
    Using formulaCalculate the confidence value D between the first kind and second classAB, wherein, pi Cluster probability between two elements in K similarity of expression corresponding to i-th of similarity, i are more than 0 and are less than or equal to K.
  5. 5. according to the method in claim 2 or 3, it is characterised in that described using each similarity in the K similarity Cluster probability between two corresponding elements, it is of a sort analogy probability to calculate the first kind and second class, including：
    Using formulaIt is of a sort analogy probability P to calculate the first kind and second classAB, wherein, piTable Show cluster probability between two elements in K similarity corresponding to i-th of similarity, i is more than 0 and is less than or equal to K.
  6. A kind of 6. clustering processing device, it is characterised in that including：
    Acquisition module, it is configured to obtain between any two similar of each element and each element in the second class in the first kind Degree, obtains M*N similarity, wherein, the first kind includes M element, the second class includes N number of element, and M, N are whole more than 0 Number；
    Determining module, it is configured as according to the similarity from big to small, it is determined that the default K similarities, wherein, K is big In 0 integer；
    Processing module, it is configured as using cluster probability between two elements in the K similarity corresponding to each similarity, It is of a sort to calculate the confidence value between the first kind and second class and the first kind and second class Analogy probability, wherein, the cluster probability is used for the probability for indicating that two elements are identical element；
    Cluster module, it is configured as when the confidence value and the analogy probability are satisfied by preparatory condition, by described first Class and second class merge.
  7. 7. device according to claim 6, it is characterised in that also include：
    Probability acquisition module, it is configured as obtaining pending element set, calculates any two in the pending element set Cluster probability between element, the pending element set include：In the first kind in all elements and second class All elements.
  8. 8. device according to claim 7, it is characterised in that the probability acquisition module, including：
    Statistic submodule, it is configured as counting the similarity in the pending element set between any two element；
    Determination sub-module, it is configured as according to the similarity between any two element, determines between any two element Cluster probability.
  9. 9. the device according to claim 7 or 8, it is characterised in that the processing module, be configured as using formulaCalculate the confidence value D between the first kind and second classAB, wherein, piRepresent that K is individual similar Cluster probability between two elements in degree corresponding to i-th of similarity, i are more than 0 and are less than or equal to K.
  10. 10. the device according to claim 6 or 8, it is characterised in that the processing module, be configured as using formulaIt is of a sort analogy probability P to calculate the first kind and second classAB, wherein, piRepresent K similarity In cluster probability between two elements corresponding to i-th of similarity, i is more than 0 and is less than or equal to K.
  11. A kind of 11. clustering processing device, it is characterised in that including：
    Processor；
    For storing the memory of executable instruction；
    Wherein, the processor is configured as：
    Each element and each similarity of element between any two in the second class in the first kind are obtained respectively, and it is individual similar to obtain M*N Degree, wherein, the first kind includes M element, the second class includes N number of element, and M, N are the integer more than 0；
    According to the similarity from big to small, it is determined that presetting the K similarities, wherein, K is the integer more than 0；
    Using cluster probability between two elements corresponding to each similarity in the K similarity, calculate the first kind and Confidence value and the first kind and second class between second class are of a sort analogy probability, wherein, institute State cluster probability be used for indicate two elements for identical element probability；
    When the confidence value and the analogy probability are satisfied by preparatory condition, the first kind and second class are closed And.
CN201710573089.4A 2017-07-14 2017-07-14 Clustering processing method and device Active CN107423757B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710573089.4A CN107423757B (en) 2017-07-14 2017-07-14 Clustering processing method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710573089.4A CN107423757B (en) 2017-07-14 2017-07-14 Clustering processing method and device

Publications (2)

Publication Number Publication Date
CN107423757A true CN107423757A (en) 2017-12-01
CN107423757B CN107423757B (en) 2020-10-09

Family

ID=60427642

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710573089.4A Active CN107423757B (en) 2017-07-14 2017-07-14 Clustering processing method and device

Country Status (1)

Country Link
CN (1) CN107423757B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110377670A (en) * 2018-04-11 2019-10-25 腾讯大地通途（北京）科技有限公司 A kind of method, apparatus, medium and the equipment of determining road element information
CN110403582A (en) * 2019-07-23 2019-11-05 宏人仁医医疗器械设备(东莞)有限公司 A method of for analyzing pulse waveform quality
CN111444933A (en) * 2019-11-26 2020-07-24 北京邮电大学 Object classification method and device

