CN104680046A - User activity recognition method and device - Google Patents
User activity recognition method and device Download PDFInfo
- Publication number
- CN104680046A CN104680046A CN201310629291.6A CN201310629291A CN104680046A CN 104680046 A CN104680046 A CN 104680046A CN 201310629291 A CN201310629291 A CN 201310629291A CN 104680046 A CN104680046 A CN 104680046A
- Authority
- CN
- China
- Prior art keywords
- recognition result
- activity
- user
- activity recognition
- user activity
- 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
Abstract
The invention discloses a user activity recognition method and a user activity recognition device. The user activity recognition method is used for an intelligent terminal with a sensor, and includes: obtaining first data of the sensor; calculating the first data according to a first algorithm so as to obtain a first activity recognition result; using an amending model built according to training data to amend the first activity recognition result, and obtaining a second activity recognition result, wherein the training data includes at least one group of user relevance information and history data, the history data comprises a user activity recognition result and a user activity of a user label corresponding to the user activity recognition result, and the user relevance information at least includes one item of obtaining time, an obtaining position, user personal information and application use state information of sensor data used to confirm the user activity recognition result; confirming the second activity recognition result as the user activity. The user activity recognition method improves recognition accuracy for the user activity.
Description
Technical field
The present invention relates to technical field of data processing, particularly relate to a kind of User Activity recognition methods and device.
Background technology
Along with the development of mobile communication, the sensor device that several functions is powerful is integrated with in present intelligent mobile terminal, as acceleration (Accelerometer), surround lighting (Ambient Light), GPS (Global Positioning System, GPS), proximity sensor (Proximity Sensor), compass (compass), gyroscope (gyroscope), camera (Camera) etc., intelligent terminal can catch the small change of each sensor in real time, and makes corresponding reaction.
These sensors in a mobile communication device be generally applied as the application chance that data mining provides huge.Sensing data in mobile terminal analyzed and excavates to identify that the action of user and activity are the subjects of current hot topic, having great value in science with commercial.Such as, utilize smart mobile phone to be in children and the activity of school detects, detect the exception of children ' s activity early, or to the head of a family and teacher, early warning etc. is carried out to the danger that may face etc.
In prior art, intelligent terminal can identify the activity of active user according to the sensing data of current acquisition and the knowledge base set up in advance, comprise the training data of sensing data and the corresponding User Activity formerly collected in this knowledge base.But the method is completely based on sensing data and limited training data, lower to the recognition accuracy of User Activity.
Summary of the invention
Provide a kind of User Activity recognition methods and device in the embodiment of the present invention, the recognition accuracy of User Activity can be improved.
In order to solve the problems of the technologies described above, the embodiment of the invention discloses following technical scheme:
First aspect, provide a kind of User Activity recognition methods, be applied to the intelligent terminal being provided with sensor, described method comprises:
Obtain the first data of described sensor;
According to the first algorithm, calculating acquisition first activity recognition result is carried out to described first data;
The correction model set up according to training data is adopted to revise described first activity recognition result, obtain the second activity recognition result, described training data comprises at least one group of user related information and historical data, described historical data comprises the User Activity of the user annotation of User Activity recognition result and correspondence thereof, and described user related information at least comprises one in acquisition time of the sensing data for determining described User Activity recognition result and place, userspersonal information, application using state information;
Described second activity recognition result is defined as User Activity.
In conjunction with above-mentioned first aspect, in the first implementation, described first activity recognition result is the vector of a n kind activity probability;
The correction model that described employing is set up according to training data is revised described first activity recognition result, obtains the second activity recognition result, comprising:
Described correction model is expressed as binary set;
Calculate the described vector of n kind activity probability and the product of described binary set, obtain probability component;
In the described probability component obtained, the activity corresponding to probability component selecting numerical value maximum is as described second activity recognition result.
In conjunction with above-mentioned first aspect, or the first implementation of first aspect, in the second implementation, described described second activity recognition result is defined as User Activity before, also comprise:
Judge whether the probability component of described second activity recognition result is more than or equal to threshold value;
Described described second activity recognition result is defined as User Activity, comprises:
When the probability component of described second activity recognition result is more than or equal to described threshold value, then described second activity recognition result is defined as User Activity.
In conjunction with above-mentioned first aspect, or the first implementation of first aspect, or the second implementation of first aspect, in the third implementation, also comprise:
When the probability component of described second activity recognition result is less than described threshold value, export described second activity recognition result to user;
Receive the feedback information to described second activity recognition result of described user input;
According to described feedback information determination User Activity.
In conjunction with above-mentioned first aspect, or the first implementation of first aspect, or the second implementation of first aspect, or the third implementation of first aspect, in the 4th kind of implementation, also comprise:
When comprising the User Activity of user annotation in described feedback information, the User Activity of the user annotation in described first activity recognition result and described feedback information is increased in described training data as one group of historical data, and according to adding correction model described in the training data correction after historical data.
In conjunction with above-mentioned first aspect, or the first implementation of first aspect, or the second implementation of first aspect, or the third implementation of first aspect, or the 4th of first aspect the kind of implementation, in the 5th kind of implementation, described described second activity recognition result is defined as User Activity after, also comprise:
According to described User Activity, in knowledge base, search the application program corresponding with described User Activity, wherein, in described knowledge base, store the corresponding relation of multiple User Activity and application program;
Start described corresponding with described User Activity application program.
Second aspect, provides a kind of User Activity recognition device, and this application of installation is in the intelligent terminal being provided with sensor, and described device comprises:
Acquiring unit, for obtaining the first data of described sensor;
Recognition unit, carries out calculating acquisition first activity recognition result for the first data obtained described acquiring unit according to the first algorithm;
Amending unit, for adopting the correction model set up according to training data, the first activity recognition result that described recognition unit identifies is revised, obtain the second activity recognition result, described training data comprises at least one group of user related information and historical data, described historical data comprises the User Activity of the user annotation of User Activity recognition result and correspondence thereof, and described user related information at least comprises one in acquisition time of the sensing data for determining described User Activity recognition result and place, userspersonal information, application using state information;
Determining unit, is defined as User Activity for the second activity recognition result described amending unit correction obtained.
In conjunction with above-mentioned second aspect, in the first implementation, described first activity recognition result is the vector of a n kind activity probability;
Described amending unit comprises:
Characterize subelement, for described correction model is expressed as binary set;
Computation subunit, the product of the binary set that vector and described sign subelement for calculating described n kind activity probability characterize, obtains probability component;
Chooser unit, in the described probability component that obtains in described computation subunit, the activity corresponding to probability component selecting numerical value maximum is as described second activity recognition result.
In conjunction with above-mentioned second aspect, or the first implementation of second aspect, in the second implementation, also comprise:
First judging unit, for judging whether the probability component of described second activity recognition result is more than or equal to threshold value;
Described determining unit, during specifically for judging that the probability component of described second activity recognition result is more than or equal to described threshold value when described first judging unit, is defined as User Activity by described second activity recognition result.
In conjunction with above-mentioned second aspect, or the first implementation of second aspect, or the second implementation of second aspect, in the third implementation, also comprise:
Output unit, during for judging that the probability component of described second activity recognition result is less than described threshold value when described first judging unit, exports described second activity recognition result to user;
Information receiving unit, for receiving the feedback information to the described second activity recognition result that described output unit exports of described user input;
Described determining unit, the feedback information determination User Activity also for receiving according to described information receiving unit.
In conjunction with above-mentioned second aspect, or the first implementation of second aspect, or the second implementation of second aspect, or the third implementation of second aspect, in the 4th kind of implementation, also comprise:
Second judging unit, for judging the User Activity whether comprising user annotation in the feedback information that described information receiving unit receives;
Data adding device, for when described second judging unit judges the User Activity comprising user annotation in described feedback information, the User Activity of described first activity recognition result and described user annotation is increased in described training data as one group of historical data, and according to adding correction model described in the training data correction after historical data.
In conjunction with above-mentioned second aspect, or the first implementation of second aspect, or the second implementation of second aspect, or the third implementation of second aspect, or the 4th of second aspect the kind of implementation, in the 5th kind of implementation, also comprise:
Search unit, for according to the determined User Activity of described determining unit, in knowledge base, search the application program corresponding with described User Activity, wherein, in described knowledge base, store the corresponding relation of User Activity and application program;
Start unit, for searching the application program corresponding with described User Activity that unit finds described in starting.
The embodiment of the present invention sets up correction model by the relevant information and historical data adopting user, revising, can improve the accuracy rate to User Activity identification to calculating according to sensing data the activity recognition result obtained.
Accompanying drawing explanation
In order to be illustrated more clearly in the technical scheme of the embodiment of the present invention, be briefly described to the accompanying drawing used required in embodiment or description of the prior art below, apparently, for those of ordinary skills, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is a kind of User Activity recognition methods of embodiment of the present invention process flow diagram;
Fig. 2 is to the method flow diagram that the first activity recognition result is revised in the embodiment of the present invention;
Fig. 3 is the method flow diagram that embodiment of the present invention another kind identifies User Activity;
Fig. 4 is a kind of structural representation identifying User Activity device of the embodiment of the present invention;
Fig. 5 is the structural representation of a kind of amending unit in the embodiment of the present invention;
Fig. 6 is the structural representation of the another kind of User Activity recognition device of the embodiment of the present invention;
Fig. 7 is the structural representation of the another kind of User Activity recognition device of the embodiment of the present invention.
Embodiment
Technical scheme in the embodiment of the present invention is understood better in order to make those skilled in the art person, and enable the above-mentioned purpose of the embodiment of the present invention, feature and advantage become apparent more, below in conjunction with accompanying drawing, technical scheme in the embodiment of the present invention is described in further detail.
See Fig. 1, it is a kind of User Activity recognition methods of embodiment of the present invention process flow diagram.
The present embodiment method is applied to the intelligent terminal being provided with at least one sensor, and this User Activity recognition methods can comprise:
Step 101, obtains the first data of sensor.
First User Activity recognition device obtains the first data that sensor gathers.Wherein, sensor can be Gravity accelerometer, linear acceleration sensors, gyroscope, one or more in the sensors such as range sensor, the first data gathered can be speed and variable quantity thereof, sound, light, Wireless Fidelity (Wireless-Fidelity, Wifi) data such as device identification (Identity, ID) and signal intensity.
Step 102, carries out calculating acquisition first activity recognition result to the first data according to the first algorithm.
For the ease of calculating, can be first the attribute for calculating needed for the first activity recognition result by this first data processing.Then adopt the model and algorithm set up in advance to calculate the first data, obtain the first activity recognition result, this algorithm can be Active Learning Algorithm, also can be other existing algorithms, will not enumerate herein.This first activity recognition result can be expressed as the vector of a n kind activity probability, as (p1, p2 ..., pn), wherein, pj is the probability of movable j, j=1 ... n.
Wherein, being the attribute for calculating needed for the first activity recognition result by this first data processing, such as, gathering the Gravity accelerometer data on intelligent terminal, obtaining the serial number sequence of an acceleration transducer on x, y, z tri-directions.The data window of t time width is got forward from current time T.Thus obtain (T-t, T) sensing data in the time, to x in this time window, y, z tri-directions, and the data vector on three directions and the data on direction calculating mean value respectively, standard deviation, maximal value, minimum value, and the amplitude of all directions component after discrete Fourier transformation.Using the value after above-mentioned calculating as (T-t, T) during this period of time in the attributive character of acceleration transducer.In addition, property value can also by the gps data in (T-t, T), sound, light, and the service condition of application does corresponding calculating and conversion obtains corresponding property value.Especially, can to the POI (attribute in place) in place corresponding to gps data as attribute for gps data, the site identifications that also GPS of user can be carried out giving after cluster is as attribute.
In the embodiment of the present invention, the algorithm obtaining the first activity recognition result can be any effective sorting algorithm.For Active Learning sorting algorithm, the advantage of Active Learning Algorithm is when having a large amount of unlabeled data, algorithm can be picked out has the data of lifting to allow user or expert come manually to mark to classifying quality, thus obtains main labeled data to strengthen the classifying quality of algorithm.Simultaneously because algorithm can be selected unlabeled data, the workload of data mark can be reduced so widely.By the mark allowing user feedback come general data, thus the accuracy of Optimum Classification can be carried out by minimum cost in the present invention.
Can adopt the algorithm of support vector machine for the unlabeled data being difficult to classify, this algorithm can calculate the support vector of classification, thus determines a best classification plane.For unlabeled data, calculate the distance W that each unlabeled data separates class plane.Separate the impact of the nearer data of class plane on classifying quality larger, therefore unlabeled data W being less than certain value e is picked out, user is allowed to carry out feeding back the mark obtaining this partial data, then to the training that model carries out again, thus make classification plane more accurate, thus the effect of classification is got a promotion.Algorithm of support vector machine is traditional algorithm, and those skilled in the relevant art should understand.
Step 103, adopts the correction model set up according to training data to revise the first activity recognition result, obtains the second activity recognition result.
The embodiment of the present invention and the most important difference of prior art are, are not that the activity choosing maximum probability in the first activity recognition result is defined as final User Activity, but revise the first activity recognition result further.
Before revising the first activity recognition result, first setting up correction model, set up the process of correction model, can be first gather training data.This training data comprises at least one group of user related information and historical data, this historical data comprises the User Activity of the user annotation of User Activity recognition result and correspondence thereof, wherein, the User Activity recognition result of mutual correspondence and the User Activity of user annotation can refer to the same activity for user, carry out calculating the User Activity recognition result and the User Activity that goes out of user's Direct Mark that obtain according to sensing data.User related information at least comprises one in acquisition time of the sensing data for determining the User Activity recognition result in above-mentioned historical data and place, userspersonal information, application using state information.Userspersonal information can comprise the sex, age, occupation etc. of user.These training datas can be stored in special knowledge base; Then correction model (or modification rule) is set up according to these training datas, make user related information can be adopted to revise the User Activity recognition result in training data based on this correction model, correction result is the User Activity of user annotation corresponding with this User Activity recognition result in training data.After acquisition correction model, can revise the first activity recognition result that upper step obtains based on this correction model.
Wherein, correction model can be predefined rule base, revises recognition result according to the reasoning on regular Sum fanction.Rule base and reasoning can use Ontology(ontology) mode represents rule and uses existing Ontology inference engine to carry out reasoning.First recognition result is represented be the probability vector of n class activity, then show that the n class activity vector may be engaged under locality, event and user's current state is comprehensive by inference engine.For " having a meal ", this activity is defined as a kind of activity occurred in locality.Other restrictive conditions may be also had for this activity.And " dining room " is defined as one " room " as the three unities and exists " dining table ".The restriction that it is movable that the position current according to user can judge that " dining room " meets " having a meal ".Therefore carry out consistency check by inference engine, thus determine that the activity of now user is " having a meal ".
In another specific implementation, if the first activity recognition result is expressed as the vector of a n kind activity probability, this as shown in Figure 2, specifically can comprise the steps: the makeover process of the first activity recognition result
Step 201, is expressed as binary set by correction model.
Step 202, calculates the vector of n kind activity probability and the product of described binary set, obtains probability component.
Step 203, in the described probability component obtained, the activity corresponding to probability component selecting numerical value maximum is as the second activity recognition result.
Such as, current according to user position defines respectively:
Eating(CURR_ACT);Restaurant(CURR_LOC);
Consistency checking is carried out in inference engine:
performedIn(CURR_ACT,CURR_LOC);isConsistent();
IsConsistent returns 1, if current location meets the condition that movable " having a meal " occurs; Otherwise return 0.Then the activity probability vector obtained by calculating learning algorithm is multiplied with the binary set that reasoning obtains.Then in probability vector be that the component of " 0 " is by reset in binary set.The activity that select probability is maximum from remaining non-" 0 " probability component is as the second activity recognition result.
Step 104, is defined as User Activity by the second activity recognition result.
The embodiment of the present invention sets up correction model by the relevant information and historical data adopting user, revising, can improve the accuracy rate to User Activity identification to calculating according to sensing data the activity recognition result obtained.
See Fig. 3, for embodiment of the present invention another kind identifies the method flow diagram of User Activity.
The present embodiment method is applied to the intelligent terminal being provided with at least one sensor equally, and the method for this identification User Activity can comprise:
Step 301, obtains the first data of sensor.
Step 302, carries out calculating acquisition first activity recognition result to the first data according to the first algorithm.
Step 303, adopts the correction model set up according to training data to revise the first activity recognition result, obtains the second activity recognition result.
Step 301 is similar to 103 with the step 101 in previous embodiment to 303, and wherein, the process obtaining the second activity recognition result in step 303 can with reference to step 201 to 203.
After the probability component of acquisition second activity recognition result, perform step 304 further.
Step 304, judges whether the probability component of the second activity recognition result is more than or equal to threshold value.
