CN111126294A - Method and server for recognizing gait of terminal user based on mobile terminal data - Google Patents

Method and server for recognizing gait of terminal user based on mobile terminal data Download PDF

Info

Publication number
CN111126294A
CN111126294A CN201911359210.9A CN201911359210A CN111126294A CN 111126294 A CN111126294 A CN 111126294A CN 201911359210 A CN201911359210 A CN 201911359210A CN 111126294 A CN111126294 A CN 111126294A
Authority
CN
China
Prior art keywords
gait
data
user
mobile terminal
ltoreq
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
CN201911359210.9A
Other languages
Chinese (zh)
Other versions
CN111126294B (en
Inventor
董霖
杨玉春
曹克丹
叶新江
方毅
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.)
Zhejiang Meiri Interdynamic Network Technology Co ltd
Original Assignee
Zhejiang Meiri Interdynamic Network Technology 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 Zhejiang Meiri Interdynamic Network Technology Co ltd filed Critical Zhejiang Meiri Interdynamic Network Technology Co ltd
Priority to CN201911359210.9A priority Critical patent/CN111126294B/en
Publication of CN111126294A publication Critical patent/CN111126294A/en
Application granted granted Critical
Publication of CN111126294B publication Critical patent/CN111126294B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/20—Movements or behaviour, e.g. gesture recognition
    • G06V40/23—Recognition of whole body movements, e.g. for sport training
    • G06V40/25—Recognition of walking or running movements, e.g. gait recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching
    • G06F2218/16—Classification; Matching by matching signal segments
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Psychiatry (AREA)
  • Social Psychology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Theoretical Computer Science (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

The invention relates to a method for identifying gait of a terminal user based on mobile terminal data, which comprises the following steps: step 100, obtaining at least one motion data C of the mobile terminal1(n)、C2(n)、…、Cm(n) wherein the x-th motion data Cx(n) is a time period T ═ T1,t2]The mobile terminal collects the signals according to a first preset collection frequency F, wherein x is more than or equal to 1 and less than or equal to m; step 200, respectively setting C at a preset sliding length delta W by using a rectangular window Rt with a window length W1(n)、C2(n)、…、Cm(n) sliding up to obtain first step profile D of said end-user0,D1,…,DuWherein D isiComprising rectangular windows Rt respectively at C1(n)、C2(n)、…、Cm(n) i is 0. ltoreq. i.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.<W; step 300, based on the first step characteristic data D0,D1,…,DuAnd identifying the gait of the terminal user by a user gait identification model; further, W is larger than or equal to the standard gait cycle length, and the standard gait cycle length is the maximum value of a plurality of known gait cycle lengths.