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070086627A1 (en) * 2005-10-18 2007-04-19 Samsung Electronics Co., Ltd. Face identification apparatus, medium, and method
TW201118754A (en) * 2009-11-17 2011-06-01 Nat Cheng Kong University Face clustering system using PLSA model
US20120328167A1 (en) * 2011-06-23 2012-12-27 Peng Wu Merging face clusters
CN103324949A (en) * 2012-03-21 2013-09-25 阿里巴巴集团控股有限公司 Method and device for recognizing object in image
CN105608430A (en) * 2015-12-22 2016-05-25 小米科技有限责任公司 Face clustering method and device
CN105844283A (en) * 2015-01-16 2016-08-10 阿里巴巴集团控股有限公司 Method for identifying category of image, image search method and image search device
CN106778501A (en) * 2016-11-21 2017-05-31 武汉科技大学 Video human face ONLINE RECOGNITION method based on compression tracking with IHDR incremental learnings

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070086627A1 (en) * 2005-10-18 2007-04-19 Samsung Electronics Co., Ltd. Face identification apparatus, medium, and method
TW201118754A (en) * 2009-11-17 2011-06-01 Nat Cheng Kong University Face clustering system using PLSA model
US20120328167A1 (en) * 2011-06-23 2012-12-27 Peng Wu Merging face clusters
CN103324949A (en) * 2012-03-21 2013-09-25 阿里巴巴集团控股有限公司 Method and device for recognizing object in image
CN105844283A (en) * 2015-01-16 2016-08-10 阿里巴巴集团控股有限公司 Method for identifying category of image, image search method and image search device
CN105608430A (en) * 2015-12-22 2016-05-25 小米科技有限责任公司 Face clustering method and device
CN106778501A (en) * 2016-11-21 2017-05-31 武汉科技大学 Video human face ONLINE RECOGNITION method based on compression tracking with IHDR incremental learnings

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
ALBERT CLAPES: "Multi-modal user identification and object recognition surveillance system", 《PATTERN RECOGNITION LETTERS》 *
王蕾: "人脸识别系统与人脸检测算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 *
田丰: "采用类别相似度聚合的关联文本分类方法", 《西安交通大学学报》 *
贾海龙: "视频人脸识别中基于聚类中心LLE的特征相似性融合方法", 《科学技术与工程》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110377670A (en) * 2018-04-11 2019-10-25 腾讯大地通途（北京）科技有限公司 A kind of method, apparatus, medium and the equipment of determining road element information
CN110377670B (en) * 2018-04-11 2021-11-26 腾讯大地通途（北京）科技有限公司 Method, device, medium and equipment for determining road element information
CN110403582A (en) * 2019-07-23 2019-11-05 宏人仁医医疗器械设备(东莞)有限公司 A method of for analyzing pulse waveform quality
CN110403582B (en) * 2019-07-23 2021-12-03 宏人仁医医疗器械设备(东莞)有限公司 Method for analyzing pulse wave form quality
CN111444933A (en) * 2019-11-26 2020-07-24 北京邮电大学 Object classification method and device
CN111444933B (en) * 2019-11-26 2023-10-10 北京邮电大学 Object classification method and device

Also Published As

Publication number Publication date
CN107423757B (en) 2020-10-09

Similar Documents

Publication Publication Date Title
CN104156947B (en) Image partition method, device and equipment
CN106372204A (en) Push message processing method and device
CN110188236A (en) A kind of recommended method of music, apparatus and system
CN108304475A (en) Data query method, apparatus and electronic equipment
CN105260732A (en) Image processing method and device
CN104809157B (en) Number identification method and device
TW202109360A (en) Image processing method and device, electronic equipment and storage medium
CN104834559A (en) Notification message display method and device
CN105426878A (en) Method and device for face clustering
CN105205494A (en) Similar picture identification method and device
TWI779449B (en) Object counting method electronic equipment computer readable storage medium
CN105354560A (en) Fingerprint identification method and device
CN107423757A (en) clustering processing method and device
CN111259967A (en) Image classification and neural network training method, device, equipment and storage medium
CN106534951A (en) Method and apparatus for video segmentation
CN105678266A (en) Method and device for combining photo albums of human faces
CN105335714A (en) Photograph processing method, device and apparatus
CN105208284A (en) Photographing reminding method and device
CN105098882B (en) input current distribution method and device
CN106469041A (en) The method and device of PUSH message, terminal unit
CN105608430A (en) Face clustering method and device
CN105205153A (en) File storage method and device
CN107402767A (en) The method and apparatus for showing PUSH message
CN104268149A (en) Clustering method and clustering device
CN105551047A (en) Picture content detecting method and device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant