JP6801902B1 - Child Abuse Sign Identification Program and System - Google Patents

Abstract

PROBLEM TO BE SOLVED: To determine the possibility of abuse with high accuracy and automatically without relying more on human hands. SOLUTION: In a child abuse sign discrimination program for discriminating signs of abuse of a child, an information acquisition step of acquiring image information of a child to be discriminated, reference image information of a child captured in the past, and the image information thereof. Discrimination step to determine the possibility of child abuse based on the reference image information according to the image information acquired in the above information acquisition step using the degree of association with the possibility of child abuse at least three levels. And let the computer do it. [Selection diagram] Fig. 3

Description

The present invention relates to a child abuse sign discrimination program and system suitable for discriminating the signs of child abuse by a parent with high accuracy.

In recent years, parental abuse of children has increased, and numerous tragic incidents have been reported that sometimes lead to the death of children. Child guidance centers and others are making efforts to detect such abuse of children by parents as soon as possible and prevent disastrous consequences.

However, there are oversights and limitations in detecting abuse by humans such as child guidance center staff and school teachers. For this reason, there has been an increasing social demand for improving the accuracy of detection and protecting children by obtaining assistance from artificial intelligence.

Therefore, the present invention has been devised in view of the above-mentioned problems, and an object of the present invention is child abuse that can automatically determine the signs of parental abuse of a child with high accuracy. To provide a symptom discrimination program and system.

The child abuse sign discrimination program according to the present invention is a child abuse sign discrimination program that discriminates signs of child abuse by a parent, and targets individual children to be discriminated to target faces, limbs, and clothes in chronological order. An information acquisition step in which images are taken multiple times and the change tendency of the image over time is acquired as image information, and the face, limbs, and clothes are time-series by targeting individual children in the past. It was acquired in the above information acquisition step by using the reference image information consisting of the tendency of the image to change over time and the possibility of abuse of the child in three or more stages. Based on the reference image information according to the image information, the computer is made to perform a determination step for determining the possibility of child abuse.

Even if you do not have any special skills or experience, you can accurately and automatically determine the signs of parental abuse of your child without resorting to human resources.

It is a block diagram which shows the whole structure of the system to which this invention is applied. It is a figure which shows the specific configuration example of a search device. It is a figure for demonstrating the operation of this invention. It is a figure for demonstrating the operation of this invention. It is a figure for demonstrating the operation of this invention. It is a figure for demonstrating the operation of this invention. It is a figure for demonstrating the operation of this invention. It is a figure for demonstrating the operation of this invention. It is a figure for demonstrating the operation of this invention.

Hereinafter, the child abuse sign discrimination program to which the present invention is applied will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an overall configuration of a child abuse sign discrimination system 1 to which a child abuse sign discrimination program to which the present invention is applied is implemented. The child abuse sign discrimination system 1 includes an information acquisition unit 9, a discrimination device 2 connected to the information acquisition unit 9, and a database 3 connected to the discrimination device 2.

The information acquisition unit 9 is a device for a person who uses this system to input various commands and information, and specifically, is composed of a keyboard, buttons, a touch panel, a mouse, a switch, and the like. The information acquisition unit 9 is not limited to a device for inputting text information, and may be configured by a device such as a microphone that can detect voice and convert it into text information. Further, the information acquisition unit 9 may be configured as an image pickup device capable of capturing an image of a camera or the like. The information acquisition unit 9 may be configured by a scanner having a function of recognizing a character string from a paper-based document. Further, the information acquisition unit 9 may be integrated with the discrimination device 2 described later. The information acquisition unit 9 outputs the detected information to the determination device 2. Further, the information acquisition unit 9 may be configured by means for specifying the position information by scanning the map information. Further, the information acquisition unit 9 may be composed of an illuminance sensor for measuring a temperature sensor, a humidity sensor, and a wind direction sensor. Further, the information acquisition unit 9 may be composed of a communication interface for acquiring data about the weather from the Japan Meteorological Agency or a private weather forecast company. Further, the information acquisition unit 9 may be composed of a body sensor that is attached to the body to detect body data, and this body sensor detects, for example, body temperature, heart rate, blood pressure, number of steps, walking speed, and acceleration. It may be composed of a sensor for the purpose. Further, the body sensor may acquire biological data of animals as well as humans. Further, the information acquisition unit 9 may be configured as a device that acquires information such as drawings by scanning or reading it from a database. In addition to these, the information acquisition unit 9 may be configured by an odor sensor that detects an odor or scent.

Database 3 stores various information necessary for discriminating signs of child abuse. Information necessary for discriminating signs of child abuse includes reference image information of children captured in the past and reference image information consisting of moving images of children captured in the past, which are applied to the types of child behavior patterns. To obtain reference behavior pattern information, reference attribute information about the child captured in obtaining reference image information, reference home environment information regarding the child's home environment captured in obtaining reference image information, and reference image information. Reference listening information and the like regarding the content of the abuse heard from the child imaged in the above are accumulated in relation to the degree of possibility of the child being abused as output data.

That is, in the database 3, in addition to such reference image information, any one or more of reference behavior pattern information, reference attribute information, reference home environment information, and reference listening information, and abuse of children The degree of possibility is memorized in association with each other.

The discrimination device 2 is composed of, for example, an electronic device such as a personal computer (PC), but is embodied in any other electronic device such as a mobile phone, a smartphone, a tablet terminal, a wearable terminal, etc., in addition to the PC. It may be converted. The user can obtain a search solution by the discrimination device 2.

FIG. 2 shows a specific configuration example of the discrimination device 2. The discrimination device 2 performs wired communication or wireless communication with a control unit 24 for controlling the entire discrimination device 2 and an operation unit 25 for inputting various control commands via operation buttons, a keyboard, or the like. A communication unit 26 for the purpose, a determination unit 27 for making various determinations, and a storage unit 28 for storing a program for performing a search to be executed represented by a hard disk or the like are connected to the internal bus 21, respectively. .. Further, a display unit 23 as a monitor that actually displays information is connected to the internal bus 21.

The control unit 24 is a so-called central control unit for controlling each component mounted in the discrimination device 2 by transmitting a control signal via the internal bus 21. Further, the control unit 24 transmits various control commands via the internal bus 21 in response to the operation via the operation unit 25.

The operation unit 25 is embodied by a keyboard or a touch panel, and an execution command for executing a program is input from the user. When the execution command is input by the user, the operation unit 25 notifies the control unit 24 of the execution command. Upon receiving this notification, the control unit 24, including the determination unit 27, executes a desired processing operation in cooperation with each component. The operation unit 25 may be embodied as the information acquisition unit 9 described above.

The determination unit 27 determines the search solution. The discriminating unit 27 reads out various information stored in the storage unit 28 and various information stored in the database 3 as necessary information when executing the discriminating operation. The discriminating unit 27 may be controlled by artificial intelligence. This artificial intelligence may be based on any well-known artificial intelligence technique.

The display unit 23 is composed of a graphic controller that creates a display image based on the control by the control unit 24. The display unit 23 is realized by, for example, a liquid crystal display (LCD) or the like.

When the storage unit 28 is composed of a hard disk, predetermined information is written to each address based on the control by the control unit 24, and is read out as needed. Further, the storage unit 28 stores a program for executing the present invention. This program is read by the control unit 24 and executed.

The operation in the child abuse sign discrimination system 1 having the above-described configuration will be described.

In the child abuse sign determination system 1, for example, as shown in FIG. 3, it is premised that three or more levels of association between the reference image information and the possibility of child abuse are set in advance. The reference image information is information obtained by capturing an image of a child, and also includes information obtained for the first time by analyzing this image information. This image may be a moving image as well as a still image. This image may be taken through a camera installed in a school classroom, a camera installed in a schoolyard, or the like, or is an image taken by a camera by targeting an individual child. May be good. This reference image information may identify the state of scratches, swelling, and bruises on the child's face or limbs by analyzing images taken of the child's face or exposed limbs. Further, the reference image information may be used to capture the state of the child's clothing. Then, the state of dirt on the clothes may be detected by analyzing the reference image information obtained by capturing the image of the clothes.

The possibility of child abuse is the possibility of abuse that has been investigated at child guidance centers, municipalities, local governments, schools, etc. in the past, and is expressed as a percentage from the one with the highest possibility of abuse. Alternatively, it may be displayed in two simple stages, "with signs of abuse" and "without signs of abuse". The possibility of such abuse may be utilized if the risk level, etc. indicating whether or not it was a dangerous case is described from the report, etc. in the past child guidance center. If such a risk level is not described, it may be ranked according to the final course (for example, whether it was killed or was transported to an ambulance, the number of days of hospitalization, etc.). .. In addition, the possibility of this abuse may be automatically classified by reading the description contents of past reports and reports through OCR etc. and analyzing the sentences. In such cases, the existing artificial intelligence technology may be used. You may use it.

In addition, the likelihood of child abuse may be judged at its level based on previous experience by the evaluator, or by actually reading past case material and multiple evaluators assessing the level. However, they may be statistically analyzed to determine the degree of potential for abuse.

In the example of FIG. 3, it is assumed that the input data is, for example, reference image information P01 to P03. The reference image information P01 to P03 as such input data is linked to the possibility of child abuse as output. In this output, the probabilities A to D of child abuse as output solutions are displayed as percentages according to the height of the signs.

The reference image information is associated with each other through three or more levels of association with the abuse possibilities A to D as the output solution. The reference image information is arranged on the left side through this degree of association, and each possibility of abuse is arranged on the right side through this degree of association. The degree of association indicates the degree of the possibility of abuse and the degree of relevance to the reference image information arranged on the left side. In other words, this degree of association is an index showing what kind of abuse possibility each reference image information is likely to be associated with, and is used to select the most probable abuse possibility from the reference image information. It shows the accuracy in. In the example of FIG. 3, w13 to w19 are shown as the degree of association. These w13 to w19 are shown in 10 stages as shown in Table 1 below, and the closer to 10 points, the higher the degree of relevance of each combination as an intermediate node to the possibility of abuse as an output. On the contrary, the closer to one point, the lower the degree of relevance of each combination as an intermediate node to the price as an output.

The discriminating device 2 acquires in advance the degree of association w13 to w19 of three or more stages shown in FIG. That is, the discriminating device 2 accumulates a past data set as to which of the reference image information and the possibility of abuse in that case is adopted and evaluated in discriminating the actual search solution, and stores these. By analyzing and analyzing, the degree of association shown in FIG. 3 is created.

For example, it is assumed that the abusive possibility A is highly evaluated as the abusive possibility of the reference image information imaged on the child in the past. By collecting and analyzing such a data set, the degree of association with the reference image information is strengthened.

This analysis and analysis may be performed by artificial intelligence. In such a case, for example, in the case of reference image information P01, analysis is performed from various data as a result of evaluating the possibility of abuse in the past. In the case of reference image information P01, if there are many cases of abuse possibility A, the degree of association that leads to the evaluation of this abuse possibility is set higher, and if there are many cases of abuse possibility B, Set a higher degree of association that leads to an assessment of this potential abuse. For example, in the example of the reference image information P01, the abuse possibility A and the abuse possibility C are linked, but from the previous case, the degree of association of w13 leading to the abuse possibility A is set to 7 points, and the abuse possibility C is set. The degree of association of w14 that leads to is set to 2 points.

Further, the degree of association shown in FIG. 3 may be composed of the nodes of the neural network in artificial intelligence. That is, the weighting coefficient for the output of the node of this neural network corresponds to the above-mentioned degree of association. Further, the network is not limited to a neural network, and may be composed of all decision-making factors constituting artificial intelligence.
In such a case, as shown in FIG. 4, reference image information is input as input data, the possibility of abuse is output as output data, and at least one or more hidden layers are provided between the input node and the output node. , Machine learning may be made. The above-mentioned degree of association is set in either one or both of the input node and the hidden layer node, and this is the weight of each node, and the output is selected based on this. Then, when this degree of association exceeds a certain threshold value, the output may be selected.

Such a degree of association is what is called learned data in artificial intelligence. After creating such learned data through a data set of images of children to be evaluated before and the possibility of abuse actually discriminated and evaluated, the possibility of abuse will be newly determined from now on. Above, the above-mentioned learned data will be used to search for the possibility of abuse. In such a case, the image information obtained by actually capturing the image of the child in the area to be discriminated is newly acquired. The newly acquired image information is input by the information acquisition unit 9 described above. The image information is acquired by imaging the appearance of the child. This determination method may be performed by the same method as the reference image information described above.

Based on the image information newly acquired in this way, the possibility of abuse is determined. In such a case, the degree of association shown in FIG. 3 (Table 1) acquired in advance is referred to. For example, when the newly acquired image information is the same as or similar to P02, the abuse possibility B is associated with w15 and the abuse possibility C is associated with the association degree w16 through the degree of association. In such a case, the abuse possibility B having the highest degree of association is selected as the optimum solution. However, it is not essential to select the one with the highest degree of association as the optimum solution, and the abuse possibility C, which has the lowest degree of association but the association itself is recognized, may be selected as the optimum solution. In addition to this, it goes without saying that an output solution to which the arrows are not connected may be selected, and as long as it is based on the degree of association, it may be selected in any other priority.

In this way, the most suitable possibility of abuse can be searched for from the newly acquired image information and displayed to the user. By looking at the results of this search, signs of child abuse can be discerned before any incident occurs, and the safety of the child can be ensured.

As the above-mentioned image, in addition to the image captured by a normal camera, any one or more of a spectrum image and an ultrasonic image may be acquired. In such a case, it is necessary to capture one or more of a spectrum image, a visible image, and an ultrasonic image according to the image information to be acquired as reference image information.

Further, in the present invention, as reference image information, a state of scratches, swelling, bruises on the face or limbs, or a state of dirt on clothes may be detected. Since such scratches, swelling, and stains indicate signs of abuse of children, the accuracy of discrimination can be improved by extracting them. Such scratches, swelling, and stains can be extracted by image analysis of reference image information. As a specific extraction method, it can be extracted from an image by using a pattern matching technique or a deep learning technique of artificial intelligence.

As reference image information, the condition of scratches, swelling, bruises, or stains on clothes of the face or limbs is detected, and in relation to this, the above-mentioned is described from the data set with the data of past possibility of abuse. Form the degree of association and use this as the trained model. Then, when actually determining the possibility of abuse of a child, the state of scratches, swelling, bruises, or dirt on clothes of the face or limbs is detected from the image information captured for the child. At this time, among the scratches, swelling, bruises, and stains on clothes of the face or limbs, those matched to the trained model are detected. If, as a trained model, the possibility of abuse and the degree of association are formed in relation to the swelling of the face extracted from the reference image information, the swelling of the face is similarly extracted from the image information. That is, scratches, swelling, etc. of the face or limbs as detection targets are extracted by associating the image information with the reference image information and making them consistent with each other. Based on the scratches, swelling, etc. of the face or limbs extracted from such image information, it is possible to determine the actual possibility of abuse by referring to the degree of association formed between the reference image information and the possibility of abuse. It will be possible.

In the example of FIG. 5, the degree of association is formed between the reference behavior pattern information applied to the type of the child's behavior pattern from the reference image information consisting of the moving images of the child captured in the past and the possibility of abuse. is there.

The behavior pattern information referred to here is obtained by analyzing a moving image of a child as reference image information. Behaviors that show signs of abuse include, for example, poor involvement with others, frightened attitudes, looking at the complexion of parents and adults, abnormal dietary attachment, and abnormal anxiety when undressing. If it is considered that the behavior is extremely restless, violent and emotionally unstable, or that the person does not have a basic lifestyle, such behavior should be categorized and patterned in advance.

If the pattern is such that the neck sways back and forth or the limbs are poorly shaken as a sign that the attitude is frightening, it is determined by analyzing the moving image whether or not it is applied to the pattern. In the process of this determination, an existing image analysis technique or deep learning technique may be used. Then, when an action applied to this pattern is detected, it is applied to the reference behavior pattern information classified according to each behavior pattern.

A trained model is created by forming a similar degree of association between such reference behavior pattern information and the possibility of abuse. Then, when actually determining the possibility of child abuse, by performing image analysis in the same manner from the image information captured for the child, it is applied to various types of pre-patterned behaviors, and this is applied to the behavior. Use pattern information. The type of behavior pattern to be applied in this behavior pattern information should conform to the above-mentioned reference behavior pattern information.

Based on the behavior pattern information extracted from the image information in this way, it is possible to determine the actual possibility of abuse by referring to the degree of association formed between the reference behavior pattern information and the possibility of abuse. .. The process of determining the possibility of abuse is the same as the process of determining the possibility of abuse based on the image information shown in FIG. 3 described above.

In the example of FIG. 6, the degree of association of the combination of the reference image information and the reference attribute information is formed. The reference attribute information is information on the age, gender, etc. of the child. Since the specific content of abuse may differ depending on age and gender, the possibility of abuse should be determined with higher accuracy by combining this with learning data and making a judgment together with reference image information. Can be done. Therefore, in addition to the reference image information, the reference attribute information is combined to form the above-mentioned degree of association.

In the example of FIG. 6, it is assumed that the input data is, for example, reference image information P01 to P03 and reference attribute information P14 to 17. The intermediate node shown in FIG. 6 is a combination of reference image information and reference attribute information as such input data. Each intermediate node is further linked to the output. In this output, the possibility of abuse as an output solution is displayed.

Each combination (intermediate node) of the reference image information and the reference attribute information is associated with each other through three or more levels of association with the possibility of abuse as this output solution. The reference image information and the reference attribute information are arranged on the left side through this degree of association, and the possibility of abuse is arranged on the right side through this degree of association. The degree of association indicates the degree of possibility of abuse and the degree of relevance to the reference image information and the reference attribute information arranged on the left side. In other words, this degree of association is an index indicating what kind of abuse possibility each reference image information and reference attribute information is likely to be associated with, and from the reference image information and reference attribute information. It shows the accuracy in selecting the most probable abuse potential. Therefore, the optimum possibility of abuse is searched for by combining the reference image information and the reference attribute information.

In the example of FIG. 6, w13 to w22 are shown as the degree of association. As shown in Table 1, these w13 to w22 are shown in 10 stages, and the closer to 10 points, the higher the degree of relation between each combination as an intermediate node with the output, and conversely, 1 point. The closer they are, the less relevant each combination as an intermediate node is to the output.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. That is, the discriminating device 2 accumulates past data as to which of the reference image information, the reference attribute information, and the possibility of abuse in that case is suitable for discriminating the actual search solution. By analyzing and analyzing these, the degree of association shown in FIG. 6 is created.

For example, it is assumed that the reference image information in the actual case in the past is the image data α. Further, it is assumed that the reference attribute information is (boy, 8 years old). In such a case, the possibility of abuse, which indicates how much the possibility of abuse was actually, is learned as a data set and defined in the form of the above-mentioned degree of association. In addition, such reference image information and reference attribute information may be extracted from a management database managed by a child guidance center or the like.

This analysis and analysis may be performed by artificial intelligence. In such a case, for example, in the case of the reference image information P01 and the reference attribute information P16, the possibility of abuse is analyzed from the past data. If there are many cases of abuse possibility A, the degree of association leading to this abuse possibility A is set higher, and if there are many cases of abuse possibility B and there are few cases of abuse possibility A, abuse Set the degree of association that leads to possibility B to be high and the degree of association that leads to possibility of abuse A to be low. For example, in the example of the intermediate node 61a, the output of the abuse possibility A and the quality B is linked, but from the previous case, the degree of association of w13 that leads to the abuse possibility A is 7 points, and the connection degree of w13 that leads to the abuse possibility B is w14. The degree of association is set to 2 points.

Further, the degree of association shown in FIG. 6 may be composed of the nodes of the neural network in artificial intelligence. That is, the weighting coefficient for the output of the node of this neural network corresponds to the above-mentioned degree of association. Further, the network is not limited to a neural network, and may be composed of all decision-making factors constituting artificial intelligence. Other than that, the configuration related to artificial intelligence is the same as the description in FIG.

In the example of the degree of association shown in FIG. 6, the node 61b is a node in which the reference attribute information P14 is combined with the reference image information P01, the degree of association of the abuse possibility C is w15, and the abuse possibility E The degree of association is w16. The node 61c is a node that is a combination of the reference attribute information P15 and P17 with respect to the reference image information P02, and the degree of association of the abuse possibility B is w17 and the degree of association of the abuse possibility D is w18. ..

Such a degree of association is what is called learned data in artificial intelligence. After creating such learned data, when actually determining the possibility of abuse from now on, the above-mentioned learned data will be used. In such a case, the image information of the child who actually tries to determine the possibility of abuse and the attribute information of the child are input or selected.

Based on the newly acquired image information and attribute information in this way, the optimum possibility of abuse is searched for. In such a case, the degree of association shown in FIG. 6 (Table 1) acquired in advance is referred to. For example, when the newly acquired image information is the same as or similar to P02 and the attribute information is P17, the node 61d is associated via the degree of association, and this node In 61d, the abuse possibility C is associated with w19, and the abuse possibility D is associated with the association degree w20. In such a case, the abuse possibility C having the highest degree of association is selected as the optimum solution. However, it is not essential to select the one with the highest degree of association as the optimum solution, and the possibility of abuse D in which the degree of association is low but the association itself is recognized may be selected as the optimum solution. In addition to this, it goes without saying that an output solution to which the arrows are not connected may be selected, and as long as it is based on the degree of association, it may be selected in any other priority.

In addition, Table 2 below shows examples of the degrees of association w1 to w12 extending from the input.

The intermediate node 61 may be selected based on the degree of association w1 to w12 extending from this input. That is, the larger the degree of association w1 to w12, the heavier the weighting in the selection of the intermediate node 61 may be. However, the association degrees w1 to w12 may all have the same value, and the weightings in the selection of the intermediate node 61 may all be the same.

In addition to the above-mentioned reference image information, instead of the above-mentioned reference attribute information, a combination with the reference spectrum information and the possibility of abuse of the combination may be set to three or more levels of association. Good.

By the way, the above-mentioned reference image information is not limited to the case where the image data at a certain temporary point is captured for the child, and the image is captured a plurality of times in chronological order and the change in chronological order. You may try to get the trend. The degree of association is formed as reference image information including the change tendency of such time-series image data, and the time-series change of the image obtained by capturing the image of the child to be discriminated multiple times in time series. When the tendency is acquired, it is determined by inputting it as input data.

In FIG. 7, in addition to the above-mentioned reference image information, a combination with the reference home environment information instead of the above-mentioned reference attribute information and the possibility of abuse of the combination are set to three or more levels of association. An example is shown.

This reference home environment information, which is added as an explanatory variable instead of the reference attribute information, is any information about the home environment of the child whose reference image information is to be imaged. Reference home environment information includes the child's family structure, whether the child's mother is pregnant, whether the guardian is chasing a disability or illness, whether they live with their grandparents, and whether the care recipient is at home. Whether or not the child is affected, whether or not the parents are divorced, income (household income), whether or not there is livelihood protection, whether or not there are siblings and their age structure, whether or not the parents or one of the parents is the real child , Etc. All of these are obtained through declarations and submitted documents from each household, and in order to actually implement this program, it is a prerequisite that these reference household environment information is converted into electronic data.

Since such a home environment also affects the possibility of abuse, the accuracy of discrimination can be improved by determining the possibility of abuse through the degree of association in combination with the reference image information.

In the example of FIG. 7, it is assumed that the input data is, for example, reference image information P01 to P03 and reference home environment information P18 to 21. The intermediate node shown in FIG. 6 is a combination of reference image information and reference home environment information as such input data. Each intermediate node is further linked to the output. In this output, the possibility of abuse as an output solution is displayed.

Each combination (intermediate node) of the reference image information and the reference home environment information is associated with each other through three or more levels of association with the possibility of abuse as this output solution. The reference image information and the reference home environment information are arranged on the left side through this degree of association, and the possibility of abuse is arranged on the right side through this degree of association. The degree of association indicates the degree of possibility of abuse and the degree of relevance to the reference image information and the reference home environment information arranged on the left side. In other words, this degree of association is an index indicating what kind of abuse possibility each reference image information and reference home environment information is likely to be associated with, and is a reference image information and reference home environment. It shows the accuracy in selecting the most probable abuse potential from the information.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. 7. That is, in determining the actual search solution, the discrimination device 2 determines which of the reference image information, the reference home environment information obtained when the reference image information is acquired, and the possibility of abuse in that case. Whether it was suitable or not, the past data is accumulated, and these are analyzed and analyzed to create the degree of association shown in FIG. 7.

For example, at the time of evaluation of the actual possibility of abuse in the past, it is assumed that the reference home environment information has a history of divorce of parents and the household income is ●● 10,000 yen for a certain reference image information. In such a case, if there are many cases where the possibility of abuse is determined to be A, these are trained as a data set and defined in the form of the above-mentioned degree of association.

This analysis and analysis may be performed by artificial intelligence. In such a case, for example, when the reference image information P01 is the reference home environment information P20, the possibility of abuse is analyzed from the past data. When there are many cases of abuse possibility A, the degree of association that this abuse possibility leads to A is set higher, and when there are many cases of abuse possibility B and there are few cases of abuse possibility A, , The degree of association that the possibility of abuse leads to B is increased, and the degree of association that the possibility of abuse leads to A is set low. For example, in the example of the intermediate node 61a, the output of the abuse possibility A and the abuse possibility B is linked, but from the previous case, the degree of association of w13 leading to the abuse possibility A is set to 7 points, and the abuse possibility B is set. The degree of association of the connected w14 is set to 2 points.

Further, the degree of association shown in FIG. 7 may be composed of the nodes of the neural network in artificial intelligence. That is, the weighting coefficient for the output of the node of this neural network corresponds to the above-mentioned degree of association. Further, the network is not limited to a neural network, and may be composed of all decision-making factors constituting artificial intelligence. Other than that, the configuration related to artificial intelligence is the same as the description in FIG.

In the example of the degree of association shown in FIG. 7, the node 61b is a node that is a combination of the reference image information P01 and the reference home environment information P18, and the degree of association of the abuse possibility C is w15 and the abuse possibility E. The degree of association is w16. The node 61c is a node that is a combination of the reference home environment information P19 and P21 with respect to the reference image information P02, and the degree of association of the abuse possibility B is w17 and the degree of association of the abuse possibility D is w18. There is.

Such a degree of association is what is called learned data in artificial intelligence. After creating such learned data, when actually searching for the possibility of abuse from now on, the above-mentioned learned data will be used. In such a case, the image information to be determined for the possibility of abuse and the home environment information are actually acquired. Here, the home environment information is newly acquired for the child whose image information is to be acquired when the possibility of abuse is newly determined, and the acquisition method is the same as the above-mentioned reference home environment information.

Based on the newly acquired image information and home environment information in this way, the optimum possibility of abuse is searched for. In such a case, the degree of association shown in FIG. 7 (Table 1) acquired in advance is referred to. For example, when the newly acquired image information is the same as or similar to P02 and the home environment information is the same as or similar to P21, the node 61d is associated via the degree of association. In this node 61d, the abuse possibility C is associated with w19, and the abuse possibility D is associated with the association degree w20. In such a case, the abuse possibility C having the highest degree of association is selected as the optimum solution. However, it is not essential to select the one with the highest degree of association as the optimum solution, and the possibility of abuse D in which the degree of association is low but the association itself is recognized may be selected as the optimum solution. In addition to this, it goes without saying that an output solution to which the arrows are not connected may be selected, and as long as it is based on the degree of association, it may be selected in any other priority.

In FIG. 8, in addition to the above-mentioned reference image information, a combination with the reference listening information instead of the above-mentioned reference attribute information and the possibility of abuse of the combination are set to three or more levels of association. An example is shown.

This reference listening information, which is added as an explanatory variable instead of the reference attribute information, is a data of what was actually heard through an interview from a child who captured the reference image information, and was violent, for example. A number of abuses, such as being ranted, being taken out, etc., are categorized. Since the content of listening included in such reference listening information also affects the possibility of abuse, it is possible to improve the discrimination accuracy by determining the possibility of abuse through the degree of association in combination with the reference image information. it can.

In the example of FIG. 8, it is assumed that the input data is, for example, reference image information P01 to P03 and reference listening information P18 to 21. The intermediate node shown in FIG. 8 is a combination of reference listening information and reference image information as such input data. Each intermediate node is further linked to the output. In this output, the possibility of abuse as an output solution is displayed.

Each combination (intermediate node) of the reference image information and the reference listening information is associated with each other through three or more levels of association with the possibility of abuse as this output solution. The reference image information and the reference listening information are arranged on the left side through this degree of association, and the possibility of abuse is arranged on the right side through this degree of association. The degree of association indicates the degree of possibility of abuse and the degree of relevance to the reference image information and the reference listening information arranged on the left side. In other words, this degree of association is an index indicating what kind of abuse possibility each reference image information and reference listening information is likely to be associated with, and from the reference image information and reference listening information. It shows the accuracy in selecting the most probable abuse potential.

The discriminating device 2 acquires in advance the degree of association w13 to w22 of three or more stages shown in FIG. That is, in discriminating the actual search solution, the discriminating device 2 prefers the reference image information, the reference listening information obtained when the reference image information is acquired, and the possibility of abuse in that case. Or, by accumulating past data and analyzing and analyzing these, the degree of association shown in FIG. 7 is created.

This analysis and analysis may be performed by artificial intelligence. In such a case, for example, when the reference image information P01 is the reference listening information P20, the possibility of abuse is analyzed from the past data. When there are many cases of abuse possibility A, the degree of association that this abuse possibility leads to A is set higher, and when there are many cases of abuse possibility B and there are few cases of abuse possibility A, , The degree of association that the possibility of abuse leads to B is increased, and the degree of association that the possibility of abuse leads to A is set low. For example, in the example of the intermediate node 61a, the output of the abuse possibility A and the abuse possibility B is linked, but from the previous case, the degree of association of w13 leading to the abuse possibility A is set to 7 points, and the abuse possibility B is set. The degree of association of the connected w14 is set to 2 points.

Further, the degree of association shown in FIG. 8 may be composed of the nodes of the neural network in artificial intelligence. That is, the weighting coefficient for the output of the node of this neural network corresponds to the above-mentioned degree of association. Further, the network is not limited to a neural network, and may be composed of all decision-making factors constituting artificial intelligence. Other than that, the configuration related to artificial intelligence is the same as the description in FIG.

In the example of the degree of association shown in FIG. 8, the node 61b is a node in which the reference listening information P18 is combined with the reference image information P01, the degree of association of the abuse possibility C is w15, and the association of the abuse possibility E. The degree is w16. The node 61c is a node in which the reference listening information P19 and P21 are combined with the reference image information P02, and the degree of association of the abuse possibility B is w17 and the degree of association of the abuse possibility D is w18. ..

Such a degree of association is what is called learned data in artificial intelligence. After creating such learned data, when actually searching for the possibility of abuse from now on, the above-mentioned learned data will be used. In such a case, the image information to be determined for the possibility of abuse and the listening information are actually acquired. Here, the hearing information is a data of the content of the abuse heard from the child who actually judges the possibility of abuse, and may be applied to the content of the abuse categorized in advance. The method of acquiring the listening information is the same as the above-mentioned reference listening information. The listening information and the listening information for reference may be acquired by inputting the keyboard to a device such as a PC or a smartphone, or by analyzing the character information in which the content of the interview from the child is described.

Based on the newly acquired image information and listening information in this way, the optimum possibility of abuse is searched for. In such a case, the degree of association shown in FIG. 8 (Table 1) acquired in advance is referred to. For example, when the newly acquired image information is the same as or similar to P02 and the listening information is the same as or similar to P21, the node 61d is associated via the degree of association. The node 61d is associated with the abusive possibility C by w19 and the abusive possibility D by the degree of association w20. In such a case, the abuse possibility C having the highest degree of association is selected as the optimum solution. However, it is not essential to select the one with the highest degree of association as the optimum solution, and the possibility of abuse D in which the degree of association is low but the association itself is recognized may be selected as the optimum solution. In addition to this, it goes without saying that an output solution to which the arrows are not connected may be selected, and as long as it is based on the degree of association, it may be selected in any other priority.

In the above-mentioned degree of association, the degree of association is expressed by a 10-step evaluation, but it is not limited to this, and it may be expressed by a degree of association of 3 or more levels, and conversely, it may be expressed by 3 or more levels. For example, 100 steps or 1000 steps may be used. On the other hand, this degree of association does not include those expressed in two stages, that is, whether or not they are related to each other, either 1 or 0.

According to the present invention having the above-described configuration, anyone can easily determine and search for the possibility of abuse without any special skill or experience. Further, according to the present invention, it is possible to determine the search solution with higher accuracy than that performed by a human being. Further, by configuring the above-mentioned degree of association with artificial intelligence (neural network or the like), it is possible to further improve the discrimination accuracy by learning this.

Since there are many cases where the input data and the output data described above do not exist exactly the same in the process of learning, the input data and the output data may be classified by type. That is, the information P01, P02, ... P15, 16, ... That constitute the input data are classified according to the criteria classified in advance on the system side or the user side according to the content of the information, and the classified inputs. A data set may be created between the data and the output data and trained.

In the above-mentioned degree of association, in addition to the reference image information, an example is taken in the case where the reference behavior pattern information, the reference attribute information, the reference home environment information, and the reference listening information are combined. I explained, but it is not limited to this. That is, the degree of association may be composed of a combination of any two or more of the reference behavior pattern information, the reference attribute information, the reference home environment information, and the reference listening information in addition to the reference image information. In addition to the reference image information, the degree of association is one or more of the reference behavior pattern information, the reference attribute information, the reference home environment information, and the reference listening information, and other factors are added to this combination. The degree of association may be formed.

In either case, data is input according to the reference information of the degree of association, and the possibility of abuse is calculated using the degree of association.

Further, as shown in FIG. 9, the present invention determines the possibility of abuse based on the degree of association between two or more types of information, reference information U and reference information V. The reference information Y is the reference image information, and the reference information V is the reference attribute information, the reference home environment information, the reference listening information, the reference biological information, the reference breeding environment information, and the reference food information. It shall be either.

At this time, as shown in FIG. 9, the output obtained for the reference information U is used as input data as it is, and is associated with the output (possibility of abuse) via the intermediate node 61 in combination with the reference information V. May be good. For example, for reference information U (reference image information), after an output solution is output as shown in FIG. 3, this is used as an input as it is, and the degree of association with other reference information V is used. You may want to search for output (potential abuse).

Further, according to the present invention, instead of obtaining the possibility of abuse as an output solution, it is possible to transmit alert information such as an alarm, an alarm, etc. based on the possibility of abuse. The higher the possibility of abuse, the higher the degree of alerting of the alerting information. As a result, it is possible to efficiently alert the outside that the child is in a dangerous state.

Further, according to the present invention, it is characterized in that an optimum solution search is performed through the degree of association set in three or more stages. The degree of association can be described by, for example, a numerical value from 0 to 100% in addition to the above-mentioned 10 steps, but the degree of association is not limited to this, and any step can be described as long as it can be described by a numerical value of 3 or more steps. It may be configured.

By determining the most probable abuse possibility based on the degree of association expressed by the numerical values of three or more levels, the degree of association is high in a situation where there are multiple possible candidates for the search solution. It is also possible to search and display in order. If the users can be displayed in descending order of the degree of association in this way, it is possible to preferentially display more probable search solutions.

In addition to this, according to the present invention, it is possible to judge without overlooking the discrimination result of the output having an extremely low degree of association such as 1%. Remind the user that even a judgment result with an extremely low degree of association is connected as a slight sign, and may be useful as the judgment result once every tens or hundreds of times. be able to.

Further, according to the present invention, there is an advantage that the search policy can be determined by the method of setting the threshold value by performing the search based on the degree of association of three or more stages. If the threshold value is lowered, even if the above-mentioned degree of association is 1%, it can be picked up without omission, but it is unlikely that a more appropriate discrimination result can be detected favorably, and a lot of noise may be picked up. is there. On the other hand, if the threshold value is raised, there is a high possibility that the optimum search solution can be detected with high probability, but the degree of association is usually low and it is passed through, but it is suitable to appear once in tens or hundreds of times. Sometimes the solution is overlooked. It is possible to decide which one should be emphasized based on the ideas of the user side and the system side, but it is possible to increase the degree of freedom in selecting the points to be emphasized.

Further, in the present invention, the above-mentioned degree of association may be updated. This update may reflect information provided via a public communication network such as the Internet. In addition, when each reference information such as reference image information is acquired and knowledge, information, and data regarding the possibility of abuse and improvement measures are acquired, the degree of association is increased or decreased according to these.

In other words, this update corresponds to learning in the sense of artificial intelligence. It can be said that it is a learning act because it acquires new data and reflects it in the learned data.

In addition, this update of the degree of association is not based on information that can be obtained from the public communication network, but is also updated by the system side or the user side based on the contents of research data, papers, conference presentations, newspaper articles, books, etc. by experts. It may be updated artificially or automatically. Artificial intelligence may be utilized in these update processes.

Further, the process of first creating the trained model and the above-mentioned update may use not only supervised learning but also unsupervised learning, deep learning, reinforcement learning, and the like. In the case of unsupervised learning, instead of reading and learning the data set of input data and output data, information corresponding to the input data is read and trained, and the degree of association related to the output data is self-formed from there. You may let it.

1 Child abuse sign discrimination system 2 Discrimination device 21 Internal bus 23 Display unit 24 Control unit 25 Operation unit 26 Communication unit 27 Discrimination unit 28 Storage unit 61 node

Claims (7)

In a child abuse sign determination program that determines signs of parental abuse of a child
An information acquisition step in which the face, limbs, and clothes are photographed multiple times in chronological order by targeting each child to be discriminated , and the change tendency of the image in chronological order is acquired as image information .
In the past, the face, limbs, and clothes were imaged multiple times in time series by targeting each child, and reference image information consisting of the change tendency of the time series image and abuse of the child. Based on the reference image information according to the image information acquired in the above information acquisition step, using the degree of association with the possibility of 3 or more levels, the computer is provided with a discrimination step for determining the possibility of child abuse. A child abuse symptom determination program characterized by running. In the above information acquisition step, as the above image information, the state of scratches, swelling, bruises, or dirt on clothes of the face or limbs is detected.
In the determination step, as the reference image information, the state of scratches, swelling, bruises, or stains on the face or limbs according to the image information is detected, and the possibility of child abuse against these is detected. The child abuse sign determination program according to claim 1, wherein the degree of association of three or more levels is used. In the above information acquisition step, a moving image of the child to be discriminated is imaged a plurality of times in chronological order, and a type of behavior pattern in which signs of abuse of the child appear from the image information consisting of the change tendency of the image in chronological order. Generates behavior pattern information applied to
In the above discrimination step, a moving image of a child is imaged multiple times in a time series in the past, and a type of behavior pattern in which signs of abuse of the child appear from reference image information consisting of a change tendency of the image in the time series. Based on the above behavior pattern information applied from the image information acquired in the above information acquisition step, using the three or more levels of association between the reference behavior pattern information applied to and the possibility of abuse of the child, the child The child abuse sign determination program according to claim 1, wherein the possibility of abuse of the child is determined. In the above information acquisition step, the attribute information regarding the child to be discriminated is further acquired.
In the determination step, there are three or more stages of association between the combination of the reference image information, the reference attribute information about the child captured in obtaining the reference image information, and the possibility of abuse of the child. The child abuse sign according to claim 1 or 2, which is characterized in that the possibility of child abuse is determined based on the reference attribute information according to the attribute information acquired in the above information acquisition step. Discrimination program. The above information acquisition step further acquires home environment information regarding the home environment of the child to be discriminated.
In the above-mentioned determination step, the combination having the above-mentioned reference image information and the reference home environment information regarding the child's home environment captured in obtaining the reference image information, and the possibility of abuse of the child 3 Claim 1 or claim 1, which is characterized in that the possibility of child abuse is determined based on the reference home environment information according to the home environment information acquired in the above information acquisition step by using the degree of association of one level or higher. 2 The child abuse sign discrimination program described. The above information acquisition step further acquires hearing information regarding the content of abuse heard from the child to be discriminated.
In the determination step, the combination having the reference image information, the reference listening information regarding the abuse content heard from the child captured in obtaining the reference image information, and the possibility of the child being abused. Claim 1 or 2 characterized in that the possibility of child abuse is determined based on the reference listening information corresponding to the listening information acquired in the above information acquisition step by using three or more levels of association. The listed child abuse sign discrimination program. The child abuse sign discrimination program according to any one of claims 1 to 6, wherein in the discrimination step, the degree of association corresponding to the weighting coefficient of each output of the node of the neural network in artificial intelligence is used. ..

Images