If be more than or equal to threshold value, then perform step 305, if be less than threshold value, then perform step 306.Wherein, the size of this threshold value can need to arrange according to user, does not limit herein.
Step 305, is defined as User Activity by the second activity recognition result.
After determining User Activity, proceed to step 310 to 311.
Step 306, exports the second activity recognition result to user.
If the probability component of this second activity recognition result is less than threshold value, then needs to determine that whether this recognition result is correct further, now first export this second activity recognition result to user, judge for user.
Step 307, receives the feedback information to the second activity recognition result of user's input.
User judges that whether the second activity recognition result is correct, if correctly, then can directly feed back for representing correct information; If incorrect, user can the incorrect information of feedback representation, and in feedback information, mark out User Activity further, also can only feed back the User Activity marked out.
Step 308, according to feedback information determination User Activity.
After receiving the feedback information of user, if user feedback is correct, then the second activity recognition result is defined as User Activity, the User Activity that if user feedback is incorrect or user feedback marks out, then the User Activity of this user annotation is defined as final User Activity, and performs step 309 further.
Step 309, when comprising the User Activity of user annotation in feedback information, the User Activity of the first activity recognition result and user annotation is increased in training data as one group of historical data, and according to adding the aforementioned correction model of training data correction after historical data.
After the User Activity obtaining user annotation, new one group of corresponding data calculating the User Activity of User Activity recognition result and the user annotation obtained according to sensing data can be obtained, this group corresponding data is increased in training data as historical data, and then the correction model can set up according to the new training data correction adding above-mentioned historical data, to make correction model more accurate.
After step 305 and step 308, step 310 can also be performed further to 311.
Step 310, according to User Activity, searches the application program corresponding with User Activity in knowledge base, wherein, stores the corresponding relation of User Activity and application program in knowledge base.
Set up the corresponding relation of User Activity and application program in advance, such as, sleep (belonging to User Activity) corresponding quiet setting (belonging to application program) etc.This corresponding relation can be that user pre-sets according to the demand of oneself.This corresponding relation can be stored in knowledge base together with training data.
After upper step determines final User Activity, this corresponding relation can be searched and obtain application program corresponding to this User Activity, and perform step 311.
Step 311, starts the application program corresponding with User Activity.
The embodiment of the present invention does not set up correction model by means of only the relevant information and historical data that adopt user, revise calculating according to sensing data the activity recognition result obtained, improve the accuracy rate to User Activity identification, and corresponding application service can be provided according to User Activity.
Such as, the probability component calculating " lying " in acquisition first activity recognition result according to the first data of sensor is maximum, then according to the relevant information of user as the age: 30, place: bedroom, the correction model that time: 12:30 etc. and historical data are set up, revises the first activity recognition result, the the second activity recognition result obtained after revising is " sleep ", now, the quiet of sleep correspondence pre-set can be arranged applications trigger, the sound of intelligent terminal is changed into quiet.Again such as, the probability component calculating " lying " in acquisition first activity recognition result according to the first data of sensor is maximum, then according to the relevant information of user as the age: 70, place: dining room, the correction model that time: 8::00 etc. and historical data are set up, first activity recognition result is revised, the the second activity recognition result obtained after revising is " falling down ", now, the corresponding warning of falling down pre-set can be arranged applications trigger, send warning the tinkle of bells by intelligent terminal or send alarming short message, phone etc.
Be more than the description to the inventive method embodiment, below the device realizing said method be introduced.
See Fig. 4, identify the structural representation of User Activity device for the embodiment of the present invention is a kind of.
This application of installation, in the intelligent terminal being provided with sensor, also can be exactly the intelligent terminal itself being provided with sensor.This device can comprise:
Acquiring unit 401, for obtaining the first data of described sensor.
Recognition unit 402, carries out calculating acquisition first activity recognition result for the first data obtained described acquiring unit 401 according to the first algorithm.
Amending unit 403, for adopting the correction model set up according to training data, the first activity recognition result that described recognition unit 402 identifies is revised, obtain the second activity recognition result, described training data comprises at least one group of user related information and historical data, described historical data comprises the User Activity of the user annotation of User Activity recognition result and correspondence thereof, described user related information at least comprises acquisition time and the place of the sensing data for determining described User Activity recognition result, userspersonal information, one in application using state information.
Determining unit 404, is defined as User Activity for the second activity recognition result described amending unit 403 being revised acquisition.
After acquiring unit 401 obtains the first data of sensor, by recognition unit 402 such as, according to certain algorithm, Active Learning Algorithm, calculate acquisition first activity recognition result, this the first activity recognition result can be expressed as the vector of a n kind activity probability, as (p1, p2, pn), wherein, pj is the probability of movable j, j=1 ... n.Then, amending unit 403 is revised this first activity recognition result according to the correction model set up in advance, obtains the second activity recognition result, and wherein, correction model is set up according at least one group of user related information and historical data.Finally, by determining unit 404, second activity recognition result is defined as User Activity.
The embodiment of the present invention adopts the relevant information of user and historical data to set up correction model by said units, revising, can improve the accuracy rate to User Activity identification to calculating according to sensing data the activity recognition result obtained.
In another embodiment of the invention, this first activity recognition result can be expressed as the vector of a n kind activity probability; As shown in Figure 5, amending unit 403 may further include:
Characterize subelement 501, for described correction model is expressed as binary set;
Computation subunit 502, the product of the binary set that vector and described sign subelement 501 for calculating described n kind activity probability characterize, obtains probability component;
Chooser unit 503, in the described probability component that obtains in computation subunit 502, the activity corresponding to probability component selecting numerical value maximum is as described second activity recognition result.
See Fig. 6, it is the structural representation of the another kind of User Activity recognition device of the embodiment of the present invention.
