CN109492595B - Behavior prediction method and system suitable for fixed group - Google Patents

Behavior prediction method and system suitable for fixed group Download PDF

Info

Publication number
CN109492595B
CN109492595B CN201811374961.3A CN201811374961A CN109492595B CN 109492595 B CN109492595 B CN 109492595B CN 201811374961 A CN201811374961 A CN 201811374961A CN 109492595 B CN109492595 B CN 109492595B
Authority
CN
China
Prior art keywords
behavior
data
information
abnormal
behavior data
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.)
Active
Application number
CN201811374961.3A
Other languages
Chinese (zh)
Other versions
CN109492595A (en
Inventor
陈阳
卓荣庆
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 University of Media and Communications
Original Assignee
Zhejiang University of Media and Communications
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 University of Media and Communications filed Critical Zhejiang University of Media and Communications
Priority to CN201811374961.3A priority Critical patent/CN109492595B/en
Publication of CN109492595A publication Critical patent/CN109492595A/en
Application granted granted Critical
Publication of CN109492595B publication Critical patent/CN109492595B/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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
    • G06V40/16—Human faces, e.g. facial parts, sketches or expressions
    • G06V40/172—Classification, e.g. identification

Abstract

The embodiment of the invention discloses a behavior prediction method and a behavior prediction system suitable for a fixed group, wherein the method comprises the following steps: matching the current behavior data of the behavior object acquired by the behavior acquisition device with the behavior data in the behavior database to obtain the identity information of the behavior object; matching facial features of facial image data of the behavior object acquired by the image acquisition device based on a preset image database, and determining the corresponding relation between the identity information and the behavior object according to the result of the facial feature matching; and when the corresponding relation indicates that the behavior object is matched with the identity information, performing behavior prediction on the acquired real-time behavior data of the behavior object by adopting a machine learning model to obtain the behavior data of the behavior object at the next moment. By adopting the method and the device, through the double identity authentication of the behavior object, when the abnormal behavior of the behavior object is predicted to possibly occur, the efficiency of timely eliminating the abnormal behavior by related personnel can be improved.

Description

Behavior prediction method and system suitable for fixed group
Technical Field
The invention relates to the technical field of internet, in particular to a behavior prediction method and system suitable for a fixed group.
Background
With the rapid development of internet technology, human behavior is no longer unpredictable at will. The scientific prediction of human behaviors has important commercial value and practical significance.
In the prior art, when an abnormal behavior object is determined, a human behavior analysis and prediction system reminds related personnel of attention by sending alarm information, for example, by analyzing the current behavior of a person handling business in a bank and predicting the next action of the person, and when a customer about to take an abnormal behavior (for example, robbing the bank) is predicted, security personnel are reminded of paying close attention, so that the safety of social personnel and property is guaranteed. However, when there are many monitored behavior objects within a certain range, the related personnel need to determine the target of the impending abnormal behavior through human screening, which reduces the efficiency of eliminating the abnormal behavior.
Disclosure of Invention
The embodiment of the invention provides a behavior prediction method and a behavior prediction system suitable for a fixed group, which can improve the efficiency of relevant personnel for timely eliminating abnormal behaviors when the abnormal behaviors possibly appear in behavior objects are predicted through dual identity authentication of the behavior objects.
A first aspect of an embodiment of the present invention provides a behavior prediction method applicable to a fixed group, where the method includes:
matching the current behavior data of the behavior object acquired by the behavior acquisition device with the behavior data in the behavior database to obtain the identity information of the behavior object;
matching facial features of facial image data of the behavior object acquired by the image acquisition device based on a preset image database, and determining the corresponding relation between the identity information and the behavior object according to the result of the facial feature matching;
and when the corresponding relation indicates that the behavior object is matched with the identity information, performing behavior prediction on the acquired real-time behavior data of the behavior object by adopting a machine learning model to obtain the behavior data of the behavior object at the next moment.
In an alternative embodiment, the method further comprises:
and acquiring behavior data of each member in the fixed group based on a behavior acquisition device to form a behavior database.
In an alternative embodiment, the method further comprises:
and training a machine learning model aiming at the behavior database by adopting a preset learning algorithm.
In an alternative embodiment, the method further comprises:
and determining the identity information of the behavior object according to the identity identification input by the behavior object.
In an alternative embodiment, the behavior acquisition apparatus includes at least two behavior acquisition devices, and the at least two behavior acquisition devices acquire current behavior data and/or real-time behavior data of the behavior object from at least two angles.
In an alternative embodiment, the method further comprises:
acquiring at least two real-time angle behavior data of behavior objects acquired by at least two behavior acquisition devices;
and performing fusion analysis on the at least two real-time angle behavior data by adopting a data fusion algorithm to generate real-time behavior data of the behavior object.
In an alternative embodiment, the method further comprises:
judging whether the predicted behavior data of the behavior object at the next moment belongs to abnormal behavior data or not based on the object information of the behavior object, wherein the object information indicates the object attribute of the behavior object;
and outputting abnormal alarm information when the judgment result is yes, wherein the abnormal alarm information carries the identity information of the behavior object.
In an alternative embodiment, the method further comprises:
and when the behavior data of at least two behavior objects in the same area at the next moment is predicted to be abnormal behavior data, outputting emergency alarm information, wherein the emergency alarm information carries the identity information of the at least two behavior objects.
A second aspect of an embodiment of the present invention provides a behavior prediction system suitable for a fixed group, which may include:
the information acquisition module is used for matching the current behavior data of the behavior object acquired by the behavior acquisition device with the behavior data in the behavior database to obtain the identity information of the behavior object;
the relationship determination module is used for matching facial features of the facial image data of the behavior object acquired by the image acquisition device based on a preset image database and determining the corresponding relationship between the identity information and the behavior object according to the result of the facial feature matching;
and the behavior prediction module is used for performing behavior prediction on the acquired real-time behavior data of the behavior object by adopting a machine learning model when the corresponding relation indicates that the behavior object is matched with the identity information to obtain the behavior data of the behavior object at the next moment.
In an alternative embodiment, the system further comprises:
and the database acquisition module is used for acquiring behavior data of each member in the fixed group based on the behavior acquisition device to form a behavior database.
In an alternative embodiment, the system further comprises:
and the model training module is used for training a machine learning model aiming at the behavior database by adopting a preset learning algorithm.
In an alternative embodiment, the system further comprises:
and the information determining module is used for determining the identity information of the behavior object according to the identity identification input by the behavior object.
In an alternative embodiment, the behavior acquisition apparatus includes at least two behavior acquisition devices, and the at least two behavior acquisition devices acquire current behavior data and/or real-time behavior data of the behavior object from at least two angles.
In an alternative embodiment, the system further comprises: the angle data acquisition module is used for acquiring at least two real-time angle behavior data of the behavior object acquired by at least two behavior acquisition devices;
and the data generation module is used for performing fusion analysis on the at least two real-time angle behavior data by adopting a data fusion algorithm to generate the real-time behavior data of the behavior object.
In an alternative embodiment, the system further comprises:
the abnormal behavior judging module is used for judging whether the predicted behavior data of the behavior object at the next moment belongs to the abnormal behavior data or not based on the object information of the behavior object, and the object information indicates the object attribute of the behavior object;
and the alarm information output module is used for outputting abnormal alarm information when the judgment result is yes, wherein the abnormal alarm information carries the identity information of the behavior object.
In an alternative embodiment, the system further comprises:
and the emergency information output module is used for outputting emergency alarm information when the behavior data of at least two behavior objects in the same region at the next moment is predicted to be abnormal behavior data, wherein the emergency alarm information carries the identity information of the at least two behavior objects.
In the embodiment of the invention, the identity information of the behavior object is preliminarily analyzed through the current behavior data of the behavior object, the corresponding relation between the behavior object and the identity information is confirmed again through the matching of the face image, and the behavior data of the behavior object at the next moment is predicted on the basis of ensuring the corresponding relation. When abnormal behaviors occur, the identities of the behavior objects with the abnormal behaviors are accurately output according to the corresponding relation between the behavior objects and the identity information, and the efficiency of relevant personnel for timely eliminating the abnormal behaviors is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for persons of ordinary skill in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a schematic flow chart of a behavior prediction method for a fixed group according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of another behavior prediction system suitable for a fixed group according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The behavior prediction method suitable for the fixed group provided by the embodiment of the invention can be applied to an application scene for predicting behaviors of various behavioral objects in the fixed group (for example, behavior prediction of the old in a nursing home or behavior prediction of examinees in an unmanned examination room).
The behavior prediction system (hereinafter, simply referred to as a behavior prediction system) applicable to a fixed group according to an embodiment of the present invention may include a behavior acquisition device, a plurality of image acquisition devices, a card punch, an intelligent terminal (terminal devices such as a tablet computer, a Personal Computer (PC), a smart phone, a palmtop computer, and a Mobile Internet Device (MID)), a data processing server, and the like, where the image acquisition device may also serve as the behavior acquisition device.
The following describes a behavior prediction method applicable to a fixed group according to an embodiment of the present invention in detail with reference to fig. 1.
Referring to fig. 1, a flow chart of a behavior prediction method suitable for a fixed group is provided for the embodiment of the present invention. As shown in fig. 1, the method of an embodiment of the present invention may include the following steps S101 to S103.
And S101, matching the current behavior data of the behavior object acquired by the behavior acquisition device with the behavior data in the behavior database to obtain the identity information of the behavior object.
It should be noted that the behavior prediction system may form a behavior database by collecting behavior data of each member in the fixed group based on the behavior collection device, the behavior collection device may be a device that collects behavior data of a behavior object when the behavior object enters a predictable behavior region, and may be a Kinect sensor, an infrared sensor, a camera, or the like, and the predictable behavior region may be a range region of activities of the fixed group, for example, a school, a classroom, an old care home, or the like. For example, when a student walks into a class, the Kinect sensor extracts the student's skeletal node data at the classroom doorway, which may indicate behavioral characteristics of the student, such as characteristics of pace; or by a camera taking an image of the person as he enters the classroom, which image may indicate the physical characteristics of the person (tall, short, fat, thin, or other characteristics). The behavior database may be a database composed of behavior data of all members in a fixed group, for example, bone node data or images of all students in a class.
In an alternative embodiment, the behavior acquisition apparatus includes at least two behavior acquisition devices, and the at least two behavior acquisition devices can acquire at least two current angle behavior data of the behavior object from at least two angles. The behavior prediction system can adopt a data fusion algorithm to perform fusion analysis on at least two current angle behavior data to generate current behavior data of the behavior object. Through multi-angle collection and fusion analysis, the accuracy of the acquired current behavior data can be ensured.
Specifically, the behavior prediction system may match the current behavior data of the behavior object with the behavior data in the behavior database through the data processing server to obtain the identity information of the behavior object. It can be understood that, when the behavior database is generated, the identity information of the behavior object is also stored correspondingly. The identity information may be information uniquely identifying the identity of the behavioral object, and may be, for example, a name, an academic number, a class, an identification number, or other numbers.
In an alternative embodiment, the behavior prediction system may determine the identity information of the behavior object according to the identity input by the behavior object. The identity may be a symbolic item representing the identity of the behavioral object, such as a student card or an elderly card.
S102, matching facial features of facial image data of the behavior object acquired by the image acquisition device based on a preset image database, and determining the corresponding relation between the identity information and the behavior object according to the result of the facial feature matching.
It is understood that the image capturing device may capture facial image data of each member in the fixed group in advance, and store the facial image data in a preset image database. The image capture device may capture facial image data of the behavioral object when the behavioral object is active within a predictable behavioral region.
Specifically, the behavior prediction system may perform facial feature matching on the facial image data with facial image data in a preset image database through the data processing server.
Further, the behavior prediction system may determine a correspondence between the identity information and the behavior object according to the result of the facial feature matching, and the correspondence may indicate whether the behavior object is actually a behavior object corresponding to the identity information.
In the embodiment of the invention, the identity of the behavior object is confirmed again through the matching of the facial image data, so that the condition of identity error of the behavior object which is confirmed for the first time due to the existence of similar posture characteristics or actions of swiping a card by others instead can be avoided, and the action of confirming the identity of the behavior object for the first time refers to the process of completing identity confirmation by matching the current behavior data with the behavior data in the behavior database.
S103, when the corresponding relation indicates that the behavior object is matched with the identity information, performing behavior prediction on the acquired real-time behavior data of the behavior object by adopting a machine learning model to obtain the behavior data of the behavior object at the next moment.
It is understood that when the correspondence indicates that the behavior object matches the identity information, the behavior object indicated by the identity information may be considered as the behavior object corresponding to the correspondence, for example, after matching facial features, it is determined that a student with a fat height 170 is a young king rather than a young king with a similar posture.
It should be noted that the behavior prediction system may use a preset learning algorithm to train a machine learning model for the behavior database, where the preset learning algorithm may be a support vector machine, a decision tree, a bayesian classification, or a deep neural network algorithm, and the specific algorithm selection may be determined according to the requirements of developers. The machine learning model can predict the behavior data of the behavior object at the next moment by analyzing the current behavior data of the behavior object.
Specifically, the behavior prediction system may perform behavior prediction on the acquired real-time behavior data of the behavior object by using a machine learning model, so as to obtain the behavior data of the behavior object at the next moment. It is to be understood that the real-time behavior data may be real-time behavior data of a behavior object acquired by the behavior acquisition device within a predictable behavior region.
In an alternative embodiment, the behavior acquisition apparatus includes at least two behavior acquisition devices, and the at least two behavior acquisition devices can acquire at least two real-time angle behavior data of the behavior object from at least two angles. The behavior prediction system can adopt a data fusion algorithm to perform fusion analysis on the at least two real-time angle behavior data to generate the real-time behavior data of the behavior object. Through multi-angle collection and fusion analysis, the accuracy of the acquired real-time behavior data can be ensured.
In an optional embodiment, the behavior prediction system may determine whether the predicted behavior data of the behavior object at the next time belongs to the abnormal behavior data based on the object information of the behavior object, and the object information may be stored in the behavior database in correspondence with the identity information or may be stored separately in correspondence with the identity information. The object information may include attribute information of the behavior object, and the attribute information may be information that can be a distinctive feature point of the behavior object, such as a physical attribute or an identity attribute of the behavior object. (e.g., information on diseases, information on physical health, information on the quality of learning, etc.). The abnormal behavior data may indicate that a behavior object may have a bad situation when the behavior exists, for example, a student in an examination room who learns poorly may have eastern and western behaviors, indicating that the student may want to cheat (behavior at the next moment), and an old person in a nursing home with heart disease may have an action of touching the chest, indicating that the old person may possibly have a disease (behavior at the next moment). Whether abnormal behaviors occur in the predicted behavior object is determined through the attribute information of the behavior object, and the reliability of confirming the abnormal behaviors of the behavior object is guaranteed.
Further, when the behavior prediction system determines that abnormal behavior data will occur in the behavior object, abnormal alarm information may be output. Alternatively, the behavior prediction system may output the abnormal alert information to a client (e.g., a teacher's computer). The abnormal alarm information carries the identity information of the behavior object, and therefore a guardian or a invigilator can conveniently and accurately position the behavior object with the abnormal behavior.
In an optional embodiment, when the behavior prediction system predicts that the behavior data of the behavior object at the next time is abnormal behavior data, the time and/or the place where the behavior object is located may be obtained, and the abnormal behavior data is distinguished and corrected according to the time and/or the place. The above-described discrimination correction is to further confirm whether or not the data predicted and determined as abnormal behavior data is abnormal. It is understood that the behavior data of the behavior object at the next moment may be considered as abnormal only when the behavior object is at a specific time and/or place. For example, when a current behavior of which attention is not specific appears in a classroom and a test time of the student a, the behavior data of the student a at the next moment can be predicted to be abnormal behavior data (e.g., cheating); however, when the student a performs the above-described behavior in the classroom during the next class, the behavior data of the student a at the next time cannot be predicted as abnormal behavior data.
In an optional embodiment, when the abnormal behavior data of the behavior object is predicted, the behavior prediction system may obtain facial expression data of the behavior object, and determine whether the predicted abnormal behavior data is really abnormal according to the facial expression data, for example, the old a has an action of touching the front chest with a hand, but the facial expression is light or has no abnormality, and may consider that the old a scratches or arranges clothes, and is not a heart problem, and may not output abnormal warning information.
In an optional embodiment, when the abnormal behavior data of the behavior object is predicted, the behavior prediction system may obtain the current behavior data of the monitoring object around the behavior object, and determine whether the abnormal alarm information needs to be output or the level of the abnormal alarm information needs to be output according to the current behavior data of the monitoring object, where the current behavior data of the monitoring object carries the identity information of the monitoring object. For example, if the behavior prediction system acquires that there is a doctor around the elderly person a and the doctor is taking a rescue measure, the system does not need to output abnormal warning information, or if the system acquires that there is another person around the elderly person a other than the doctor (e.g., a nursing home attendant) can output lower-level danger warning information to a person monitoring the system.
By performing anomaly confirmation on the predicted behavior data that has been confirmed to be anomalous, the accuracy of performing anomaly confirmation on the predicted behavior data is increased.
In an alternative embodiment, the behavior prediction system may select a level of the output abnormal warning information according to the reconfirmed abnormal result, where the abnormal warning level may be classified into at least two levels of warning ratios, optionally, the first level warning information may be output when both reconfirmed results indicate that the predicted behavior data is abnormal, and the second level warning information may be output when only one reconfirmed result indicates that the predicted behavior data is abnormal.
In an optional embodiment, when it is predicted that behavior data of at least two behavior objects at the next moment in the same area is abnormal behavior data, the behavior prediction system may output emergency alert information, which may be first-level alert information. It is understood that the emergency alert system may represent the level of the above-mentioned abnormal alert information, may be the abnormal alert information of the highest level, and may carry the identity information of at least two behavior objects. For example, in an examination room without invigilation, when the behavior prediction system predicts that abnormal behaviors such as cheating may occur in the behaviors of a plurality of students at the next moment, the behavior prediction system may output the abnormal alarm information at the highest level, that is, the emergency alarm information, to prompt the invigilator to move to the examination room for processing at the first time.
In the embodiment of the invention, the identity information of the behavior object is preliminarily analyzed through the current behavior data of the behavior object, the corresponding relation between the behavior object and the identity information is confirmed again through the matching of the face image, and the behavior data of the behavior object at the next moment is predicted on the basis of ensuring the corresponding relation. When abnormal behaviors occur, the identities of the behavior objects with the abnormal behaviors are accurately output according to the corresponding relation between the behavior objects and the identity information, and the efficiency of relevant personnel for timely eliminating the abnormal behaviors is improved.
In an optional embodiment, the behavior prediction system may also perform multiple authentications on the identity of the behavior object by using other devices, so as to further ensure the accuracy of the identity information of the confirmed behavior object.
A behavior prediction system suitable for a fixed group according to an embodiment of the present invention will be described in detail below with reference to fig. 2. It should be noted that the behavior prediction system shown in fig. 2 is used for executing the method according to the embodiment of the present invention shown in fig. 1, and for convenience of description, only the portion related to the embodiment of the present invention is shown, and details of the specific technology are not disclosed, please refer to the embodiment of the present invention shown in fig. 1.
Referring to fig. 2, a schematic structural diagram of a behavior prediction system according to an embodiment of the present invention is provided. As shown in fig. 2, the behavior prediction system 100 of the embodiment of the present invention may include: the system comprises an information acquisition module 101, a relation determination module 102, a behavior prediction module 103, a database acquisition module 104, a model training module 105, an information determination module 106, an angle data acquisition module 107, a data generation module 108, an abnormal behavior judgment module 109, an alarm information output module 110 and an emergency information output module 111.
The information obtaining module 101 is configured to match the current behavior data of the behavior object obtained by the behavior collecting device with the behavior data in the behavior database to obtain the identity information of the behavior object.
Before the execution information obtaining module 101 obtains the identity information of the behavior object, the database collecting module 104 may collect behavior data of each member in the fixed group to form a behavior database based on the behavior collecting device, where the behavior collecting device may be a device that collects behavior data of the behavior object when the behavior object enters a predictable behavior region, and may be a Kinect sensor, an infrared sensor, or a camera, and the predictable behavior region may be a range region of activities of the fixed group, for example, a school, a classroom, an old care home, and the like. For example, when a student walks into a class, the Kinect sensor extracts the student's skeletal node data at the classroom doorway, which may indicate behavioral characteristics of the student, such as characteristics of pace; or by a camera taking an image of the person as he enters the classroom, which image may indicate the physical characteristics of the person (tall, short, fat, thin, or other characteristics). The behavior database may be a database composed of behavior data of all members in a fixed group, for example, bone node data or images of all students in a class.
In an alternative embodiment, the behavior acquisition apparatus includes at least two behavior acquisition devices, and the at least two behavior acquisition devices can acquire at least two current angle behavior data of the behavior object from at least two angles. The behavior prediction system can adopt a data fusion algorithm to perform fusion analysis on at least two current angle behavior data to generate current behavior data of the behavior object. Through multi-angle collection and fusion analysis, the accuracy of the acquired current behavior data can be ensured.
In a specific implementation, the information obtaining module 101 may match the current behavior data of the behavior object with the behavior data in the behavior database through the data processing server to obtain the identity information of the behavior object. It is understood that when the database collection module 104 generates the behavior database, the identity information of the behavior object is also stored correspondingly. The identity information may be information uniquely identifying the identity of the behavioral object, and may be, for example, a name, an academic number, a class, an identification number, or other numbers.
In an alternative embodiment, the information determining module 106 may determine the identity information of the behavior object according to the identity identifier input by the behavior object. The identity may be a symbolic item representing the identity of the behavioral object, such as a student card or an elderly card.
The relationship determining module 102 is configured to perform facial feature matching on facial image data of the behavior object acquired by the image acquisition device based on a preset image database, and determine a corresponding relationship between the identity information and the behavior object according to a result of the facial feature matching.
It is understood that the image capturing device may capture facial image data of each member in the fixed group in advance, and store the facial image data in a preset image database. The image capture device may capture facial image data of the behavioral object when the behavioral object is active within a predictable behavioral region.
In a specific implementation, the relationship determining module 102 may perform facial feature matching on the facial image data and facial image data in a preset image database through a data processing server.
Further, the relationship determination module 102 may determine a correspondence between the identity information and the behavior object according to the result of the facial feature matching, and the correspondence may indicate whether the behavior object is really a behavior object corresponding to the identity information.
In the embodiment of the invention, the identity of the behavior object is confirmed again through the matching of the facial image data, so that the condition of identity error of the behavior object which is confirmed for the first time due to the existence of similar posture characteristics or actions of swiping a card by others instead can be avoided, and the action of confirming the identity of the behavior object for the first time refers to the process of completing identity confirmation by matching the current behavior data with the behavior data in the behavior database.
And the behavior prediction module 103 is configured to perform behavior prediction on the obtained real-time behavior data of the behavior object by using a machine learning model when the correspondence indicates that the behavior object matches the identity information, so as to obtain behavior data of the behavior object at the next moment.
It is understood that when the correspondence indicates that the behavior object matches the identity information, the behavior object indicated by the identity information may be considered as the behavior object corresponding to the correspondence, for example, after matching facial features, it is determined that a student with a fat height 170 is a young king rather than a young king with a similar posture.
It should be noted that the model training module 105 may train a machine learning model for the behavior database by using a preset learning algorithm, where the preset learning algorithm may be a support vector machine, a decision tree, a bayesian classification, or a deep neural network algorithm, and a specific algorithm selection may be determined according to a requirement of a developer. The machine learning model can predict the behavior data of the behavior object at the next moment by analyzing the current behavior data of the behavior object.
In a specific implementation, the behavior prediction module 103 may perform behavior prediction on the acquired real-time behavior data of the behavior object by using a machine learning model, so as to obtain the behavior data of the behavior object at the next time. It is to be understood that the real-time behavior data may be real-time behavior data of a behavior object acquired by the behavior acquisition device within a predictable behavior region.
In an optional embodiment, the behavior acquisition apparatus includes at least two behavior acquisition devices, the at least two behavior acquisition devices may acquire at least two real-time angle behavior data of the behavior object from at least two angles, and the angle data acquisition module 107 may acquire the at least two real-time angle behavior data. The data generation module can adopt a data fusion algorithm to perform fusion analysis on the at least two real-time angle behavior data to generate the real-time behavior data of the behavior object. Through multi-angle collection and fusion analysis, the accuracy of the acquired real-time behavior data can be ensured.
In an optional embodiment, the abnormal behavior determining module 109 may determine whether the predicted behavior data of the behavior object at the next time belongs to the abnormal behavior data based on the object information of the behavior object, where the object information may be stored in the behavior database in correspondence with the identity information or may be stored separately in correspondence with the identity information. The object information may include attribute information of the behavior object, and the attribute information may be information that can be a distinctive feature point of the behavior object, such as a physical attribute or an identity attribute of the behavior object. (e.g., information on diseases, information on physical health, information on the quality of learning, etc.). The abnormal behavior data may indicate that a behavior object may have a bad situation when the behavior exists, for example, a student in an examination room who learns poorly may have eastern and western behaviors, indicating that the student may want to cheat (behavior at the next moment), and an old person in a nursing home with heart disease may have an action of touching the chest, indicating that the old person may possibly have a disease (behavior at the next moment). Whether abnormal behaviors occur in the predicted behavior object is determined through the attribute information of the behavior object, and the reliability of confirming the abnormal behaviors of the behavior object is guaranteed.
Further, when the behavior prediction system determines that abnormal behavior data will occur in the behavior object, the alarm information output module 110 may output abnormal alarm information. Alternatively, the alarm information output module 110 may output the abnormal alarm information to a client (e.g., a teacher's computer). The abnormal alarm information carries the identity information of the behavior object, and therefore a guardian or a invigilator can conveniently and accurately position the behavior object with the abnormal behavior.
In an alternative embodiment, when the behavior prediction system 100 predicts that the behavior data of the behavior object at the next time is abnormal behavior data, the time and/or the place where the behavior object is located may be obtained, and the abnormal behavior data may be determined and corrected according to the time and/or the place. The above-described discrimination correction is to further confirm whether or not the data predicted and determined as abnormal behavior data is abnormal. It is understood that the behavior data of the behavior object at the next moment may be considered as abnormal only when the behavior object is at a specific time and/or place. For example, when a current behavior of which attention is not specific appears in a classroom and a test time of the student a, the behavior data of the student a at the next moment can be predicted to be abnormal behavior data (e.g., cheating); however, when the student a performs the above-described behavior in the classroom during the next class, the behavior data of the student a at the next time cannot be predicted as abnormal behavior data.
In an alternative embodiment, when the abnormal behavior data of the behavior object is predicted, the behavior prediction system 100 may obtain facial expression data of the behavior object, and determine whether the predicted abnormal behavior data is really abnormal according to the facial expression data, for example, the elderly a has an action of touching the front chest with a hand, but the facial expression is relaxed or there is no abnormality, and may consider that the elderly a is scratching or tidying clothes, and is not a heart problem, and may not output abnormal warning information.
In an optional embodiment, when the abnormal behavior data of the behavior object is predicted, the behavior prediction system 100 may obtain the current behavior data of the monitoring object around the behavior object, and determine whether to output the abnormal alarm information or output the level of the abnormal alarm information according to the current behavior data of the monitoring object, where the current behavior data of the monitoring object carries the identity information of the monitoring object. For example, if the behavior prediction system acquires that there is a doctor around the elderly person a and the doctor is taking a rescue measure, the system does not need to output abnormal warning information, or if the system acquires that there is another person around the elderly person a other than the doctor (e.g., a nursing home attendant) can output lower-level danger warning information to a person monitoring the system.
By performing anomaly confirmation on the predicted behavior data that has been confirmed to be anomalous, the accuracy of performing anomaly confirmation on the predicted behavior data is increased.
In an alternative embodiment, the behavior prediction system 100 may select a level of the output abnormal warning information according to the reconfirmed abnormal result, where the abnormal warning level may be classified into at least two levels of warning ratios, optionally, output the first level warning information when both reconfirmed results indicate that the predicted behavior data is abnormal, and output the second level warning information when only one reconfirmed result indicates that the predicted behavior data is abnormal.
In an alternative embodiment, when it is predicted that behavior data of at least two behavior objects at the next moment in the same area is abnormal behavior data, the behavior prediction system 100 may output emergency alert information, which may be first-level alert information. It is understood that the emergency alert system may represent the level of the above-mentioned abnormal alert information, may be the abnormal alert information of the highest level, and may carry the identity information of at least two behavior objects. For example, in an examination room without invigilation, when the behavior prediction system predicts that abnormal behaviors such as cheating may occur in the behaviors of a plurality of students at the next moment, the behavior prediction system may output the abnormal alarm information at the highest level, that is, the emergency alarm information, to prompt the invigilator to move to the examination room for processing at the first time.
In the embodiment of the invention, the identity information of the behavior object is preliminarily analyzed through the current behavior data of the behavior object, the corresponding relation between the behavior object and the identity information is confirmed again through the matching of the face image, and the behavior data of the behavior object at the next moment is predicted on the basis of ensuring the corresponding relation. When abnormal behaviors occur, the identities of the behavior objects with the abnormal behaviors are accurately output according to the corresponding relation between the behavior objects and the identity information, and the efficiency of relevant personnel for timely eliminating the abnormal behaviors is improved.
In an alternative embodiment, the behavior prediction system 100 may also perform multiple authentications on the identity of the behavior object by using other devices, so as to further ensure the accuracy of the identity information of the behavior object.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium and can include processes of the embodiments of the methods described above when the computer program is executed. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.

Claims (9)

1. A method of behavioral prediction applicable to a fixed population, the method comprising:
matching the current behavior data of the behavior object acquired by the behavior acquisition device with the behavior data in a behavior database to obtain the identity information of the behavior object;
matching facial features of facial image data of the behavior object acquired by an image acquisition device based on a preset image database, and determining the corresponding relation between the identity information and the behavior object according to the result of the facial feature matching;
when the corresponding relation indicates that the behavior object is matched with the identity information, performing behavior prediction on the acquired real-time behavior data of the behavior object by adopting a machine learning model to obtain behavior data of the behavior object at the next moment;
judging whether the predicted behavior data of the behavior object at the next moment belongs to abnormal behavior data corresponding to the object information or not based on the object information of the behavior object, wherein the object information indicates the object attribute of the behavior object;
and outputting abnormal alarm information when the judgment result is yes, wherein the abnormal alarm information carries the identity information of the behavior object.
2. The method of claim 1, further comprising:
and acquiring behavior data of each member in the fixed group based on a behavior acquisition device to form a behavior database.
3. The method of claim 2, further comprising:
and training a machine learning model aiming at the behavior database by adopting a preset learning algorithm.
4. The method of claim 1, further comprising:
and determining the identity information of the behavior object according to the identity identification input by the behavior object.
5. The method according to claim 1, characterized in that the behavior acquisition arrangement comprises at least two behavior acquisition devices which acquire current behavior data and/or real-time behavior data of the behavior object from at least two angles.
6. The method of claim 5, further comprising:
acquiring at least two real-time angle behavior data of the behavior object acquired by at least two behavior acquisition devices;
and performing fusion analysis on the at least two real-time angle behavior data by adopting a data fusion algorithm to generate the real-time behavior data of the behavior object.
7. The method of claim 6, further comprising:
and when the behavior data of at least two behavior objects at the next moment in the same area is predicted to be abnormal behavior data, outputting emergency alarm information, wherein the emergency alarm information carries the identity information of the at least two behavior objects.
8. A behavior prediction system adapted for use in a fixed population, the system comprising:
the information acquisition module is used for matching the current behavior data of the behavior object acquired by the behavior acquisition device with the behavior data in the behavior database to obtain the identity information of the behavior object;
the relationship determination module is used for matching facial features of the facial image data of the behavior object acquired by the image acquisition device based on a preset image database and determining the corresponding relationship between the identity information and the behavior object according to the result of the facial feature matching;
the behavior prediction module is used for predicting the behavior of the acquired real-time behavior data of the behavior object by adopting a machine learning model when the corresponding relation indicates that the behavior object is matched with the identity information to obtain the behavior data of the behavior object at the next moment;
the abnormal behavior judging module is used for judging whether the predicted behavior data of the behavior object at the next moment belongs to the abnormal behavior data corresponding to the object information or not based on the object information of the behavior object, and the object information indicates the object attribute of the behavior object;
and outputting abnormal alarm information when the judgment result is yes, wherein the abnormal alarm information carries the identity information of the behavior object.
9. The system of claim 8, further comprising:
the database acquisition module is used for acquiring behavior data of each member in the fixed group to form a behavior database based on the behavior acquisition device;
and the model training module is used for training a machine learning model aiming at the behavior database by adopting a preset learning algorithm.
CN201811374961.3A 2018-11-19 2018-11-19 Behavior prediction method and system suitable for fixed group Active CN109492595B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811374961.3A CN109492595B (en) 2018-11-19 2018-11-19 Behavior prediction method and system suitable for fixed group

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811374961.3A CN109492595B (en) 2018-11-19 2018-11-19 Behavior prediction method and system suitable for fixed group

Publications (2)

Publication Number Publication Date
CN109492595A CN109492595A (en) 2019-03-19
CN109492595B true CN109492595B (en) 2020-10-16

Family

ID=65696833

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811374961.3A Active CN109492595B (en) 2018-11-19 2018-11-19 Behavior prediction method and system suitable for fixed group

Country Status (1)

Country Link
CN (1) CN109492595B (en)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110197171A (en) * 2019-06-06 2019-09-03 深圳市汇顶科技股份有限公司 Exchange method, device and the electronic equipment of action message based on user
CN110363220B (en) * 2019-06-11 2021-08-20 北京奇艺世纪科技有限公司 Behavior class detection method and device, electronic equipment and computer readable medium
CN110443408A (en) * 2019-07-04 2019-11-12 特斯联（北京）科技有限公司 Travel forecasting approaches and device
CN110334669B (en) * 2019-07-10 2021-06-08 深圳市华腾物联科技有限公司 Morphological feature recognition method and equipment
CN110647806B (en) * 2019-08-13 2022-05-03 浙江大华技术股份有限公司 Object behavior monitoring method, device, equipment, system and storage medium
CN110516599A (en) * 2019-08-27 2019-11-29 中国科学院自动化研究所 Group behavior identification model and its training method based on gradual relational learning
CN110974215B (en) * 2019-12-20 2022-06-03 首都医科大学宣武医院 Early warning system and method based on wireless electrocardiogram monitoring sensor group
CN111476202B (en) * 2020-04-30 2021-05-25 浙江申汇金融服务外包有限公司 User behavior analysis method and system of security system

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104463174A (en) * 2014-12-16 2015-03-25 广州南方电力集团科技发展有限公司 Multi-angle dynamic people recognition and behavior prediction system
US11734958B2 (en) * 2015-06-19 2023-08-22 eConnect, Inc. Predicting behavior from surveillance data
CN106778546A (en) * 2016-11-29 2017-05-31 聚鑫智能科技（武汉）股份有限公司 A kind of visual identity method and system based on visible ray and non-visible light
CN106778556A (en) * 2016-12-01 2017-05-31 大连因特视智能传感科技有限公司 Intelligent identity identification system based on dynamic human face identification and Gait Recognition fusion
CN107171872B (en) * 2017-07-19 2020-06-02 上海百芝龙网络科技有限公司 User behavior prediction method in smart home
CN108281203B (en) * 2018-02-28 2021-10-22 泰康保险集团股份有限公司 Method and device for predicting abnormal behavior

Also Published As

Publication number Publication date
CN109492595A (en) 2019-03-19

Similar Documents

Publication Publication Date Title
CN109492595B (en) Behavior prediction method and system suitable for fixed group
CN107704834B (en) Micro-surface examination assisting method, device and storage medium
US20230282090A1 (en) Utilizing artificial intelligence to detect objects or patient safety events in a patient room
US20190216333A1 (en) Thermal face image use for health estimation
US9202121B2 (en) Liveness detection
CN112328999B (en) Double-recording quality inspection method and device, server and storage medium
JP4401079B2 (en) Subject behavior analysis
Hernandez-de-Menendez et al. Biometric applications in education
US10610109B2 (en) Emotion representative image to derive health rating
CN111227789A (en) Human health monitoring method and device
Hasan et al. Pain level detection from facial image captured by smartphone
Lin et al. Looking at the body: Automatic analysis of body gestures and self-adaptors in psychological distress
US9830830B2 (en) Stimulus recognition training and detection methods
CN113017634B (en) Emotion evaluation method, emotion evaluation device, electronic device, and computer-readable storage medium
CN111222374A (en) Lie detection data processing method and device, computer equipment and storage medium
CN114764948A (en) Living body detection method, living body detection device, living body detection equipment and storage medium
Zamzami The Key Criteria for Predicting Unusual Behavior in the Elderly With Deep Learning Models Under 5G Technology
Sundharamurthy et al. Cloud‐based onboard prediction and diagnosis of diabetic retinopathy
CN111091901A (en) Old people first aid decision-making method, device, equipment and storage medium
Szankin et al. Fully Automated AI-powered Contactless Cough Detection based on Pixel Value Dynamics Occurring within Facial Regions
CN115459952A (en) Attack detection method, electronic device, and computer-readable storage medium
KR20230017454A (en) Method, Device and Computer Program For Preventing Cheating In Non-face-to-face Evaluation
KR20240046992A (en) Method for providing diagnosis and treatment of skin dermatitis based on machine-learning
Rey Assessing MABIS Mobile App Based on People at the Center of Mobile Application Development (PACMAD) Usability Model: Empirical Investigation
EP4256541A1 (en) Systems and methods for detection of subject activity by processing video and other signals using artificial intelligence

Legal Events

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