JP2019125061A - System for estimating condition of person - Google Patents

Abstract

To provide a system for estimating a condition of a person, which can accurately estimate a sensible temperature or an emotion of the person.SOLUTION: In order to solve the problem above, a condition estimation system in one aspect of this invention is a system for estimating a condition of at least one of a sensible temperature and an emotion of a person and comprises: a behavior recognition unit which acquires an image of a person and extracts movement of his or her body from the acquired image to recognize his or her behavior; and a condition estimation unit which estimates a condition of at least one of a sensible temperature and an emotion of the person on the basis of the recognized behavior.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a human state estimation system.

  An air conditioner is known which controls the operation of the air conditioner according to the state of a person. For example, Patent Document 1 describes a vehicle air conditioner that controls the operation of a vehicle air conditioner according to the temperature of the occupant's face. In addition, Patent Document 2 describes a technique for performing fashion coordination in accordance with the environment outside the house. In addition, Patent Document 3 describes an indoor environment control device that measures personnel density and illuminance with an image analysis device, detects a temperature of a vehicle with a temperature sensor, and adjusts the operation based on the detection result. . Further, Patent Document 4 describes a technology for analyzing emotion response data obtained from a user's face image in relation to an advertisement.

Unexamined-Japanese-Patent No. 2006-137364 JP, 2014-013560, A Unexamined-Japanese-Patent No. 08-193727 gazette Special publication 2017-507434 gazette

  In order to provide a comfortable air conditioning environment, it is desirable to adjust the temperature, the wind speed, the wind direction, and the like of the air conditioner according to the user's perceived temperature and how it feels. The air conditioner described in Patent Document 1 is configured to increase or decrease the air blowing amount according to the temperature of the face of the occupant based on the data of the thermal image captured by the IR camera. However, it has been difficult to accurately estimate a person's perceived temperature and how to feel it from the face temperature alone, and there is a problem that a comfortable air conditioning environment can not be provided.

  The present invention has been made in view of these problems, and an object of the present invention is to provide a human state estimation system capable of accurately estimating the bodily sensation temperature and emotion of a human being.

  In order to solve the above problems, a human state estimation system according to an aspect of the present invention is a system that estimates a state of at least one of a human's body temperature and emotion, and acquires an image of a human and acquires the image. An action recognition unit that extracts the movement of the person's body from the image and recognizes the action of the person; and a state estimation unit that estimates the state of at least one of the temperature and emotion of the person based on the recognized action And.

  According to this aspect, it is possible to estimate at least one of the body temperature and emotion of the person based on the action of the person recognized according to the movement of the body. By controlling the target device based on these estimation results, it is possible to control the device to an appropriate setting. This estimation result can be used as a setting of an air conditioner or a control reference of various types of devices and systems.

  Another aspect of the present invention is also a human state estimation system. This state estimation system is a system that estimates the state of at least one of a person's perceived temperature and emotion, acquires an image of a person, and based on the acquired image, the movement, body shape, moving speed, and the like of the person's body. An activity recognition unit that extracts at least one or more clothes and recognizes an activity of the person; and a state estimation unit that estimates at least one of the temperature of the person's body temperature and emotion based on the recognized activity. Prepare.

  According to this aspect, in addition to the movement of the human body, the temperature and feeling of the person's body temperature, emotion based on the person's behavior recognized according to the person's body type, moving speed, and at least one or more extracted information of clothes. Since at least one of the states is estimated, it is possible to more accurately estimate at least one of the perceived temperature of the person and the emotion.

  Yet another aspect of the present invention is also a human condition estimation system. This state estimation system is a system for estimating the state of at least one of the temperature and emotion of the group of people, acquiring an image of the group of people, and extracting the movement of the group of people from the acquired image. And a state estimation unit that estimates a state of at least one of a sock temperature and an emotion of the group based on the recognized behavior.

  According to this aspect, it is possible to estimate at least one of the sensory temperature and the emotion of the group of people based on the group of people's behavior. For example, at least one of the sensible temperature and emotion of a group of people accommodated in the entire building can be estimated, and the air conditioner can be controlled to an appropriate setting according to the estimation result.

  It should be noted that any combination of the above-described components, or any of the components or expressions of the present invention mutually replaced by a method, an apparatus, a program, a temporary or non-temporary storage medium recording a program, a system, etc. Are also effective as an aspect of the present invention.

  According to the present invention, it is possible to provide a state estimation system capable of accurately estimating the sensible temperature of a person.

It is a block diagram showing an example of a sensible temperature estimation system concerning a 1st embodiment. It is a block diagram showing an example of a sensible temperature estimation system concerning a 2nd embodiment. It is a block diagram showing an example of a power consumption prediction system concerning an application example. It is a block diagram which shows the example which applied the power consumption prediction system of FIG. 3 to a building.

  Hereinafter, the present invention will be described based on preferred embodiments with reference to the drawings. In the embodiment and the modification, the same or equivalent constituent elements and members are denoted by the same reference numerals, and duplicating descriptions will be appropriately omitted. In addition, dimensions of members in each drawing are shown appropriately enlarged or reduced for easy understanding. In each drawing, a part of members which are not important in describing the embodiment is omitted and displayed.

First Embodiment
A first embodiment of the present invention is a system for estimating the sensible temperature of a person. FIG. 1 is a block diagram showing an example of a sensory temperature estimation system 100 according to the first embodiment. In the first embodiment, a sensible temperature estimation system 100 is illustrated as a human state estimation system. In this example, the sensible temperature estimation system 100 can use the estimation result of the sensible temperature as a reference for controlling the setting of the air conditioner. The estimation result of the sensible temperature estimation system 100 (hereinafter simply referred to as “estimation result”) can be used to control not only the air conditioner but various devices. Further, according to the sensible temperature estimation system 100, it is possible to estimate an emotion such as a person's sense of tension or calmness based on a change in the person's sensible temperature. In addition, it is possible to directly estimate human emotion based on the behavior of the person recognized according to the movement of the human body. Emotions estimated based on changes in the sock temperature and emotions estimated directly based on human behavior (hereinafter referred to as "estimated emotions") can be used to control various devices.

  For example, by utilizing the estimated emotion for controlling the projection mapping device, the projection mapping device can provide an image that imparts a gentle feeling of color to the person viewing it. Also, for example, by using the estimated emotion to control the BGM device, the BGM device can provide music that feels comfortable to the person who is listening. Also, for example, by utilizing the estimated emotion for controlling the scent generating device, the scent generating device can provide a scent that feels comfortable to a person who is touching the scent. Also, for example, by utilizing the estimated emotion for controlling the vibration generating device, the vibration generating device can provide the vibration that feels comfortable to the person receiving the vibration. The control of the vibration generator may be to control 1 / f fluctuation of the vibration.

  That is, according to the present invention, by using the estimated emotion for controlling the device for exerting various actions on the person, the device outputs an action that feels comfortable to the person receiving the action. Can. This idea includes control to feed back the estimated emotion to the operation of the device to make the output action of the device more comfortable.

  The sensational temperature estimation system 100 mainly includes an appearance / action recognition unit 10, a sensational temperature estimation unit 12, an estimation result output unit 14, an action DB 16, and an image sensor 20. In the present specification, "DB" is used in the meaning of "database". The sensible temperature estimation system 100 controls the target device 30 according to the movement of the body of a person (hereinafter, referred to as a “target person”) that is the target of estimation. The image sensor 20 detects the appearance and behavior of the subject, and outputs the detection result to the appearance / action recognition unit 10. In this example, the image sensor 20 includes a zoom mechanism (not shown), and can detect the condition of the subject in detail as needed. The zoom mechanism may be controlled by the appearance / action recognition unit 10.

(Appearance / action recognition department)
The appearance / action recognition unit 10 will be described. In this example, the appearance / action recognition unit 10 is illustrated as the action recognition unit. The appearance / action recognition unit 10 recognizes the appearance and behavior (hereinafter referred to as “recognition behavior”) of the subject from the image acquired by the image sensor 20. Cognitive behavior can be extracted based on an image of a subject, and the motion of the subject can be extracted, and the appearance and behavior of the subject can be recognized by an artificial intelligence capable of recognizing. When there are a plurality of subjects, the motion of each person may be extracted to recognize the recognition action.