This device, except can comprising acquiring unit 601, recognition unit 602, amending unit 603 and determining unit 604, can further include:
First judging unit 605, for judging whether the probability component of described second activity recognition result is more than or equal to threshold value;
Output unit 606, during for judging that the probability component of described second activity recognition result is less than described threshold value when the first judging unit 605, exports described second activity recognition result to user;
Information receiving unit 607, for receiving the feedback information to the described second activity recognition result that described output unit 606 exports of user's input;
Described determining unit 604, specifically for when described first judging unit 605 judges that the probability component of described second activity recognition result is more than or equal to described threshold value, is defined as User Activity by described second activity recognition result; Also for when the first judging unit 605 judges that the probability component of the second activity recognition result is less than described threshold value, according to the described feedback information determination User Activity that information receiving unit 607 receives.
Second judging unit 608, for judging the User Activity whether comprising user annotation in the feedback information that described information receiving unit 607 receives.
Data adding device 609, during for comprising the User Activity of user annotation in judging when the second judging unit 608 the described feedback information that information receiving unit 607 receives, the User Activity of described first activity recognition result and described user annotation is increased in described training data as one group of historical data, and according to adding correction model described in the training data correction after historical data.
Search unit 610, for the User Activity determined according to described determining unit 604, in knowledge base, search the application program corresponding with described User Activity, wherein, in described knowledge base, store the corresponding relation of User Activity and application program;
Start unit 611, for starting the application program corresponding with described User Activity of searching unit 610 and finding.
The embodiment of the present invention not only adopts the relevant information of user and historical data to set up correction model by said units, revise calculating according to sensing data the activity recognition result obtained, improve the accuracy rate to User Activity identification, and corresponding application service can be provided according to User Activity.
As shown in Figure 7, the embodiment of the present invention additionally provides another kind of User Activity recognition device, and this device can be applied in and be provided with on the intelligent terminal of sensor, and this device comprises processor 710, storer 720 and transceiver 730.
Wherein, processor 710, storer 720, transceiver 730 are interconnected by bus 740; Bus 740 can be isa bus, pci bus or eisa bus etc.Described bus can be divided into address bus, data bus, control bus etc.
Transceiver 730 is for receiving the first data of described sensor.
Training data and one section of program code is stored in storer 720, wherein, described training data comprises at least one group of user related information and historical data, described historical data comprises the User Activity of the user annotation of User Activity recognition result and correspondence thereof, and described user related information at least comprises one in acquisition time of the sensing data for determining described User Activity recognition result and place, userspersonal information, application using state information; Particularly, described program code comprises computer-managed instruction.Storer 720 may comprise high-speed RAM storer, still may comprise nonvolatile memory (non-volatile memory), such as at least one magnetic disk memory.
Processor 710, for reading the program code in storer 720, and performs following steps:
According to the first algorithm, calculating acquisition first activity recognition result is carried out to described first data;
The correction model set up according to training data is adopted to revise described first activity recognition result;
Described second activity recognition result is defined as User Activity.
Those of ordinary skill in the art can recognize, in conjunction with unit and the algorithm steps of each example of embodiment disclosed herein description, can realize with the combination of electronic hardware or computer software and electronic hardware.These functions perform with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme.Professional and technical personnel can use distinct methods to realize described function to each specifically should being used for, but this realization should not thought and exceeds scope of the present invention.
Those skilled in the art can be well understood to, and for convenience and simplicity of description, the specific works process of the system of foregoing description, device and unit, with reference to the corresponding process in preceding method embodiment, can not repeat them here.
In several embodiments that the application provides, should be understood that disclosed system, apparatus and method can realize by another way.Such as, device embodiment described above is only schematic, such as, the division of described unit, be only a kind of logic function to divide, actual can have other dividing mode when realizing, such as multiple unit or assembly can in conjunction with or another system can be integrated into, or some features can be ignored, or do not perform.Another point, shown or discussed coupling each other or direct-coupling or communication connection can be by some interfaces, and the indirect coupling of device or unit or communication connection can be electrical, machinery or other form.
The described unit illustrated as separating component or can may not be and physically separates, and the parts as unit display can be or may not be physical location, namely can be positioned at a place, or also can be distributed in multiple network element.Some or all of unit wherein can be selected according to the actual needs to realize the object of the present embodiment scheme.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, also can be that the independent physics of unit exists, also can two or more unit in a unit integrated.
If described function using the form of SFU software functional unit realize and as independently production marketing or use time, can be stored in a computer read/write memory medium.Based on such understanding, the part of the part that technical scheme of the present invention contributes to prior art in essence in other words or this technical scheme can embody with the form of software product, this computer software product is stored in a storage medium, comprising some instructions in order to make a computer equipment (can be personal computer, server, or the network equipment etc.) or processor (processor) perform all or part of step of method described in each embodiment of the present invention.And aforesaid storage medium comprises: USB flash disk, portable hard drive, ROM (read-only memory) (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disc or CD etc. various can be program code stored medium.
The above; be only the specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, is anyly familiar with those skilled in the art in the technical scope that the present invention discloses; change can be expected easily or replace, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should described be as the criterion with the protection domain of claim.
Claims (12)
1. a User Activity recognition methods, is characterized in that, is applied to the intelligent terminal being provided with sensor, and described method comprises:
Obtain the first data of described sensor;
According to the first algorithm, calculating acquisition first activity recognition result is carried out to described first data;
The correction model set up according to training data is adopted to revise described first activity recognition result, obtain the second activity recognition result, described training data comprises at least one group of user related information and historical data, described historical data comprises the User Activity of the user annotation of User Activity recognition result and correspondence thereof, and described user related information at least comprises one in acquisition time of the sensing data for determining described User Activity recognition result and place, userspersonal information, application using state information;
Described second activity recognition result is defined as User Activity.
2. method according to claim 1, is characterized in that, described first activity recognition result is the vector of a n kind activity probability;
The correction model that described employing is set up according to training data is revised described first activity recognition result, obtains the second activity recognition result, comprising:
Described correction model is expressed as binary set;
Calculate the described vector of n kind activity probability and the product of described binary set, obtain probability component;
In the described probability component obtained, the activity corresponding to probability component selecting numerical value maximum is as described second activity recognition result.
3. method according to claim 2, is characterized in that, described described second activity recognition result is defined as User Activity before, also comprise:
Judge whether the probability component of described second activity recognition result is more than or equal to threshold value;
Described described second activity recognition result is defined as User Activity, comprises:
When the probability component of described second activity recognition result is more than or equal to described threshold value, described second activity recognition result is defined as User Activity.
4. method according to claim 3, is characterized in that, also comprises:
When the probability component of described second activity recognition result is less than described threshold value, export described second activity recognition result to user;
Receive the feedback information to described second activity recognition result of described user input;
According to described feedback information determination User Activity.
5. method according to claim 4, is characterized in that, also comprises:
When comprising the User Activity of user annotation in described feedback information, the User Activity of the user annotation in described first activity recognition result and described feedback information is increased in described training data as one group of historical data, and according to adding correction model described in the training data correction after historical data.
6. method as claimed in any of claims 1 to 5, is characterized in that, described described second activity recognition result is defined as User Activity after, also comprise:
According to described User Activity, in knowledge base, search the application program corresponding with described User Activity, wherein, in described knowledge base, store the corresponding relation of User Activity and application program;
Start described corresponding with described User Activity application program.
7. a User Activity recognition device, is characterized in that, this application of installation is in the intelligent terminal being provided with sensor, and described device comprises:
Acquiring unit, for obtaining the first data of described sensor;
Recognition unit, carries out calculating acquisition first activity recognition result for the first data obtained described acquiring unit according to the first algorithm;
Amending unit, for adopting the correction model set up according to training data, the first activity recognition result that described recognition unit identifies is revised, obtain the second activity recognition result, described training data comprises at least one group of user related information and historical data, described historical data comprises the User Activity of the user annotation of User Activity recognition result and correspondence thereof, and described user related information at least comprises one in acquisition time of the sensing data for determining described User Activity recognition result and place, userspersonal information, application using state information;
Determining unit, is defined as User Activity for the second activity recognition result described amending unit correction obtained.
8. device according to claim 7, is characterized in that, described first activity recognition result is the vector of a n kind activity probability;
Described amending unit comprises:
Characterize subelement, for described correction model is expressed as binary set;
Computation subunit, the product of the binary set that vector and described sign subelement for calculating described n kind activity probability characterize, obtains probability component;
Chooser unit, in the described probability component that obtains in described computation subunit, the activity corresponding to probability component selecting numerical value maximum is as described second activity recognition result.
9. device according to claim 8, is characterized in that, also comprises:
First judging unit, for judging whether the probability component of described second activity recognition result is more than or equal to threshold value;
Described determining unit, during specifically for judging that the probability component of described second activity recognition result is more than or equal to described threshold value when described first judging unit, is defined as User Activity by described second activity recognition result.
10. device according to claim 9, is characterized in that, also comprises:
Output unit, during for judging that the probability component of described second activity recognition result is less than described threshold value when described first judging unit, exports described second activity recognition result to user;
Information receiving unit, for receiving the feedback information to the described second activity recognition result that described output unit exports of described user input;
Described determining unit, the feedback information determination User Activity also for receiving according to described information receiving unit.
11. devices according to claim 10, is characterized in that, also comprise:
Second judging unit, for judging the User Activity whether comprising user annotation in the feedback information that described information receiving unit receives;
Data adding device, for when described second judging unit judges the User Activity comprising user annotation in described feedback information, the User Activity of described first activity recognition result and described user annotation is increased in described training data as one group of historical data, and according to adding correction model described in the training data correction after historical data.
12., according to the device in claim 7 to 11 described in any one, is characterized in that, also comprise:
Search unit, for according to the determined User Activity of described determining unit, in knowledge base, search the application program corresponding with described User Activity, wherein, in described knowledge base, store the corresponding relation of User Activity and application program;
Start unit, for searching the application program corresponding with described User Activity that unit finds described in starting.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310629291.6A CN104680046B (en) | 2013-11-29 | 2013-11-29 | A kind of User Activity recognition methods and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310629291.6A CN104680046B (en) | 2013-11-29 | 2013-11-29 | A kind of User Activity recognition methods and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104680046A true CN104680046A (en) | 2015-06-03 |
| CN104680046B CN104680046B (en) | 2018-09-07 |
Family
ID=53315078
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310629291.6A Active CN104680046B (en) | 2013-11-29 | 2013-11-29 | A kind of User Activity recognition methods and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104680046B (en) |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105468161A (en) * | 2016-01-21 | 2016-04-06 | 北京百度网讯科技有限公司 | Instruction execution method and device |
| CN105873110A (en) * | 2016-05-27 | 2016-08-17 | 北京灵龄科技有限责任公司 | User behavior correcting method and device |
| CN107146445A (en) * | 2017-06-30 | 2017-09-08 | 北京奇虎科技有限公司 | Mobile terminal traffic information reminding method, device and mobile terminal |
| CN107277154A (en) * | 2017-06-30 | 2017-10-20 | 北京奇虎科技有限公司 | Information-pushing method, device and mobile terminal |
| CN107341226A (en) * | 2017-06-30 | 2017-11-10 | 北京奇虎科技有限公司 | Information displaying method, device and mobile terminal |
| CN107368850A (en) * | 2017-06-30 | 2017-11-21 | 北京奇虎科技有限公司 | User Activity state switching method, device and mobile terminal based on mobile terminal |
| CN107391603A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | User's portrait method for building up and device for mobile terminal |
| CN107391605A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | Information-pushing method, device and mobile terminal based on geographical position |
| CN107396306A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | User Activity state identification method, device and mobile terminal based on mobile terminal |
| CN107391604A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | Map styles display methods, device and mobile terminal based on active state |
| CN107529135A (en) * | 2016-06-20 | 2017-12-29 | 同济大学 | User Activity type identification method based on smart machine data |
| CN109165057A (en) * | 2017-06-28 | 2019-01-08 | 华为技术有限公司 | A kind of method and apparatus that smart machine executes task |
| CN109242025A (en) * | 2018-09-14 | 2019-01-18 | 北京旷视科技有限公司 | Model iterative correction methods, apparatus and system |
| CN110672086A (en) * | 2018-07-03 | 2020-01-10 | 百度在线网络技术（北京）有限公司 | Scene recognition method, device, equipment and computer readable medium |
| CN111351524A (en) * | 2018-12-21 | 2020-06-30 | 亚玛芬体育数字服务公司 | Sensor data management |
| CN112232424A (en) * | 2020-10-21 | 2021-01-15 | 成都商汤科技有限公司 | Identity recognition method and device, electronic equipment and storage medium |
| CN113269129A (en) * | 2021-06-11 | 2021-08-17 | 成都商汤科技有限公司 | Identity recognition method and device, electronic equipment and storage medium |
| US11137820B2 (en) | 2015-12-01 | 2021-10-05 | Amer Sports Digital Services Oy | Apparatus and method for presenting thematic maps |
| US11145272B2 (en) | 2016-10-17 | 2021-10-12 | Amer Sports Digital Services Oy | Embedded computing device |
| US11144107B2 (en) | 2015-12-01 | 2021-10-12 | Amer Sports Digital Services Oy | Apparatus and method for presenting thematic maps |
| US11210299B2 (en) | 2015-12-01 | 2021-12-28 | Amer Sports Digital Services Oy | Apparatus and method for presenting thematic maps |
| US11215457B2 (en) | 2015-12-01 | 2022-01-04 | Amer Sports Digital Services Oy | Thematic map based route optimization |
| US11284807B2 (en) | 2015-12-21 | 2022-03-29 | Amer Sports Digital Services Oy | Engaging exercising devices with a mobile device |
| CN114416389A (en) * | 2021-12-01 | 2022-04-29 | 荣耀终端有限公司 | Activity identification method and related equipment |
| US11541280B2 (en) | 2015-12-21 | 2023-01-03 | Suunto Oy | Apparatus and exercising device |
| US11587484B2 (en) | 2015-12-21 | 2023-02-21 | Suunto Oy | Method for controlling a display |
| US11607144B2 (en) | 2015-12-21 | 2023-03-21 | Suunto Oy | Sensor based context management |
| US11838990B2 (en) | 2015-12-21 | 2023-12-05 | Suunto Oy | Communicating sensor data in wireless communication systems |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101989127A (en) * | 2009-08-03 | 2011-03-23 | 索尼公司 | Data processing apparatus, data processing method, and program |
| CN102314474A (en) * | 2010-06-24 | 2012-01-11 | 索尼公司 | Signal conditioning package, information handling system, information processing method and program |
| JP2013131150A (en) * | 2011-12-22 | 2013-07-04 | Dainippon Printing Co Ltd | Portable terminal having personal authentication function and application program |
| CN103414930A (en) * | 2012-07-27 | 2013-11-27 | Tcl集团股份有限公司 | Remote control system for identifying and sensing user and method thereof |
-
2013
- 2013-11-29 CN CN201310629291.6A patent/CN104680046B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101989127A (en) * | 2009-08-03 | 2011-03-23 | 索尼公司 | Data processing apparatus, data processing method, and program |
| CN102314474A (en) * | 2010-06-24 | 2012-01-11 | 索尼公司 | Signal conditioning package, information handling system, information processing method and program |
| JP2013131150A (en) * | 2011-12-22 | 2013-07-04 | Dainippon Printing Co Ltd | Portable terminal having personal authentication function and application program |
| CN103414930A (en) * | 2012-07-27 | 2013-11-27 | Tcl集团股份有限公司 | Remote control system for identifying and sensing user and method thereof |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11144107B2 (en) | 2015-12-01 | 2021-10-12 | Amer Sports Digital Services Oy | Apparatus and method for presenting thematic maps |
| US11215457B2 (en) | 2015-12-01 | 2022-01-04 | Amer Sports Digital Services Oy | Thematic map based route optimization |
| US11210299B2 (en) | 2015-12-01 | 2021-12-28 | Amer Sports Digital Services Oy | Apparatus and method for presenting thematic maps |