Description

Method and server for recognizing gait of terminal user based on mobile terminal data
Technical Field
The invention relates to a gait recognition technology, in particular to a method and a server for recognizing gait of a terminal user.
Background
The identification of individuals or groups of end users is of high interest, wherein identification of individuals or groups based on biometric features unique to the end user, such as facial features, fingerprint features, and gait of the user, is an important identification method in the prior art, and identification of individuals or groups based on gait is a relatively new part of current identification techniques.
Gait recognition is the process of recognizing a travel gesture (a travel gesture is a gesture specific to a user or group of users in a particular travel pattern, which may include walking, running, etc., for example), and may also be referred to as individual recognition or group recognition since gait is specific to an end user or group of end users. In the prior art, a method for recognizing a walking gesture (for example, a walking gesture) according to all frame images in a video is currently used, but because the number of frame images generally contained in the video is large, the amount of data required to be input when gait recognition is performed is large, which limits the application of gait recognition to a certain extent. In order to solve the problem of excessive input data volume, technical contents for performing gait recognition based on acceleration data of an end user appear: estimating the gait cycle of the end user by adopting a peak detection method (the gait cycle is a time period from the rise of a landing foot to the landing of the landing foot again is one gait cycle), then segmenting the acceleration data of the user based on the estimated gait cycle, and then carrying out gait recognition based on segmented synthetic acceleration data, wherein the segmented synthetic acceleration data is the gait feature data of the user carrying out the gait recognition. However, the method for obtaining the estimated gait cycle based on the peak detection method is only suitable for the laboratory environment, that is, under the ideal environment, the method can have better identification effect; in an actual environment, there are various situations such as uneven roads (for example, uneven roads) and the like, if a peak detection method is used to estimate the gait cycle of a user, the estimated gait cycle and the accurate gait cycle have large errors due to a false peak, and at this time, if segmented synthetic acceleration data (namely, gait feature data of the user) is acquired according to the estimated gait cycle, the acquired gait feature data of the user is not complete gait feature data in one gait cycle of the user, so that the gait recognition effect of the user is affected, and the gait recognition effect of the user cannot meet the actual requirements.
Disclosure of Invention
In order to solve the technical problem, the invention discloses a method for identifying the gait of a terminal user based on mobile terminal data, which comprises the following steps: step 100, obtaining at least one motion data C of the mobile terminal1(n)、C2(n)、…、Cm(n) wherein the x-th motion data Cx(n) is a time period T ═ T1,t2]The mobile terminal collects the signals according to a first preset collection frequency F, wherein x is more than or equal to 1 and less than or equal to m; step 200, respectively setting C at a preset sliding length delta W by using a rectangular window Rt with a window length W1(n)、C2(n)、…、Cm(n) sliding up to obtain first step profile D of said end-user0,D1,…,DuWherein D isiComprising rectangular windows Rt respectively at C1(n)、C2(n)、…、Cm(n) i is 0. ltoreq. i.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.<W，
Figure BDA0002336738260000021
I.e. the ratio of u to Δ W is rounded down; step 300, based on the first step characteristic data D0,D1,…,DuAnd identifying the gait of the terminal user by a user gait identification model; further, W is larger than or equal to the standard gait cycle length, and the standard gait cycle length is the maximum value of a plurality of known gait cycle lengths.
By adopting the rectangular window with the window length W being larger than the length of one gait cycle of the terminal user, the acquired gait feature data of the terminal user at least comprises all information data in one complete gait cycle of the terminal user, and more comprehensive and complete information is provided for the gait recognition of the terminal user; in addition, compared with the gait feature data which is obtained in the prior art and is in a single type and is lifted from one foot to the landing of the foot, the sliding length delta W of the rectangular window in the invention is smaller than the window length W, so the method can also obtain various gait feature data with different initial states of the terminal user in a specific advancing mode (for example, the initial state of one gait feature data is that one foot is on the ground, the initial state of the other gait feature data is that the highest point of the lifted foot is the highest point of the other foot, and the like), the gait feature data types are more abundant, and the gait feature data with the abundant types can effectively improve the accuracy of identifying the gait of the terminal user.
Drawings
FIG. 1 is a flow chart of a method of identifying an end user's gait based on mobile terminal data in accordance with the invention;
FIG. 2 is another flow chart of a method of identifying an end user's gait based on mobile terminal data in accordance with the invention;
fig. 3 is a flowchart of a method for acquiring motion data of a mobile terminal according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings. This description is made by way of example and not limitation to specific embodiments consistent with the principles of the invention, the description being in sufficient detail to enable those skilled in the art to practice the invention, other embodiments may be utilized and the structure of various elements may be changed and/or substituted without departing from the scope and spirit of the invention. The following detailed description is, therefore, not to be taken in a limiting sense.
The invention discloses a method for identifying gait of a terminal user based on mobile terminal dataAs shown in fig. 1, the method comprises: step 100, obtaining at least one motion data C of the mobile terminal1(n)、C2(n)、…、Cm(n) wherein the x-th motion data Cx(n) is a time period T ═ T1,t2]The mobile terminal collects the signals according to a first preset collection frequency F, and x is more than or equal to 1 and less than or equal to m. In the present invention, the motion data may be acceleration data of the mobile terminal in different directions, motion speed data of the mobile terminal in different directions, or a rotation angle of the mobile terminal, and the like, and the above-mentioned examples are not intended to be specific limitations on the motion data of the present invention, and further include other motion data about the mobile terminal that can be acquired in the future. Exemplarily, m is 3, wherein C1(n) is acceleration data of the mobile terminal on the X axis of the mobile terminal coordinate system, C2(n) is acceleration data of the mobile terminal on the Y axis of the mobile terminal coordinate system, C3And (n) is the acceleration data of the mobile terminal on the Z axis of the mobile terminal coordinate system.
According to the present invention, the mobile terminal is a mobile phone and/or a PAD device, and it is known to those skilled in the art that the mobile terminal is integrated with commonly used sensor elements, such as an acceleration sensor for measuring acceleration of the mobile terminal in multiple directions, a gyroscope sensor for measuring a rotation angle of the mobile terminal, and the like; and the mobile terminal can acquire the data of the above-mentioned sensors by calling a specific interface.
According to the invention, t2-t1The value of (a) and the value of the first preset acquisition frequency F can be set by self-definition, and in one embodiment of the present invention, t is2-t1Such that acceleration data within at least one rectangular window of the end user can be obtained within said time period T, exemplary T2-t1Is in the range of [10 seconds, 70 seconds ]]Preferably 60 seconds, and when said t is2-t1When the value of (2) is larger, the acquired motion data is increased, so that the sliding times of the rectangular window on the motion data are increased, and further, the sliding length of the rectangular window is smaller than the window length of the rectangular window, so that the method is beneficial to the aspect that the sliding length of the rectangular window is smaller than the window length of the rectangular windowWhen the rectangular window slides on the motion data, a plurality of gait feature data with different initial states of the terminal user under a specific travelling mode can be obtained (for example, the initial state of one gait feature data is that one foot is on the ground, the initial state of the other gait feature data is that the foot is at the highest point of being lifted, and the like), so that the gait feature data available for the user during gait recognition is more sufficient and comprehensive. According to the invention, the gait feature data can reflect the individual advancing gesture features of the terminal users or the group advancing gesture features of the terminal users. In the invention, based on the gait feature data distributed with different labels and the gait recognition models with different targets, the individual gait of the terminal user can be recognized, and the gait of the terminal user group can also be recognized.
The value of the first preset acquisition frequency F is considered both data redundancy and data integrity, and when the value of the first preset acquisition frequency F is large, the acquired data volume is too large, so that the data redundancy is high, namely, the acquired data contains too much useless or repeated information; conversely, when the value of the first preset acquisition frequency F is small, the amount of data acquired in the time period T is too small, and the acquired gait feature data of the terminal user is incomplete. In one embodiment, the first preset acquisition frequency F has a value range of [80,150]And more preferably 100, i.e. 100 data acquisitions per second. For example, assume that the time period T is [11:00,11:01 ═]F is 100, m is 3 and C1(n)、C2(n)、C3When (n) is acceleration data of the mobile terminal in X, Y, Z three different directions of the mobile terminal coordinate system, the value range of n is 1,2, …,6000, where 6000 is the product of F and the time difference of the time period T, i.e. C1(n)＝{x1,x2,…,x6000Is acceleration data of the mobile terminal in the X-axis direction, C2(n)＝{y1,y2,…,y6000Is acceleration data of the mobile terminal in the Y-axis direction, C3(n)＝{z1,z2,…,z6000Is the acceleration number of the mobile terminal in the Z-axis directionAccordingly. As will be appreciated by those skilled in the art, the motion data C1(n)、C2(n)、…、CmAnd (n) is a time domain discrete signal of different mobile terminals.
Step 200, respectively setting C at a preset sliding length delta W by using a rectangular window Rt with a window length W1(n)、C2(n)、…、Cm(n) sliding up to obtain first step profile D of said end-user0,D1,…,DuWherein D isiComprising rectangular windows Rt respectively at C1(n)、C2(n)、…、Cm(n) i is 0. ltoreq. i.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.<W，
Figure BDA0002336738260000041
I.e. the ratio of u to Δ W is rounded down; further, W is larger than or equal to the standard gait cycle length, and the standard gait cycle length is the maximum value of a plurality of known gait cycle lengths.
According to the present invention, the processing of the data in the rectangular window Rt may be a functional relationship processing commonly used in the art, for example, the processing is to multiply the data in the rectangular window Rt by a fixed value, so as to facilitate the extraction of the gait feature data of the user.
According to the invention, W and the preset sliding length Δ W can be set by users. In the present invention, the standard gait cycle length can be obtained in a variety of ways: on one hand, the gait cycle length of the mobile terminal can be acquired according to the gait cycle length of the terminal client of the mobile terminal reported by the mobile terminal, in the mode, the terminal user of the mobile terminal actively or passively uploads the gait cycle length of the mobile terminal, and the maximum value is selected as the standard gait cycle length based on the acquired gait cycle lengths of all the terminal clients; on the other hand, the gait cycle length of the mobile terminal in the laboratory environment is measured, and the maximum value is taken as the standard gait cycle length; finally, it may also be that the standard gait cycle length is obtained based on a sufficient number of known end user gait cycle lengths, again with the maximum of the known end user gait cycle lengths as the standard gait cycle length. Further, in the present invention, the value of W ranges from [4 seconds, 10 seconds ], preferably 5 seconds. According to the invention, W is a function of Δ W, preferably, W ═ Δ W/2, where the preset sliding length Δ W takes a value in the range of [2 seconds, 5 seconds ], preferably 2.5 seconds.
According to the invention, the rectangular window Rt is a rectangular window with a fixed value, which may be, for example, 1. When the rectangular window Rt is in motion data Cx(n), when sliding on the rectangular window Rt, the starting point of the first sliding on the rightmost end of the rectangular window Rt is W × F/2, where W × F is the amount of data contained in the rectangular window Rt. To illustrate more clearly the rectangular window Rt is in at least one motion datum C1(n)、C2(n)、…、Cm(n) principle of sliding up, we set: the rectangular window Rt is a rectangular window with a value of 1, and the processing of the data in the rectangular window Rt is to multiply the data in the rectangular window Rt by 1, m is 2, C1(n)＝{1,2,3,4,5}，C2And (n) {6,7,8,9,10}, where W ═ Δ W/2, and values of W and F are such that the amount of data contained in the rectangular window Rt is 2, the amount of data contained in the preset sliding length Δ W for each sliding of the rectangular window Rt is 1. When the rectangular window Rt is at C for 0 th time1(n) or C2(n) when sliding on (i.e. the first sliding), the starting points of the rightmost ends of the rectangular window Rt are 1 and 6 respectively, and the sliding length is 1, and it can be seen that the data contained in the rectangular window Rt after sliding is 1,2 or 6,7 respectively, that is, in one embodiment of the invention, D is0＝[1,2，6,7]Preferably, D0＝[1,2；6,7](ii) a When the rectangular window Rt is at C for the 1 st time1(n) when the window is slid upwards, the starting point of the rightmost end of the rectangular window Rt is the position where the rightmost end of the rectangular window stays last time, and is 2, after the window is slid, the data contained in the rectangular window Rt are respectively 2 and 3, and so on. Similarly, the rectangular window is at Rt at C2(n) sliding mode and its sliding on C1The same way of sliding on (n) is used. Then, according to the above, D1＝[2,3,7,8]Or D is1＝[2,3；7,8]. Those skilled in the art will appreciate that the foregoing is illustrativeAre exemplary only and not intended as the only examples to limit the invention.
Step 300, based on the first step characteristic data D0,D1,…,DuAnd the user gait recognition model recognizes the gait of the end user. According to the invention, the user gait recognition model is obtained by training of SVM, neural network and the like, preferably, the user gait recognition model is obtained by training of an adaptive lifting method AdaBoost. Specifically, in the present invention, first step characteristic data of each mobile terminal in a training sample library for a user gait recognition model in a target traveling mode and a corresponding user gait tag thereof are obtained according to the foregoing steps 100 to 200, wherein the user gait tag corresponds to each mobile terminal, and when recognizing the gait of an end user group, the user gait tag is related to whether the end user belongs to the end user group, that is, when the end user belongs to the end user group, the user gait tag at this time may be set to 1, otherwise, the user gait tag is set to 0; or when the end user belongs to the category 1 end user group, the current user gait label may be set to 1, when the end user belongs to the category 2 end user group, the current user gait label may be set to 2, and so on, and those skilled in the art will know that the above user gait label setting method is only an exemplary method and is not a sole example for limiting the present invention; when the gait of the individual terminal user is identified, the user gait label is related to the individual terminal user, for example, if the training sample library for the user gait identification model contains num mobile terminals, the user gait labels can be set to 1,2, … and num, respectively, and when the first step feature data is the user gait feature data of the qth mobile terminal in the target advancing mode, the user gait label corresponding to the first step feature data is q, and q is greater than or equal to 1 and less than or equal to num. As will be appreciated by those skilled in the art, the SVM, neural network, or AdaB is trained using the first-state feature data of each mobile terminal in the training sample library in the target travel mode and the corresponding user gait signature thereof using the user gait recognition model obtained according to the foregoing steps 100-200oost, namely obtaining the user gait recognition model.
According to the invention, the first step profile D is used0,D1,…,DuIs inputted into the user gait recognition model, and as the input data of the user gait recognition model, the first step characteristic data D can be obtained0,D1,…,DuEach corresponding to a user gait tag. Specifically, in one embodiment of the present invention, the first step profile D is compared with the first step profile D0,D1,…,DuAny one of the user gait tags is used as the gait recognition result of the end user, and illustratively, 10 pieces of first step-state feature data D of the end user with the user gait tag of 5 are obtained0,D1,…,D10Wherein, the gait recognition model of the user can correctly recognize 8 first step state feature data as 5, and recognize the rest 2 as 1 and 3 respectively, when selecting 10 first step state feature data D0,D1,…,D10When any one of the user gait tags is used as the gait recognition result of the terminal user, the user gait recognition accuracy of the embodiment is 80%; in a preferred embodiment of the present invention, the class with the highest vote is the gait recognition result of the end user, that is, when the gait recognition result of the end user is obtained by using the voting method, it is known that in the above example, the gait recognition result of the end user is 5, and the output result is correct, that is, the voting method can further improve the accuracy of the gait recognition result of the user.
According to the content, the rectangular window with the window length W larger than the length of one gait cycle of the terminal user is adopted, so that the acquired gait feature data of the terminal user at least comprises all information data in one complete gait cycle of the terminal user, and more comprehensive and complete information is provided for the gait recognition of the terminal user; in addition, compared with the gait feature data which is acquired in the prior art and is in a single type and is lifted from one foot to the landing of the foot, because the sliding length delta W of the rectangular window in the invention is smaller than the window length W, the method can also acquire various gait feature data of different initial states of the terminal user in a specific advancing mode (for example, the initial state of one type of gait feature data is that one foot is on the ground, the initial state of the other type of gait feature data is that the foot is at the highest point of the lifted foot, and the like), the gait feature data types are more abundant, and the gait of the terminal user can be effectively identified based on the gait feature data with abundant types.
According to a preferred embodiment of the present invention, the step 300 comprises (as shown in fig. 2):
step 301, based on the first step state feature data D0,D1,…,DuAcquiring second-step state feature data VD of the terminal user by the gait abnormal model of the terminal user1,VD2,…,VDsWherein, the gait abnormal model of the terminal user is used for judging the first step characteristic data D of the terminal useriWhether the data is normal data in a target traveling mode or not, and the second-step state feature data VD1,VD2,…,VDsAre normal data under the target traveling mode, and s is less than or equal to u + 1. According to the invention, said step 301 comprises:
step 401, the first step state feature data D is processed0,D1,…,DuInputting the gait abnormal model of the end user to obtain the first step state feature data D0,D1,…,DuIs abnormal identification information L0,L1,…,LuWherein L isiIs DiThe abnormality identification information of (1). In the present invention, the normal data in the target traveling mode includes first step characteristic data obtained by the terminal user under different conditions in the traveling process according to the target traveling mode (for example, when the target traveling mode is walking, the different conditions in the traveling process according to the target traveling mode may include various conditions that the mobile terminal is always located in a clothing pocket, a user hand or a user bag during the walking process of the terminal user), and the abnormal data in the target traveling mode includes first step characteristic data obtained by the terminal user under different conditions in the traveling process not according to the target traveling modeThe state feature data (according to the above example, the abnormal data in the target traveling mode includes the first-step feature data acquired under different conditions during the running, jumping, standing and the like of the user, such as the mobile terminal running in the hand, the mobile terminal standing in the pocket of clothes and the like).
According to the invention, all first step characteristic data of each mobile terminal in the mobile terminal group for training the gait abnormal model of the terminal user in the target advancing mode and all first step characteristic data of each mobile terminal in the non-target advancing mode are obtained by adopting the method shown in the steps 100-200, and all positive samples and all negative samples are input into the models of SVM, BP neural network, sequence to sequence and the like for model training so as to obtain the gait abnormal model of the terminal user. And those skilled in the art will appreciate that DiD can be output after being input into the abnormal gait model of the end useriIs abnormal identification information LiWherein L isiCan be defined as desired, for example when Li1 represents the first step profile DiFor normal data in the target travel pattern, when L isi0 represents the first step profile DiIs abnormal data in the target traveling manner. And as known to those skilled in the art, LiThere may be other definitions, and LiIs defined in a manner that does not affect the scope of the present invention.
Step 402, based on D0,D1,…,DuAnd L0,L1,…,LuAcquiring second-step state feature data VD of the terminal user1,VD2,…,VDs. Specifically, in the present invention, when an abnormality flag L is presentiRepresents DiReserving D for normal data of terminal user in target traveling modeiAt this time, all the retained first-step feature data constitute the second-step feature data VD of the end user1,VD2,…,VDs。
Step 302, the second step state characteristic data VD is processed1,VD2,…,VDsAnd as input data of a user gait recognition model, recognizing the gait of the terminal user by using the user gait recognition model. In the invention, the second step state characteristic data VD1,VD2,…,VDsThe method for obtaining the user gait recognition result by using the first-step characteristic data as the input data of the user gait recognition model is the same as the method for obtaining the user gait recognition result by using the first-step characteristic data as the input data of the user gait recognition model, and the method is not repeated herein.
As can be seen from the above, by inputting the acquired first-step feature data into the gait abnormal model of the end user, abnormal data in the first-step feature data, for example, abnormal data that is not the user in the target traveling mode, can be removed, and the gait recognition effect of the end user can be further improved.
According to the invention, said at least one motion data C1(n)、C2(n)、…、Cm(n) obtained by the third thread of the mobile terminal by receiving and processing data collected by the second thread of the mobile terminal according to the first preset collection frequency F from a first array set so as to store the main thread of the mobile terminal for a period of time P ═ P1,p2]The data of at least one target object in the mobile terminal are acquired according to a second preset acquisition frequency E, the array size of the first array is the total number of data acquired by the at least one target object every time the main thread of the mobile terminal polls, and the at least one target object is related to the motion of the mobile terminal.
According to the present invention, the at least one target object is a sensor related to the motion of the mobile terminal, such as an acceleration sensor, a gravity sensor, a geomagnetic sensor, a gyroscope sensor, etc., and those skilled in the art will appreciate that the above examples are not meant to be limiting examples of the present invention, and may also include other elements related to the motion of the mobile terminal in the future. In the present invention, the at least one motion data C1(n)、C2(n)、…、Cm(n) the same target object is in different parts of the coordinate system of the mobile terminalData acquired upward and/or one-dimensional data acquired by another target object. Exemplarily, the at least one motion data C1(n)、C2(n)、…、CmAnd (n) is data respectively collected by the acceleration sensor in the mobile terminal in three directions of a coordinate system of the mobile terminal. According to the invention, the second preset acquisition frequency E is a function of the first preset acquisition frequency F, in a preferred embodiment, the second preset acquisition frequency E is k times the first preset acquisition frequency F, and k has a value range of [1,6 ]]Preferably, k has a value of 3.
According to an embodiment of the present invention, the present invention further discloses a server for identifying gait of an end user based on mobile terminal data, which includes a processor and a non-transitory computer-readable storage medium storing a computer program, wherein when the computer program is executed by the processor, the method for identifying gait of an end user based on mobile terminal data as described above is implemented, and details are not repeated herein. And those skilled in the art will recognize that the server may be any server in the prior art, and those skilled in the art will understand that the type, brand, and/or configuration of the server, etc. are not used as conditions for limiting the scope of the present invention.
According to another embodiment of the present invention, the present invention further discloses a storage medium storing a computer program, which when executed, implements any one of the aforementioned methods for identifying a gait of an end user based on mobile terminal data, and will not be described herein again.
Further, in the present invention, a method for acquiring motion data of a mobile terminal is also disclosed, the method comprising (as shown in fig. 3):