(Sensible temperature estimation unit)
The sensory temperature estimation unit 12 will be described. In this example, the sensible temperature estimation unit 12 is illustrated as the state estimation unit. The sensible temperature estimation unit 12 estimates the sensible temperature of the subject based on the action pattern information stored in the action DB 16 based on the recognition action recognized by the appearance / action recognition unit 10.

(Action DB)
The behavior DB 16 stores, as behavior pattern information, correlations between image and moving image data of behavior to be performed when a person feels uncomfortable with air temperature, and sensation temperature, for a large number of samples. When the person feels uncomfortable with the temperature, look at the fan, take a coat, drink a drink, eat ice cream, wear a hat, blow a hand, turn your back, put your hand in a pocket, etc. Actions are included. Behavior pattern information can be constructed by collecting and analyzing the correlation between human behavior and sensible temperature for a large number of samples. The behavior DB 16 includes behavior pattern information constructed in this way as a neural network of a plurality of layers.

  The estimation result output unit 14 outputs the estimation result estimated by the sensible temperature estimation unit 12. The setting of the target device 30 is changed based on the estimation result from the estimation result output unit 14. As described above, the target device 30 includes an air conditioner that outputs warm air or cold air whose temperature is adjusted, a projection mapping device that produces a gentle feeling by color, a fragrance generating device, and a 1 / f fluctuation mode, and a comfortable vibration It may be a vibration generating device or the like that generates The target device 30 is a device that outputs a pleasant image such as a fun picture when estimating the frustrated emotions of the target person, a color therapy device in which the user sees colors and the mind settles, an aroma scent generator, a BGM device, VR ( It may be a device or the like where mist strikes by virtual reality). The target device 30 can be given a comfortable feeling to the target person by being controlled based on the estimation result.

  According to the sensible temperature estimation system 100 of the first embodiment configured as described above, it is possible to control the air conditioner to an appropriate setting by estimating the sensible temperature of the person based on the action of the person. The action DB 16 may accumulate actions to be taken when a person feels uncomfortable. The action taken when a person feels uncomfortable may include the action taken when being frustrated.

Second Embodiment
Next, a sensory temperature estimation system according to a second embodiment of the present invention will be described with reference to FIG. FIG. 2 is a block diagram showing an example of the configuration of a sensory temperature estimation system 200 according to the second embodiment. In the second embodiment, a sensible temperature estimation system 200 is illustrated as a human state estimation system. The sensational temperature estimation system 200 mainly includes an appearance and behavior recognition unit 10, a sensational temperature estimation unit 12, an estimation result output unit 14, a behavior DB 16, a clothes and possessions DB 18, and an image sensor 20. The sensible temperature estimation system 200 controls the target device 30 in accordance with the movement, body shape, moving speed, and clothes of the subject's body. Here, duplicate explanations are omitted, and points different from the first embodiment will be mainly described.

(Clothing, possessions DB)
The clothing and belongings DB 18 will be described. The clothing and belongings DB 18 stores, as clothing information, the correlation between the temperature of the clothes and belongings and the image / moving image data when the person feels uncomfortable with the temperature for many samples. In particular, the clothes and belongings DB 18 is accumulated as a set of behavior and clothes, and even people who are not thin and feel cold can be recognized. In particular, the clothes and belongings DB 18 stores images of clothes and belongings that may be used according to the season, temperature, humidity, weather, and body type. The clothing and belongings DB 18 may be configured to automatically update data from information on a maker, a fashion magazine, and an EC site using a communication line such as the Internet. Clothes information can be constructed by collecting and analyzing the correlations for a large number of samples. The clothes and belongings DB 18 includes clothes information constructed in this manner as a neural network of a plurality of layers.

  In addition to the extraction target of the first embodiment, the appearance / action recognition unit 10 extracts the body type, moving speed, clothes, belongings, etc. of the object person, and recognizes them using artificial intelligence.

  In addition to the operation of the first embodiment, the bodily sensation temperature estimation unit 12 senses the sensation of the subject based on the movement, body shape, clothes, belongings, etc. of the subject recognized by the appearance / action recognition unit 10. Estimate the temperature.

  The estimation result output unit 14 may present the estimation result to the outside in addition to the operation of the first embodiment. As one example, the estimation result output unit 14 may present the estimation result to the specific area via the Internet as reference information at the time of going out of the specific area. This reference information at the time of going out may include recommendation information on clothes.

Third Embodiment
Next, a body temperature estimation system according to a third embodiment of the present invention will be described. In the first and second embodiments, an example is described in which the behavior and the like are identified for one or a plurality of persons, respectively, and the bodily sensation temperature is estimated, but the present invention is not limited to this. . The human state estimation system according to the third embodiment is a system for estimating the state of at least one of the sensory temperature and emotion of a group of persons. The system acquires an image of a group of people, extracts a movement of the group of people from the acquired image, and recognizes an action of the group of people; And a state estimation unit configured to estimate a state of at least one of the sensational temperature and the emotion of a group of people. According to the third embodiment, for example, it is possible to estimate at least one of the sensible temperature and the emotion of a group of people accommodated in the entire building, and control the air conditioner to an appropriate setting according to the estimation result.

  In the third embodiment, the group of persons may be, for example, a crowd housed in a building. By estimating the air conditioning state of the entire building based on the behavior of the crowd, the air conditioning equipment can be controlled to an appropriate setting. One group of people's actions are, for example, people gathering in front of a cooler store. If a group of people gather in front of the store without shopping, the group of people feel cool and it can be estimated that air conditioning of the entire building is appropriate. In addition, even when a group of people gather in the shade, it is possible to estimate that the group of people have a high sensation temperature and are in an emotional state requiring coolness. On the other hand, when a group of people gather in the sun, it can be estimated that the group of people have a low sensation temperature and are in an emotional state for warmth.

(Modification 1)
The state estimation unit may estimate the understanding level / favorability of the person based on the facial expression or the behavior of the person recognized by the behavior recognition unit. In this case, it is possible to appropriately change the advertisement or the like by estimating the understanding level or preference of the advertisement or the like based on the human expression or behavior and feeding back the estimation result.

(Modification 2)
The state estimation unit may extract or estimate the lost person based on the action of the person recognized by the action recognition unit. Such actions include actions that are going back and forth in the same place and actions that look down while watching a smart phone. In this case, by presenting necessary information to the person, it is possible to present information for guiding the person to the destination. There are footsteps that are unique to that person, so you know where you are coming from.

(First application example)
The first application example will be described with reference to FIGS. 3 and 4. FIG. 3 is a block diagram showing an example of the power consumption prediction system 300 according to the first application example. FIG. 4 is a block diagram showing an example in which the power consumption prediction system 300 is applied to a building. The power consumption prediction system 300 is a system that mainly includes the illuminance sensor 62, the estimation unit 52, and the database 54, and estimates the power consumption of the building 70 with a certain accuracy without using a power meter, for example. The first application has the advantage of being able to predict power consumption in situations where a power meter can not be used.

  The illumination sensor 62 may be installed, for example, on an automatic door 60 provided at each opening of the building 70. In this example, the illuminance sensor 62 is installed on the blind indoor side and the outdoor side of the automatic door 60, respectively. The illumination sensor 62 may be provided in the vicinity of the automatic door sensor or integrally with the automatic door sensor. The illuminance sensor 62 acquires the illuminance at night inside and outside the building 70 and transmits the illuminance to the estimation unit 52. As shown in FIG. 4, the illuminance sensor 62 is provided for each of the automatic doors 60 provided at each entrance of the building, and can detect the illuminance of the entire building 70 outside and inside the room.

  The database 54 contains information on the correlation between the building attributes, the indoor / outdoor temperature and humidity, and the illuminance at night (hereinafter referred to as "building attributes etc.") and the power consumption of the entire building (hereinafter referred to as " The building pattern information is stored (stored). Building pattern information can be constructed by collecting and analyzing correlations between building attributes and the like and power consumption of the entire building for a large number of samples. The database 54 includes building pattern information constructed in this way as a multi-layered neural network. The power consumption of the entire building may be actual data measured by a power meter or may be obtained by calculation. The attributes of the building include small-scale shops such as detached houses, apartments and convenience stores, large-scale commercial facilities, buildings, types of buildings such as factories and office buildings, floor area, area, altitude, density of buildings, etc.