| US11137820B2 (en) | 2015-12-01 | 2021-10-05 | Amer Sports Digital Services Oy | Apparatus and method for presenting thematic maps |
| US11587484B2 (en) | 2015-12-21 | 2023-02-21 | Suunto Oy | Method for controlling a display |
| US11541280B2 (en) | 2015-12-21 | 2023-01-03 | Suunto Oy | Apparatus and exercising device |
| US11284807B2 (en) | 2015-12-21 | 2022-03-29 | Amer Sports Digital Services Oy | Engaging exercising devices with a mobile device |
| US11838990B2 (en) | 2015-12-21 | 2023-12-05 | Suunto Oy | Communicating sensor data in wireless communication systems |
| US11607144B2 (en) | 2015-12-21 | 2023-03-21 | Suunto Oy | Sensor based context management |
| CN105468161A (en) * | 2016-01-21 | 2016-04-06 | 北京百度网讯科技有限公司 | Instruction execution method and device |
| CN105468161B (en) * | 2016-01-21 | 2019-03-12 | 北京百度网讯科技有限公司 | Instruction executing method and device |
| CN105873110A (en) * | 2016-05-27 | 2016-08-17 | 北京灵龄科技有限责任公司 | User behavior correcting method and device |
| CN105873110B (en) * | 2016-05-27 | 2020-01-10 | 集道成科技(北京)有限公司 | User behavior correction method and device |
| CN107529135A (en) * | 2016-06-20 | 2017-12-29 | 同济大学 | User Activity type identification method based on smart machine data |
| US11145272B2 (en) | 2016-10-17 | 2021-10-12 | Amer Sports Digital Services Oy | Embedded computing device |
| CN109165057A (en) * | 2017-06-28 | 2019-01-08 | 华为技术有限公司 | A kind of method and apparatus that smart machine executes task |
| CN109165057B (en) * | 2017-06-28 | 2021-03-30 | 华为技术有限公司 | Method and device for executing task by intelligent equipment |
| CN107391604A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | Map styles display methods, device and mobile terminal based on active state |
| CN107368850A (en) * | 2017-06-30 | 2017-11-21 | 北京奇虎科技有限公司 | User Activity state switching method, device and mobile terminal based on mobile terminal |
| CN107391603B (en) * | 2017-06-30 | 2020-12-18 | 北京奇虎科技有限公司 | User portrait establishing method and device for mobile terminal |
| CN107341226B (en) * | 2017-06-30 | 2021-06-22 | 北京奇虎科技有限公司 | Information display method and device and mobile terminal |
| CN107146445A (en) * | 2017-06-30 | 2017-09-08 | 北京奇虎科技有限公司 | Mobile terminal traffic information reminding method, device and mobile terminal |
| CN107277154A (en) * | 2017-06-30 | 2017-10-20 | 北京奇虎科技有限公司 | Information-pushing method, device and mobile terminal |
| CN107341226A (en) * | 2017-06-30 | 2017-11-10 | 北京奇虎科技有限公司 | Information displaying method, device and mobile terminal |
| CN107391603A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | User's portrait method for building up and device for mobile terminal |
| CN107396306A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | User Activity state identification method, device and mobile terminal based on mobile terminal |
| CN107391605A (en) * | 2017-06-30 | 2017-11-24 | 北京奇虎科技有限公司 | Information-pushing method, device and mobile terminal based on geographical position |
| CN110672086B (en) * | 2018-07-03 | 2023-01-31 | 百度在线网络技术（北京）有限公司 | Scene recognition method, device, equipment and computer readable medium |
| CN110672086A (en) * | 2018-07-03 | 2020-01-10 | 百度在线网络技术（北京）有限公司 | Scene recognition method, device, equipment and computer readable medium |
| CN109242025A (en) * | 2018-09-14 | 2019-01-18 | 北京旷视科技有限公司 | Model iterative correction methods, apparatus and system |
| CN111351524A (en) * | 2018-12-21 | 2020-06-30 | 亚玛芬体育数字服务公司 | Sensor data management |
| CN112232424A (en) * | 2020-10-21 | 2021-01-15 | 成都商汤科技有限公司 | Identity recognition method and device, electronic equipment and storage medium |
| WO2022257306A1 (en) * | 2021-06-11 | 2022-12-15 | 成都商汤科技有限公司 | Identity identification method and apparatus, electronic device, and storage medium |
| CN113269129A (en) * | 2021-06-11 | 2021-08-17 | 成都商汤科技有限公司 | Identity recognition method and device, electronic equipment and storage medium |
| CN114416389A (en) * | 2021-12-01 | 2022-04-29 | 荣耀终端有限公司 | Activity identification method and related equipment |
| CN114416389B (en) * | 2021-12-01 | 2023-03-03 | 荣耀终端有限公司 | Activity identification method and related equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104680046B (en) | 2018-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104680046A (en) | User activity recognition method and device | |
| KR101672864B1 (en) | User activity tracking system and device | |
| EP3985559A1 (en) | Entity semantics relationship classification | |
| JP2017130182A (en) | System, method, program and device for associating image with facility | |
| CN108629043A (en) | Extracting method, device and the storage medium of webpage target information | |
| US10740216B1 (en) | Automatic bug classification using machine learning | |
| KR101220557B1 (en) | Method and system for searching mobile application using human activity knowledge database | |
| CN111340054A (en) | Data labeling method and device and data processing equipment | |
| US11144594B2 (en) | Search method, search apparatus and non-temporary computer-readable storage medium for text search | |
| CN108846097B (en) | User interest tag representation method, article recommendation device and equipment | |
| CN104915351A (en) | Picture sorting method and terminal | |
| CN103646070A (en) | Data processing method and device for search engine | |
| CN104954988B (en) | A kind of user's trip based reminding method and intelligent terminal applied to intelligent terminal | |
| JP2013250861A (en) | Information processing apparatus, information processing method and program | |
| CN105095311B (en) | The processing method of promotion message, apparatus and system | |
| US10929775B2 (en) | Statistical self learning archival system | |
| JP2013250862A (en) | Information processing apparatus, information processing method and program | |
| CN115268719B (en) | Method, medium and electronic device for positioning target element on interface | |
| KR20190031421A (en) | Method for recommending designated items | |
| CN106598379A (en) | Method and device for managing application program | |
| CN104516938A (en) | Electronic computing device and personalized information recommendation method thereof | |
| CN110472057B (en) | Topic label generation method and device | |
| JP6183489B2 (en) | Information processing system, information processing method, and program | |
| US20170212957A1 (en) | Context identification for content generation | |
| CN105049615B (en) | A kind of user's trip based reminding method and intelligent terminal applied to intelligent terminal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |