JP2021190051A - Behavior body identification system - Google Patents

Abstract

To provide a behavior body identification system capable of accurately identifying a behavior body reflected in a time series image, on the basis of a characteristic operation of the behavior body.SOLUTION: A behavior body identification system 1 comprises: a storage part 2 storing a standard operation model related to a standard operation of plural behavior bodies; a first operation acquisition part 4 acquiring an operation of a first behavior body Y1 reflected in plural first time series images X1; a registration part 5 for setting the acquired operation to a non-standard operation and associating the same with the first behavior body Y1 and registering the same, when the operation of the first behavior body Y1 is not included in the standard operation model; a second operation acquisition part 7 for acquiring an operation of the second behavior body Y2 reflected in plural second time series image X1; and an identification part 8 for identifying the second behavior body Y2 as the first behavior body Y1, when the operation of the second behavior body Y2 has prescribed coincidence or higher coincidence, to the non-standard operation.SELECTED DRAWING: Figure 3

Description

The present invention relates to a behavioral body identification system capable of identifying a behavioral body reflected in a time-series image with high accuracy based on a characteristic motion of the behavioral body.

Conventionally, the target person's motion characteristic information representing the motion characteristics of the rehabilitation target person is stored, and based on the target person's motion feature information, it is determined whether or not the target person shown in the image is the rehabilitation target person. An operational information processing device is known (see, for example, Patent Document 1).

Patent No. 6675462

However, in the above technique, for example, in the case of the subject movement feature information of "dragging the foot", the target is such that "the maximum amount of change in the y-coordinate of the foot during the movement is less than 1 cm". Since it is necessary to set a judgment standard for each person's motion characteristic information, it is not possible to specify the person to be rehabilitated in the movement for which the judgment standard is not provided.

In addition, even if the target person's movement feature has a judgment standard, if a plurality of people who meet the judgment standard are shown in the image, the rehabilitation target person is specified if other factors are not taken into consideration. Can't.

Therefore, an object of the present invention is to provide a behavioral body identification system capable of identifying a behavioral body reflected in a time-series image with high accuracy based on a characteristic motion of the behavioral body.

The present invention has a storage unit that stores standard motion models related to standard motions of a large number of behavioral bodies, and a first motion for acquiring the motion of the first behavioral object reflected in a plurality of first time-series images. When the acquisition unit and the operation of the acquired first action body are not included in the standard operation model, the registration unit registers the acquired operation as a non-standard operation in association with the first action body. And, the second motion acquisition unit that acquires the motion of the second action object reflected in the plurality of second time-series images, and the acquired motion of the second action object are equal to or more than the non-standard motion. Provided is an action body specifying system characterized by comprising a specific unit for specifying the second action body as the first action body when the second action body has the same degree of coincidence.

According to such a configuration, the behavioral body is specified based on the characteristic movement (“non-standard movement”) that is not included in the “standard movement model”, so that when the face is not clear (the behavioral body is far away). Even if it is reflected, if you are wearing a helmet or mask, if you are shaping it, etc.), even if you are erasing your fingerprints, etc., it is necessary to identify the action body reflected in the time series image with high accuracy. Is possible. Further, in the "registration stage", the predetermined "non-standard operation" is not detected, but the operation not included in the "standard operation model" is judged to be the "non-standard operation". It is not necessary to define "non-standard operation" or to set a judgment standard for judging that it is a predetermined "non-standard operation". Further, in the "specific stage", the behavioral body is not specified by whether or not the detected motion satisfies the criterion set for the predetermined "non-standard motion", but is registered in the registration unit. Since the "non-standard motion" and the motion detected by the second detection unit are directly compared, it is possible to accurately identify the action body.

Further, the registration unit registers a plurality of the non-standard movements for the movement of the first action body, and the second movement acquisition unit is a second image reflected in the plurality of second time-series images. A plurality of movements of the behavioral body of the above are acquired, and the specific unit is predetermined to have a degree of coincidence with the plurality of non-standard movements and the predetermined degree or more among the plurality of movements of the acquired second behavioral body. When there is a ratio or more, it is preferable to specify the second action body as the first action body.

According to such a configuration, when a plurality of "non-standard movements" are detected, the second action body is specified as the first action body, so that the action body can be specified more accurately. ..

Further, in the notification determination unit that determines to perform notification and the plurality of second time-series images obtained by photographing a predetermined location, support is provided to the second action body by another action body. A support information acquisition unit for acquiring support information indicating that the support is provided, the second action body to which the support is provided is specified by the specific unit as the first action body, and the support information is the support. When acquired by the information acquisition unit, the registration unit registers the registered first action body as a support target, and the second action body newly acquired by the second action acquisition unit is said. When specified by the specific unit as the first action body registered as a support target, it is preferable that the notification determination unit decides to perform notification.

According to such a configuration, the action body requiring support is automatically registered, so that it is possible to appropriately and promptly respond to the action body requiring support.

Further, the storage unit further includes a notification determination unit that determines to perform notification, and a third motion acquisition unit that acquires a plurality of motions of the third action body reflected in the plurality of third time-series images. Is further stored with a plurality of non-standard movements that require assistance, and among the plurality of movements of the acquired third action body, the plurality of non-standard movements stored in the storage unit and the plurality of non-standard movements. It is preferable that the notification determination unit decides to perform notification when there are more than a predetermined ratio of those having a predetermined degree of coincidence or more.

With such a configuration, it is possible to deal with a wide range of behavioral bodies that are likely to require support.

Further, according to another aspect of the present invention, it is a program installed in a computer in which a standard motion model for standard motions of a large number of action bodies is stored, and is reflected in a plurality of first time-series images. When the step of acquiring the motion of the first action body and the acquired motion of the first action body are not included in the standard motion model, the acquired motion is regarded as a non-standard motion and the first action is performed. The step of registering in association with the body, the step of acquiring the movement of the second action body reflected in the plurality of second time-series images, and the movement of the acquired second action body are the non-standard movements. Provided is an action body identification program characterized by comprising a step of specifying the second action body as the first action body when the degree of coincidence is equal to or higher than a predetermined value.

Further, according to another aspect of the present invention, there is a storage unit that stores standard motion models for standard motions of a large number of action bodies, and a first action body reflected in a plurality of first time-series images. When the motion of the first motion acquisition unit for acquiring the motion and the motion of the acquired first action body are not included in the standard motion model, the acquired motion is regarded as a non-standard motion and the first action body is described. A second motion acquisition unit for acquiring the motion of the second action body reflected in a plurality of second time-series images, and a second motion acquisition unit, which is provided with a registration unit for registration in association with the above-mentioned acquired second action body. When the movement of the body has a degree of coincidence with the non-standard movement of a predetermined degree or more, communication is performed between the specific unit that identifies the second action body as the first action body and the action body identification device having. It provides a non-standard operation registration device characterized by being possible.

Further, according to another aspect of the present invention, it is a program installed in a computer in which a standard motion model for standard motions of a large number of action bodies is stored, and is reflected in a plurality of first time-series images. When the step of acquiring the motion of the first action body and the acquired motion of the first action body are not included in the standard motion model, the acquired motion is regarded as a non-standard motion and the first action is performed. The computer comprises a step of registering in association with the body, and the computer has a second motion acquisition unit for acquiring the motion of the second action object reflected in a plurality of second time-series images, and the acquired unit. An action body specifying device having a specific unit that identifies the second action body as the first action body when the action of the second action body has a degree of coincidence with the non-standard movement or more. It provides a non-standard operation registration program characterized by being able to communicate with each other.

Further, according to another aspect of the present invention, there is a storage unit that stores standard motion models for standard motions of a large number of motion bodies, and a first action body reflected in a plurality of first time-series images. When the motion of the first motion acquisition unit for acquiring the motion and the motion of the acquired first action body are not included in the standard motion model, the acquired motion is regarded as a non-standard motion and the first action body is described. It is an action body identification device capable of communicating between a registration unit registered in association with and a non-standard operation registration device having the above, and the action of the second action body reflected in a plurality of second time-series images. When the movements of the second action acquisition unit and the acquired second action body have a degree of coincidence with the non-standard movement more than a predetermined degree, the second action body is referred to as the first action. It provides a behavioral body identification device characterized by having a specific part specified as a body.

Further, according to another aspect of the present invention, there is a storage unit that stores standard motion models for standard motions of a large number of action bodies, and a first action body reflected in a plurality of first time-series images. When the motion of the first motion acquisition unit for acquiring the motion and the motion of the acquired first action body are not included in the standard motion model, the acquired motion is regarded as a non-standard motion and the first action body is described. A program installed on a computer capable of communicating between a registration unit registered in association with and a non-standard operation registration device having the second action body reflected in a plurality of second time-series images. When the step of acquiring the motion of the above and the motion of the acquired second action body have a degree of coincidence with the non-standard motion of a predetermined degree or more, the second action body is specified as the first action body. It provides a behavioral body identification program characterized by having steps to do.

According to the behavioral body identification system abnormal behavior of the present invention, it is possible to identify the behavioral body reflected in the time-series image with high accuracy based on the characteristic movement of the behavioral body.

Explanatory drawing of time series image by embodiment of this invention Block diagram of behavioral body identification system according to the embodiment of the present invention Flow chart for specifying an action body according to an embodiment of the present invention Block diagram of behavioral body identification system according to a modified example of the present invention

Hereinafter, the behavioral body identification system 1 according to the embodiment of the present invention will be described with reference to FIGS. 1 to 3.

As shown in FIG. 1, the action body specifying system 1 is for specifying the action body Y reflected in a plurality of time-series images X (each frame constituting a moving image, etc.) taken by a photographing means.

In the present embodiment, a plurality of time-series images X1 (FIG. 1 (a)) taken at a certain place are acquired, and the characteristic operation of the first action body Y1 reflected in the plurality of time-series images X1 is performed. After registering, a plurality of time-series images X2 (FIG. 2A) taken at an arbitrary place are acquired, and the characteristic of the second action body Y2 reflected in the plurality of time-series images X2. By comparing the motion with the characteristic motion of the registered first action body Y1, it is determined whether or not the second action body Y2 is the first action body Y1. In the present embodiment, a human being is adopted as the action body Y, and the action body Y is simply displayed only by the skeleton for easy understanding.

As shown in FIG. 2, the behavioral body identification system 1 includes a storage unit 2, a first detection unit 3, a first motion acquisition unit 4, a registration unit 5, a second detection unit 6, and a second. The operation acquisition unit 7 of 2 and the specific unit 8 are provided. In the present embodiment, as shown in FIG. 2, a storage unit 2, a first operation acquisition unit 4, a registration unit 5, a second operation acquisition unit 7, and a specific unit 8 are provided on the server. It is assumed that communication is possible with the first detection unit 3 and the second detection unit 6 provided externally.

The storage unit 2 stores a "standard motion model" relating to standard motions of a large number of action bodies. In the present embodiment, the "standard motion model" indicates a standard model of human general motion.

As a general type of motion, "walking", "running", "sitting" and the like can be considered, and a "standard motion model" is stored in the storage unit 2 for each type of general motion.

On the other hand, the "standard movement model" is not remembered for irregular movements and habits such as "scratching the head", "touching the nose", and "putting a hand on the waist".

However, the general operation is not limited to the above-mentioned one, and a "standard operation model" of the necessary operation may be stored according to the purpose of use of the action body specifying system 1. Therefore, for example, when the operation of "scratching the head" is required, the "standard operation model" of the operation of "scratching the head" is stored.

In addition, "standard movements of many behavioral bodies" are not limited to "standard movements of all human beings", for example, "standard movements of all Japanese people" and "standard movements of all adults (children)". It suffices to memorize what suits the purpose, such as "genuine movement" and "standard movement of all the visitors to and from the facility where the behavioral body identification system 1 is used".

The "standard motion model" may be stored in the storage unit 2 in advance, but in the present embodiment, a large number of motions (plural) detected from a large number of behavioral bodies in the behavioral body identification system 1. It shall be decided based on the movement of the joints of.

In this case, the motion set as a general motion is detected from the time-series images of a large number of action objects, and for each motion, the motion having a predetermined (threshold value) or more ratio among all the detected motions is "standard". It is conceivable to determine it as an "operation model".

Further, the storage unit 2 also stores a "joint identification standard" and a "behavioral body identification standard".

"Joint identification criteria" is for identifying multiple joints (neck, right elbow, left elbow, hip, right knee, left knee, etc.) of an action body (human), and each is identified for each joint. It shows the shape, direction, size, etc.

The "action body identification criteria" are the "basic posture" of various variations of the action body (human) ("walking", "upright", etc.), "range of motion of each joint", and "between each joint in one action body". It shows the distance "etc."

The first detection unit 3 detects the movement of the first action body Y1 reflected in the plurality of first time-series images X1. In the present embodiment, after specifying the first action body Y1 reflected in the plurality of first time-series images X1, the action of the specified first action body Y1 is detected.

A known method can be used to specify the behavioral body, but in the present embodiment, after detecting a plurality of joints corresponding to the "joint identification criteria" stored in the storage unit 2, the "behavioral body" is specified. With reference to "Specific Criteria", a plurality of joints included in the behavioral body Y are identified.

To detect the movement, the movement of each joint when a predetermined movement is performed is stored in the storage unit 2, and the specified action body Y is tracked by a known tracking method or the like to track the movement of each joint. When the above is detected, it may be detected that the predetermined movement is performed, or the movement of the entire joint, the movement of the hip joint, the movement of the center of gravity (movement speed in the XYZ direction), and the like may be simply detected.

The first motion acquisition unit 4 acquires the motion of the first action body Y1 detected by the first detection unit 3.

When the acquired motion of the first action body Y1 is not included in the "standard motion model", the registration unit 5 registers the acquired motion as a "non-standard motion" in association with the first action body Y1. do.

"Non-standard movements" include, for example, "one shoulder is raised when walking (running)", "slightly dragging when walking (running)", and "with the neck and hips bent". It is conceivable that movements such as "walking (running)" and "sitting movement is slow" may be registered.

The second detection unit 6 detects the movement of the second action body Y2 reflected in the plurality of second time-series images X2. The identification of the second action body Y2 and the detection of the action can be performed by the same method as the identification of the first action body Y1 and the detection of the action in the first detection unit 3.

The second motion acquisition unit 7 acquires the motion of the second action body Y2 detected by the second detection unit 6.

The specifying unit 8 specifies the second action body Y2 as the first action body Y1 when the acquired action of the second action body Y2 has a degree of coincidence with the “non-standard action” at a predetermined level or higher.

Subsequently, the identification of the action body by the action body identification system 1 will be described with reference to the flowchart of FIG.

The behavioral body is specified in two stages, a "registration stage" and a "specific stage".

(1) Registration stage

In the "registration stage", first, the motion of the first action body Y1 reflected in the plurality of first time-series images X1 is acquired (S1). This S1 is preferably performed on a plurality of different first time-series images X1 to acquire a large number of operations.

Subsequently, when the motion of the first action body Y1 acquired in S1 is not included in the "standard motion model" (S2: YES), the acquired motion is regarded as a "non-standard motion" and the first action body Y1 Register in association with (S3).

(2) Specific stage

Subsequently, in the "specific stage", first, the motion of the second action body Y2 reflected in the plurality of second time-series images X2 is acquired (S4).

Subsequently, when the motion of the second action body Y2 acquired in S4 has a degree of coincidence equal to or higher than a predetermined degree with the "non-standard action" registered in S3 (S5: YES), the second action body Y2 is selected. It is specified as the first action body Y1 (S6).

In addition, since it may not be possible to accurately identify by simply comparing the degree of coincidence with one "non-standard motion", in S3, a plurality of "non-standard motions" are performed for the first action body Y1. Registered, in S4, a plurality of actions of the second action body Y2 are acquired, and in S5, a plurality of actions of the second action body Y2 match with a plurality of registered "non-standard actions" by a predetermined value or more. When it has a degree, it is preferable to specify the second action body Y2 as the first action body Y1. In the present embodiment, the detected second action is detected even if the plurality of actions of the detected second action body Y2 and the plurality of registered "non-standard actions" do not all match. When there is a predetermined ratio or more of the plurality of movements of the body Y2 that have a predetermined degree of coincidence with the plurality of registered "non-standard movements", the second action body Y2 is referred to as the first action body Y1. It shall be specified as.

The behavioral body identification system 1 having such a configuration can be used for various purposes.

For example, when the action body identification system 1 is used for the purpose of admission management, attendance management, personal authentication (payment system, station ticket gate, bus boarding, etc.), the action of the target action body is registered in advance at the "registration stage". However, by performing a "specific stage" at the airport, the entrance of the company, the ticket gate, etc., it is possible to manage entry and exit. In addition, by registering in advance the movement of a behavioral body that may be wandering due to dementia or the like, it is possible to discover the behavioral body from an image or the like captured by a street camera.

In addition, even when the action body identification system 1 is used for the purpose of criminal investigation, the action of the target action body is registered in advance in the "registration stage", and the "specific stage" is set at the infestation prediction point of the target action body. By doing ", it becomes possible to discover the target behavioral body.

On the other hand, it is also possible to use the action body specifying system 1 in a state where the target action body is undecided.

For example, repeaters to buses, unmanned stores, ATMs, stores and public facilities (elderly people, physically handicapped customers, etc.) may need assistance.

Therefore, by installing the action body identification system 1 on the bus or the like, the "non-standard operation" of the user of the bus or the like is automatically registered in association with the user. Then, when the bus or the like is used next time, the user can be identified, so that the staff can rush to support when getting on the bus or the like.

As described above, in the behavioral body identification system 1 according to the present embodiment, when the motion of the first behavioral entity Y1 is not included in the "standard motion model" in the "registration stage", the motion is "non-standard". When the movement of the second action body Y2 has a predetermined degree of coincidence with the "non-standard movement" in the "specific stage", it is registered as "movement" in association with the first action body Y1. The action body Y2 of 2 is specified as the first action body Y1.

According to such a configuration, the behavioral body is specified based on the characteristic movement (“non-standard movement”) that is not included in the “standard movement model”, so that when the face is not clear (the behavioral body is far away). Even if it is reflected, if you are wearing a helmet or mask, if you are shaping it, etc.), even if you are erasing your fingerprints, etc., it is necessary to identify the action body reflected in the time series image with high accuracy. Is possible. Further, in the "registration stage", the predetermined "non-standard operation" is not detected, but the operation not included in the "standard operation model" is judged to be the "non-standard operation". It is not necessary to define "non-standard operation" or to set a judgment standard for judging that it is a predetermined "non-standard operation". Further, in the "specific stage", the action body is not specified by whether or not the detected motion satisfies the determination criteria provided for the predetermined "non-standard motion", but is registered in the registration unit 5. Since the "non-standard motion" and the motion detected by the second detection unit 6 are directly compared, it is possible to accurately identify the action body.

Further, in the action body specifying system 1 according to the present embodiment, among the plurality of actions of the second action body Y2, those having a predetermined degree of coincidence with the plurality of "non-standard actions" in the "specific stage" When there is a predetermined ratio or more, the second action body Y2 is specified as the first action body Y1.

According to such a configuration, when a plurality of "non-standard motions" are detected, the second action body Y2 is specified as the first action body Y1, so that the action body can be specified more accurately. Will be.

The behavioral body identification system of the present invention is not limited to the above-described embodiment, and various modifications and improvements can be made within the scope described in the claims.

For example, in the above embodiment, the support by bus or the like has been described, but not all the identified behavioral bodies need the support. Therefore, as shown in FIG. 4, in the notification determination unit 9 that determines to perform notification and the plurality of second time-series images X2 obtained by photographing a predetermined location, the second action body Y2 is formed by another action body. A third detection unit 10 that detects that support has been provided to the second action body Y2, and a third detection unit that indicates that support has been provided to the second action body Y2 by another action body. It is conceivable that the support information acquisition unit 11 acquired from the 10 is further provided.

In this case, for example, the movement of each joint when the support operation is performed is stored in the storage unit 2, and the distance between the other action body that performed the support operation and the second action body Y2 is a predetermined distance. If it is within the range, the third detection unit 10 detects that the support has been provided to the second action body Y2 by another action body.

Then, when the second action body Y2 to which the support is provided is specified as the first action body Y1 by the specific unit 8 and the support information is acquired by the support information acquisition unit 11, the registration unit 5 is registered. The first action body Y1 is registered as a support target. After that, when the second action body Y2 newly detected by the second detection unit 6 is specified by the specific unit 8 as the "first action body Y1 registered as a support target", the notification determination unit 9 determines. Decide to make a notification. When it is decided to perform the notification, an alert may be given by voice or the like to the bus or the like in which the second detection unit 6 is installed.

According to such a configuration, the action body requiring support is automatically registered, so that it is possible to appropriately and promptly respond to the action body requiring support.

The second detection unit 6 may also serve as the third detection unit 10. Similarly, the first motion acquisition unit 4 or the second motion acquisition unit 7 may also serve as the support information acquisition unit 11.

Further, the above modification can be applied to the third action body Y3 which is not registered in the registration unit 5.

In this case, as shown in FIG. 4, a third motion acquisition unit 12 for acquiring a plurality of motions of the third action body Y3 reflected in the plurality of third time-series images X3 is further provided, and the acquired third action body Y3 is provided. When there is a predetermined ratio or more of the plurality of movements of the action body Y3 of 3 having a degree of coincidence with the plurality of non-standard movements of the first action body Y1 registered as the support target, the notification is notified. The determination unit 9 determines to perform the notification.

According to such a configuration, the action body (third action body Y3) that performs the same non-standard movement as the first action body Y1 registered as the support target is the same as the first action body Y1. Since support is often required, it is possible to provide a wide range of responses to behavioral bodies that are likely to require support.

In this case, the "degree of coincidence above a predetermined value" and the "predetermined ratio" may be the same as or different from those of the first embodiment. Further, instead of "a plurality of non-standard movements of the first action body Y1 registered as the support target", "a plurality of non-standard movements requiring support" are stored in the storage unit 2 and the said. The notification determination unit 9 may decide to perform notification when there are a predetermined ratio or more of those having a predetermined degree of coincidence with "a plurality of non-standard operations requiring support". "Multiple non-standard movements that require assistance" include, for example, "dragging legs", "putting hands on knees", "grabbing handrails", "slow walking (sitting) movements", etc. It is conceivable, but it is not limited to these, and it may be set according to the situation and purpose.

Further, the third motion acquisition unit 12 acquires a plurality of motions of the third action body Y3 from any of the first detection unit 3, the second detection unit 6, and the third detection unit 10. It may be obtained from another detection unit. Further, the first motion acquisition unit 4 or the second motion acquisition unit 7 may also serve as the third motion acquisition unit 12.

Further, in the above embodiment, the first detection unit 3 and the second detection unit 6 are included in the action body system 1, but if the action of the action body can be acquired, they are included in the action body identification system 1. It doesn't have to be. Further, the non-standard operation registration device (storage unit 2, the first operation acquisition unit 4, the registration unit 5) and the action body identification device (second operation acquisition unit 7, specific unit 8) are separated from each other. It is also conceivable to make it communicable.

Further, in the above-described embodiment, the human being is described as an example of the action body Y, but animals and robots can also be used. For example, in the case of animals, it is possible to identify individuals that damage the fields.

Further, in the above embodiment, the movement of the action body is detected based on the movement of a plurality of joints, but the movement of the action body may be detected based on other than the movement of the joint.

Further, in the above embodiment, the case where only the first action body Y1 is shown in the first time-series image X1 and only the second action body Y2 is shown in the second time-series image X2 has been described. However, even when a plurality of action bodies Y are shown, registration and identification are performed for each action body Y in the same manner. In this case, after detecting a plurality of joints corresponding to the "joint identification standard" stored in the storage unit 2, the plurality of joints included in the behavioral body Y are specified with reference to the "behavioral body identification standard". Therefore, it is possible to specify each action body Y.

Further, the present invention can be applied to a program corresponding to the processing performed by the action body identification system 1, the non-standard operation registration device, the action body identification device, and a recording medium in which the program is stored. In the case of a recording medium, the program will be installed on a computer or the like. Here, the recording medium in which the program is stored may be a non-transient recording medium. As a non-transient recording medium, a CD-ROM or the like can be considered, but the non-transitory recording medium is not limited thereto. The above-mentioned "plurality of non-standard operations requiring support" may be stored not only in the storage unit 2 but also in a computer or an action body specifying device.

1 Action body identification system 2 Storage unit 3 First detection unit 4 First operation acquisition unit 5 Registration unit 6 Second detection unit 7 Second operation acquisition unit 8 Specific unit 9 Notification determination unit 10 Third detection unit 11 Support information acquisition unit 12 Third operation acquisition unit

Claims (24)

A storage unit that stores standard motion models for standard motions of many behavioral bodies,
A first motion acquisition unit that acquires the motion of the first action body reflected in a plurality of first time-series images, and a first motion acquisition unit.
When the acquired motion of the first action body is not included in the standard motion model, a registration unit that registers the acquired motion as a non-standard motion in association with the first action body, and
A second motion acquisition unit that acquires the motion of the second action body reflected in a plurality of second time-series images, and a second motion acquisition unit.
When the acquired movement of the second action body has a degree of coincidence with the non-standard movement of a predetermined degree or more, a specific unit that specifies the second action body as the first action body and a specific unit.
A behavioral body identification system characterized by being equipped with. The registration unit registers a plurality of the non-standard movements for the movement of the first action body.
The second motion acquisition unit acquires a plurality of motions of the second action body reflected in the plurality of second time-series images.
The specific unit is the second when there is a predetermined ratio or more of the plurality of movements of the acquired second action body that have a degree of coincidence with the plurality of non-standard movements of the predetermined or higher. The action body identification system according to claim 1, wherein the action body is specified as the first action body. The notification determination unit that decides to perform notification, and
A support information acquisition unit that acquires support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images taken at a predetermined location.
Further prepare
When the second action body to which the support is provided is specified as the first action body by the specific unit and the support information is acquired by the support information acquisition unit, the registration unit is registered. Register the first action body as a support target,
When the second action body newly acquired by the second action acquisition unit is specified by the specific unit as the first action body registered as the support target, the notification determination unit performs notification. The behavioral body identification system according to claim 1 or 2, characterized in that the determination is made. The notification determination unit that decides to perform notification, and
A third motion acquisition unit that acquires a plurality of motions of the third action object reflected in the plurality of third time-series images, and a third motion acquisition unit.
Further prepare
The storage unit further stores a plurality of non-standard operations that require assistance.
The notification determination is made when, among the plurality of movements of the acquired third action body, there are more than a predetermined ratio of those having a predetermined degree of coincidence with the plurality of non-standard movements stored in the storage unit. The behavioral body identification system according to claim 1 or 2, wherein the unit determines to perform notification. A program installed on a computer that stores standard behavior models for the standard behavior of many behavioral bodies.
A step of acquiring the movement of the first action body reflected in a plurality of first time-series images, and
When the acquired motion of the first action body is not included in the standard motion model, the step of registering the acquired motion as a non-standard motion in association with the first action body, and
The step of acquiring the movement of the second action body reflected in the plurality of second time-series images, and
A step of specifying the second action body as the first action body when the movement of the acquired second action body has a degree of coincidence with the non-standard movement of a predetermined degree or more.
A behavioral body identification program characterized by being equipped with. In the registration step, a plurality of the non-standard movements are registered for the movement of the first action body.
In the step of acquiring the movement of the second action body, a plurality of movements of the second action body reflected in the plurality of second time-series images are acquired.
In the specific step, when there is a predetermined ratio or more of the plurality of movements of the acquired second action body that have a degree of coincidence with the plurality of non-standard movements and the predetermined degree or more, the second movement is performed. The action body specifying program according to claim 5, wherein the action body of the above is specified as the first action body. The steps that determine the notification and
A step of acquiring support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images obtained by photographing a predetermined place, and a step of acquiring support information.
Further prepare
In the step of registering, when the second action body to which the support is performed is specified as the first action body in the step of specifying the support information and the support information is acquired in the step of acquiring the support information, the support information is acquired. The registered first action body is registered as a support target, and
When the second action body newly acquired in the step of acquiring the action of the second action body is specified in the step of specifying as the first action body registered as the support target, the notification is given. The behavioral body identification program according to claim 5 or 6, wherein in the step of determining to perform the notification, it is determined to perform the notification. The steps that determine the notification and
A step of acquiring a plurality of actions of a third action body reflected in a plurality of third time-series images, and
Further prepare
The computer also stores a number of non-standard operations that require assistance.
When there is a predetermined ratio or more of the plurality of movements of the acquired third action body that have a predetermined degree of coincidence with the plurality of non-standard movements stored in the storage unit, the notification is given. The behavioral body identification program according to claim 5 or 6, wherein in the step of deciding to perform, it is determined to perform notification. A storage unit that stores standard motion models for standard motions of many behavioral bodies,
A first motion acquisition unit that acquires the motion of the first action body reflected in a plurality of first time-series images, and a first motion acquisition unit.
When the acquired motion of the first action body is not included in the standard motion model, a registration unit that registers the acquired motion as a non-standard motion in association with the first action body, and
Equipped with
The second motion acquisition unit that acquires the motion of the second action object reflected in the plurality of second time-series images and the acquired motion of the second action body match the non-standard motion by a predetermined value or more. A non-standard operation registration device characterized in that communication is possible between a specific unit that identifies the second action body as the first action body and the action body identification device having the degree. .. The registration unit registers a plurality of the non-standard movements for the movement of the first action body.
The second motion acquisition unit acquires a plurality of motions of the second action body reflected in the plurality of second time-series images.
Among the plurality of movements of the second action body acquired by the second movement acquisition unit, the specific unit has a predetermined ratio or more of the plurality of non-standard movements and the degree of coincidence of the predetermined or higher. The non-standard operation registration device according to claim 9, wherein the second action body is specified as the first action body in the case of the case. The notification determination unit that decides to perform notification, and
A support information acquisition unit that acquires support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images taken at a predetermined location.
Further prepare
The second motion acquisition unit is support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images obtained by photographing a predetermined place. To get more,
When the second action body to which the support is provided is specified as the first action body by the specific unit and the support information acquisition unit is acquired by the support information, the registration unit is registered. Register the first action body as a support target,
When the second action body newly acquired by the second action acquisition unit is specified by the specific unit as the first action body registered as the support target, the notification determination unit performs notification. The non-standard operation registration device according to claim 9 or 10, characterized in that the determination is made. The notification determination unit that decides to perform notification, and
A third motion acquisition unit that acquires a plurality of motions of the third action object reflected in the plurality of third time-series images, and a third motion acquisition unit.
Further prepare
The storage unit or the behavioral body identification device further stores a plurality of non-standard movements requiring assistance.
When, among the plurality of movements of the acquired third action body, there are more than a predetermined ratio of those having a predetermined degree of coincidence with the plurality of stored non-standard movements, the notification determination unit notifies. The non-standard operation registration device according to claim 9 or 10, wherein the device is determined to perform the above-mentioned method. A program installed on a computer that stores standard behavior models for the standard behavior of many behavioral bodies.
A step of acquiring the movement of the first action body reflected in a plurality of first time-series images, and
When the acquired motion of the first action body is not included in the standard motion model, the step of registering the acquired motion as a non-standard motion in association with the first action body, and
Equipped with
In the computer, the second motion acquisition unit that acquires the motion of the second action object reflected in the plurality of second time-series images, and the acquired motion of the second action body are the non-standard motions. A non-characteristic unit characterized in that communication is possible between a specific unit that identifies the second action body as the first action body and an action body identification device having the second action body when the degree of coincidence is equal to or higher than a predetermined value. Standard operation registration program. In the registration step, a plurality of the non-standard movements are registered for the movement of the first action body.
The second motion acquisition unit acquires a plurality of motions of the second action body reflected in the plurality of second time-series images.
Among the plurality of movements of the second action body acquired by the second movement acquisition unit, the specific unit has a predetermined ratio or more of the plurality of non-standard movements and the degree of coincidence of the predetermined or higher. The non-standard action registration program according to claim 13, wherein the second action body is specified as the first action body in the case of the case. The steps that determine the notification and
A step of acquiring support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images obtained by photographing a predetermined place, and a step of acquiring support information.
Further prepare
The second motion acquisition unit is support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images obtained by photographing a predetermined place. To get more,
When the second action body to which the support is provided is specified as the first action body by the specific unit and the support information is acquired in the step of acquiring the support information, in the registration step, the support information is acquired. Register the registered first action body as a support target,
When the second action body newly acquired by the second action acquisition unit is specified by the specific unit as the first action body registered as the support target, the step of determining to perform the notification. The non-standard operation registration program according to claim 13, wherein the non-standard operation registration program is characterized in that it is determined to perform notification. The steps that determine the notification and
A step of acquiring a plurality of actions of a third action body reflected in a plurality of third time-series images, and
Further prepare
The computer or the behavioral identification device further stores a plurality of non-standard actions requiring assistance.
Among the plurality of movements of the acquired third action body, when there is a predetermined ratio or more that has a degree of coincidence with the plurality of stored non-standard movements, the notification is performed. The non-standard operation registration program according to claim 13 or 14, wherein in the determination step, it is determined to perform notification. A storage unit that stores standard motion models related to standard motions of a large number of behavioral bodies, a first motion acquisition section that acquires the motion of the first behavioral body reflected in a plurality of first time-series images, and a first motion acquisition unit. When the acquired motion of the first action body is not included in the standard motion model, it has a registration unit for registering the acquired motion as a non-standard motion in association with the first action body. It is an action body identification device that can communicate with a non-standard operation registration device.
A second motion acquisition unit that acquires the motion of the second action body reflected in a plurality of second time-series images, and a second motion acquisition unit.
When the acquired movement of the second action body has a degree of coincidence with the non-standard movement of a predetermined degree or more, a specific unit that specifies the second action body as the first action body and a specific unit.
A behavioral body identification device characterized by being equipped with. The registration unit registers a plurality of the non-standard movements for the movement of the first action body.
The second motion acquisition unit acquires a plurality of motions of the second action body reflected in the plurality of second time-series images.
Among the plurality of movements of the second action body acquired by the second movement acquisition unit, the specific unit has a predetermined ratio or more of the plurality of non-standard movements and the degree of coincidence of the predetermined or higher. The action body specifying device according to claim 17, wherein the second action body is specified as the first action body. The notification determination unit that decides to perform notification, and
A support information acquisition unit that acquires support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images taken at a predetermined location.
Further prepare
When the second action body to which the support is provided is specified as the first action body by the specific unit and the support information is acquired by the support information acquisition unit, the registration unit is registered. Register the first action body as a support target,
When the second action body newly acquired by the second action acquisition unit is specified by the specific unit as the first action body registered as the support target, the notification determination unit performs notification. The behavioral body identification device according to claim 17 or 18, characterized in that the determination is made. The notification determination unit that decides to perform notification, and
A third motion acquisition unit that acquires a plurality of motions of the third action object reflected in the plurality of third time-series images, and a third motion acquisition unit.
Further prepare
A plurality of non-standard operations requiring assistance are further stored in the storage unit or a storage unit provided separately.
When, among the plurality of movements of the acquired third action body, there are more than a predetermined ratio of those having a predetermined degree of coincidence with the plurality of stored non-standard movements, the notification determination unit notifies. 17. The behavioral body identification device according to claim 17 or 18, wherein the device is determined to perform the above-mentioned method. A storage unit that stores standard motion models related to standard motions of a large number of behavioral bodies, and a first motion acquisition section that acquires the motions of the first behavioral body reflected in a plurality of first time-series images. When the acquired motion of the first action body is not included in the standard motion model, it has a registration unit for registering the acquired motion as a non-standard motion in association with the first action body. A program installed on a computer that can communicate with a non-standard operation registration device.
The step of acquiring the movement of the second action body reflected in the plurality of second time-series images, and
A step of specifying the second action body as the first action body when the movement of the acquired second action body has a degree of coincidence with the non-standard movement of a predetermined degree or more.
A behavioral body identification program characterized by being equipped with. The registration unit registers a plurality of the non-standard movements for the movement of the first action body.
In the step of acquiring the movement of the second action body, a plurality of movements of the second action body reflected in the plurality of second time-series images are acquired.
In the specific step, when there is a predetermined ratio or more of the plurality of movements of the acquired second action body that have a degree of coincidence with the plurality of non-standard movements and the predetermined degree or more, the second movement is performed. 21. The action body specifying program according to claim 21, wherein the action body is specified as the first action body. The steps that determine the notification and
A step of acquiring support information indicating that support has been provided to the second action body by another action body in the plurality of second time-series images obtained by photographing a predetermined place, and a step of acquiring support information.
Further prepare
When the second action body to which the support is provided is specified as the first action body in the step of specifying the support information and the support information is acquired in the step of acquiring the support information, the registration unit receives the support. Register the registered first action body as a support target,
When the second action body newly acquired in the step of acquiring the action of the second action body is specified in the step of specifying as the first action body registered as the support target, the notification is given. The behavioral body identification program according to claim 21 or 22, wherein in the step of determining to perform the notification, it is determined to perform the notification. The steps that determine the notification and
A step of acquiring a plurality of actions of a third action body reflected in a plurality of third time-series images, and
Further prepare
The storage unit or the computer further stores a plurality of non-standard operations that require assistance.
Among the plurality of movements of the acquired third action body, when there is a predetermined ratio or more that has a degree of coincidence with the plurality of stored non-standard movements, the notification is performed. The behavioral body identification program according to claim 21 or 22, wherein in the determination step, it is determined to perform notification.

Images