step 001, the main thread of the mobile terminal acquires a time period P ═ P according to a second preset acquisition frequency E1,p2]Storing data of at least one target object in the mobile terminal in a first array, wherein the array size of the first array is the total number of data acquired by the at least one target object by the main thread in each polling, and the at least one target object and the mobile terminalThe movement is correlated.
According to the present invention, the at least one target object is a sensor related to the motion of the mobile terminal, such as an acceleration sensor, a gravity sensor, a geomagnetic sensor, a gyroscope sensor, etc., and those skilled in the art will appreciate that the above examples are not meant to be limiting examples of the present invention, and may also include other elements related to the motion of the mobile terminal in the future. In the present invention, the motion data may be data acquired by the same target object in different directions of a mobile terminal coordinate system and/or one-dimensional data acquired by another target object, in one embodiment, the motion data is data acquired by a acceleration sensor in the mobile terminal in three directions of the mobile terminal coordinate system, respectively, and in another embodiment, the motion data is data acquired by the acceleration sensor, a gravity sensor, and the like in the mobile terminal in three directions of the mobile terminal coordinate system, respectively. Illustratively, when the motion data are data respectively acquired by an acceleration sensor in the mobile terminal in three directions of a coordinate system of the mobile terminal, the main thread respectively acquires the data of the acceleration sensor in the three directions, so that the acceleration sensor can acquire 3 data by polling once, and the size of the first array is 3.
According to the invention, the second preset acquisition frequency E is a function of the first preset acquisition frequency F, in a preferred embodiment, the second preset acquisition frequency E is k times the first preset acquisition frequency F, and k has a value range of [1,6 ]]Preferably, k has a value of 3. In the invention, the second preset acquisition frequency E is set to enable the system data callback interface of the mobile terminal to acquire enough sampling values, so that the second thread for subsequent resampling can acquire the complete enough movement data of the terminal user. Further, according to the present invention, p2-p1Value of is not less than t2-t1So as to obtain the motion data of the terminal user in the time period T.
According to the present invention, taking the at least one target object as an acceleration sensor as an example, since values in three directions of the acceleration sensor are obtained, and the size of the first array is 3, the value in the first array is updated after the main thread polls the acceleration sensor once, that is, the latest acceleration sensor data obtained for the main thread is always stored in the first array.
And 002, the second thread acquires data from the first array according to the first preset acquisition frequency F and transmits the data to the third thread. The first preset acquisition frequency F is adopted, and the acquired data redundancy and the data volume are considered.
Step 003, the third program obtains at least one motion data C of the said mobile terminal according to the data received1(n)、C2(n)、…、Cm(n) of (a). Specifically, in the present invention, the main thread, the first thread and the second thread are all threads in the mobile terminal, and the first thread and the second thread are dedicated threads set differently from the main thread. According to the invention, the third thread, based on the data received, may perform a data pre-processing to obtain said at least one motion data C1(n)、C2(n)、…、Cm(n), exemplarily, when the main thread of the mobile terminal collects data of the acceleration sensor in three different directions of the coordinate system of the mobile terminal, the motion data C1(n) collected data of the acceleration sensor on the X axis of the mobile terminal coordinate system, and motion data C2(n) is the collected data of the acceleration sensor on the Y axis of the coordinate system of the mobile terminal, and the motion data C3And (n) is the acquired data of the acceleration sensor on the Z axis of the mobile terminal coordinate system. Further, the preprocessing may further include data coordinate space conversion and the like, and the data coordinate space conversion and the like may all adopt the related prior art in the field, which is not described herein again.
According to the invention, the operating system of the mobile terminal or the SDK built in the mobile terminal APP transmits the at least one motion data C according to the preset uploading interval1(n)、C2(n)、…、Cm(n) uploading to the server. The value of the preset uploading interval can be set in a user-defined mode, and in one embodiment, the value of the preset uploading interval is set in a user-defined modeThe transmission interval has a value range of [5 minutes, 15 minutes ]]。
In a mobile terminal, a main thread often acquires a value of a target object (for example, a sensor) through a unique data callback interface in the mobile terminal, and since more than one target object is usually required to acquire the value in the same time period of the mobile terminal, even if the acquisition frequency of the main thread for acquiring the value of a specified target object is set, time intervals between target object data often acquired are different due to the unique data callback interface calling mode, sometimes, the interval between two adjacent times of value acquisition is very large, and in order to eliminate the larger interval between two adjacent times of value acquisition, in the prior art, interpolation is usually directly performed between the two to reduce noise influence, but the inaccurate data interpolation also causes certain influence on gait recognition. The invention ensures that the interval between the data acquired by the second thread is fixed interval by resampling the data acquired by the main thread with larger data sampling frequency by the second thread with smaller data sampling frequency, further ensures that the interval between the acquired motion data is fixed, reduces the adverse effect caused by data interpolation, and further ensures that the complete motion data of the terminal user in one gait cycle can be acquired by the reading content of the second thread because the data updating speed of the main thread is higher than the reading speed of the second thread.
Moreover, other implementations of the invention will be apparent to those skilled in the art from consideration of the specification of the invention disclosed herein. The embodiments and/or aspects of the embodiments can be used in the systems and methods of the present invention alone or in any combination. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.

Claims (10)

1. A method for recognizing gait of an end user based on mobile terminal data is characterized by comprising the following steps:
step 100, obtaining at least one motion data C of the mobile terminal1(n)、C2(n)、…、Cm(n) wherein the x-th motion data Cx(n) is a time period T ═ T1,t2]The mobile terminal collects the signals according to a first preset collection frequency F, wherein x is more than or equal to 1 and less than or equal to m;
step 200, respectively setting C at a preset sliding length delta W by using a rectangular window Rt with a window length W1(n)、C2(n)、…、Cm(n) sliding up to obtain first step profile D of said end-user0,D1,…,DuWherein D isiComprising rectangular windows Rt respectively at C1(n)、C2(n)、…、Cm(n) i is 0. ltoreq. i.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.W.ltoreq.u.ltoreq.<W，
Figure FDA0002336738250000011
I.e. the ratio of u to Δ W is rounded down;
step 300, based on the first step characteristic data D0,D1,…,DuAnd identifying the gait of the terminal user by a user gait identification model;
further, W is larger than or equal to the standard gait cycle length, and the standard gait cycle length is the maximum value of a plurality of known gait cycle lengths.
2. The method of identifying an end user gait according to claim 1, characterized in that said step 300 comprises:
step 301, based on the first step state feature data D0,D1,…,DuAcquiring second-step state feature data VD of the terminal user by the gait abnormal model of the terminal user1,VD2,…,VDsWherein, the gait abnormal model of the terminal user is used for judging the first step characteristic data D of the terminal useriWhether the data is normal data in a target traveling mode or not, and the second-step state feature data VD1,VD2,…,VDsThe data are normal data in the target advancing mode, and s is less than or equal to u + 1;
step 302, the second stepState characteristic data VD1,VD2,…,VDsAnd as input data of a user gait recognition model, recognizing the gait of the terminal user by using the user gait recognition model.
3. The method of identifying an end user gait as claimed in claim 2, wherein step 301 comprises:
step 401, the first step state feature data D is processed0,D1,…,DuInputting the gait abnormal model of the end user to obtain the first step state feature data D0,D1,…,DuIs abnormal identification information L0,L1,…,LuWherein L isiIs DiThe abnormality identification information of (1);
step 402, based on D0,D1,…,DuAnd L0,L1,…,LuAcquiring second-step state feature data VD of the terminal user1,VD2,…,VDs。
4. A method of identifying an end user gait according to any of claims 1 to 3, characterized in that the window length W is in the range [4 seconds, 10 seconds ].
5. A method of identifying an end user gait according to any of claims 1 to 3, characterized in that the preset sliding length aw ranges from [2 seconds, 5 seconds ].
6. A method of identifying an end user gait according to any of the claims 1-3, characterized in that the first preset acquisition frequency F has a value in the range [80,150 ].
7. A method of identifying end user gait according to any of claims 1 to 3, characterized in that t is t2-t1The value range is [10 seconds, 70 seconds]。
8. According to the claimsMethod for identifying the gait of an end user according to claim 1, characterized in that said at least one movement datum C1(n)、C2(n)、…、Cm(n) obtained by the third thread of the mobile terminal by receiving and processing data collected by the second thread of the mobile terminal according to the first preset collection frequency F from a first array set so as to store the main thread of the mobile terminal for a period of time P ═ P1,p2]The data of at least one target object in the mobile terminal are acquired according to a second preset acquisition frequency E, the array size of the first array is the total number of data acquired by the at least one target object every time the main thread of the mobile terminal polls, and the at least one target object is related to the motion of the mobile terminal.
9. A server for identifying an end user's gait based on mobile terminal data, comprising a processor and a non-transitory computer readable storage medium storing a computer program which, when executed by the processor, implements a method for identifying an end user's gait based on mobile terminal data as claimed in any one of claims 1-8.
10. A storage medium storing a computer program which, when executed, implements a method of identifying an end user's gait based on mobile terminal data according to any one of claims 1 to 8.
CN201911359210.9A 2019-12-25 2019-12-25 Method and server for identifying gait of terminal user based on mobile terminal data Active CN111126294B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911359210.9A CN111126294B (en) 2019-12-25 2019-12-25 Method and server for identifying gait of terminal user based on mobile terminal data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911359210.9A CN111126294B (en) 2019-12-25 2019-12-25 Method and server for identifying gait of terminal user based on mobile terminal data

Publications (2)

Publication Number Publication Date
CN111126294A true CN111126294A (en) 2020-05-08
CN111126294B CN111126294B (en) 2023-06-16

Family

ID=70502373

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911359210.9A Active CN111126294B (en) 2019-12-25 2019-12-25 Method and server for identifying gait of terminal user based on mobile terminal data

Country Status (1)

Country Link
CN (1) CN111126294B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112766057A (en) * 2020-12-30 2021-05-07 浙江大学 Complex scene-oriented fine-grained attribute-driven gait data set forming method

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107122641A (en) * 2017-04-25 2017-09-01 杭州安石信息技术有限公司 Smart machine owner recognition methods and owner's identifying device based on use habit
CN107831907A (en) * 2017-12-07 2018-03-23 深圳先进技术研究院 Identity identifying method and device based on Gait Recognition
CN110209560A (en) * 2019-05-09 2019-09-06 北京百度网讯科技有限公司 Data exception detection method and detection device
CN110399032A (en) * 2019-07-09 2019-11-01 北京卡路里信息技术有限公司 The control method and device of wearable device

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107122641A (en) * 2017-04-25 2017-09-01 杭州安石信息技术有限公司 Smart machine owner recognition methods and owner's identifying device based on use habit
CN107831907A (en) * 2017-12-07 2018-03-23 深圳先进技术研究院 Identity identifying method and device based on Gait Recognition
CN110209560A (en) * 2019-05-09 2019-09-06 北京百度网讯科技有限公司 Data exception detection method and detection device
CN110399032A (en) * 2019-07-09 2019-11-01 北京卡路里信息技术有限公司 The control method and device of wearable device

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112766057A (en) * 2020-12-30 2021-05-07 浙江大学 Complex scene-oriented fine-grained attribute-driven gait data set forming method
CN112766057B (en) * 2020-12-30 2022-05-13 浙江大学 Complex scene-oriented fine-grained attribute-driven gait data set forming method

Also Published As

Publication number Publication date
CN111126294B (en) 2023-06-16

Similar Documents

Publication Publication Date Title
CN111401265B (en) Pedestrian re-identification method and device, electronic equipment and computer-readable storage medium
CN107784282B (en) Object attribute identification method, device and system
CN110458061B (en) Method for identifying old people falling down and accompanying robot
WO2017024963A1 (en) Image recognition method, measure learning method and image source recognition method and device
JP6332937B2 (en) Image processing apparatus, image processing method, and program
CN106648078B (en) Multi-mode interaction method and system applied to intelligent robot
CN111401339B (en) Method and device for identifying age of person in face image and electronic equipment
CN110807491A (en) License plate image definition model training method, definition detection method and device
JP2022540101A (en) POSITIONING METHOD AND APPARATUS, ELECTRONIC DEVICE, COMPUTER-READABLE STORAGE MEDIUM
CN112101156B (en) Target identification method and device and electronic equipment
CN105590020A (en) Improved data comparison method
CN118094118B (en) Data set quality evaluation method, system, electronic equipment and storage medium
CN110516638B (en) Sign language recognition method based on track and random forest
CN110532895B (en) Method, device and equipment for detecting fraudulent behavior in face recognition process
CN111126294A (en) Method and server for recognizing gait of terminal user based on mobile terminal data
CN107103289B (en) Method and system for handwriting identification by using handwriting outline characteristics
CN116793364A (en) Indoor positioning system and method based on dead reckoning
CN111142688B (en) Method and server for identifying terminal user machine changing based on mobile terminal data
CN110990819B (en) Method and server for acquiring gait feature data of terminal user based on mobile terminal data
CN111062352B (en) Method and server for identifying gait of terminal user based on mobile terminal data
CN111062353B (en) Method and server for acquiring gait feature data of terminal user based on mobile terminal data
CN111061376B (en) Method and server for identifying terminal user machine changing based on mobile terminal data
CN115393755A (en) Visual target tracking method, device, equipment and storage medium
CN115205589A (en) Object identification method and device, electronic equipment and storage medium
CN109993312A (en) A kind of equipment and its information processing method, computer storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
CB02 Change of applicant information

Address after: 310012 Room 418, West District, Building A, 525 Xixi Road, Xihu District, Hangzhou City, Zhejiang Province

Applicant after: Daily interactive Co.,Ltd.

Address before: 310012 Room 418, West District, Building A, 525 Xixi Road, Xihu District, Hangzhou City, Zhejiang Province

Applicant before: ZHEJIANG MEIRI INTERDYNAMIC NETWORK TECHNOLOGY Co.,Ltd.

CB02 Change of applicant information
GR01 Patent grant
GR01 Patent grant