  The estimation unit 52 estimates the power consumption of the building by artificial intelligence based on the building pattern information of the database 54 based on the attributes of the building and the like and the information acquired from the illuminance sensor 62.

  The power consumption prediction system 300 configured in this manner extracts the power consumption associated with the data similar to the attribute of the target building 70, the indoor / outdoor temperature and humidity, and the illuminance at night from the database 54. The power consumption of can be estimated with a certain accuracy without using a power meter.

  The power consumption prediction system 300 can estimate the power consumption with a certain accuracy not only for one building but also for a plurality of buildings. Also, this system can estimate the power consumption with a certain accuracy even for a group of buildings included in a certain area. Furthermore, a plurality of power consumption prediction systems 300 can be connected to a server via a network, and power consumption can be estimated with a certain accuracy even for a building over a wide area.

(Second application example)
Next, a second application example will be described. The second application example is a communication speed prediction system for a building. This communication speed prediction system can estimate the actual communication speed with a certain accuracy, without using the communication traffic meter for the communication speed of the building. The second application has the advantage of being able to predict the communication speed in situations where communication traffic meters can not be used. The second application example mainly includes a communication speed estimation unit and a communication speed database.

  The communication speed database is information on the correlation between building attributes, communication methods, weather, day of the week, holidays, time (hereinafter referred to as "building characteristic information") and communication speed (hereinafter referred to as "building communication speed" Information is stored (stored). Building communication speed information can be constructed by collecting and analyzing the correlation between building characteristic information and the communication speed of the entire building for a large number of samples. The communication speed database comprises building communication speed information constructed in this way as a multi-layer neural network. The communication speed of the entire building may be actual data measured by a communication traffic meter, or may be obtained by calculation. The attributes of the building include small-scale shops such as detached houses, apartments and convenience stores, large-scale commercial facilities, buildings, types of buildings such as factories and office buildings, floor area, area, altitude, density of buildings, etc. Communication methods include wired, wireless, public line and other methods.

  The communication speed estimation unit estimates the communication speed of the building by artificial intelligence based on the building communication speed information of the communication speed database based on data similar to the attribute of the target building, weather, day of the week, holidays, and time.

  The communication speed prediction system configured in this way extracts the communication speed of the building by extracting the communication speed associated with the data of the target building's attributes, weather, day of the week, holidays, and time from the communication speed database. It can be estimated with a certain accuracy without using a communication traffic meter.

  This communication speed prediction system can estimate the communication speed with a certain accuracy not only for one building but also for a plurality of buildings. Also, this system can estimate the communication speed with a certain accuracy even for a group of buildings included in a certain area. Furthermore, a plurality of communication speed prediction systems can be connected to a server via a network, and communication speeds can be estimated with a certain accuracy even for a building over a wide area.

  This system can estimate the communication speed of a building to live from now, so it can be suitably used by people searching for a building and a real estate agent. Information such as traffic that is concentrated on the network can not be provided to new people living in the home. Artificial intelligence techniques such as machine learning and deep learning can be used to recognize building attributes.

  The present invention has been described above based on the embodiments. It is understood by those skilled in the art that each embodiment is an exemplification, and that various modifications can be made to their respective components or combinations of respective processing processes, and such modifications are also within the scope of the present invention. is there. The contents of each embodiment do not limit the technical scope of the present invention, and many design changes such as changes, additions, and deletions of components can be made without departing from the concept of the invention defined in the claims. Is possible.

  10 · · · appearance · behavior recognition unit, 12 · · · sense of temperature temperature estimation unit, 14 · · · estimation result output unit, 16 · · action DB, 18 · · clothing · possession goods DB, 20 · · · image sensor, 30 · · · Equipment · 52 · · estimation unit · 54 · · · database · 60 · · automatic door, 62 · · illuminance sensor 70 · · · building · 100 · sense of temperature estimation system, · · · sense of temperature estimation system, · · · consumption 300 Power forecasting system.

Claims (3)

A system for estimating a state of at least one of human body temperature and emotion,
An action recognition unit that acquires an image of a person, extracts movement of the body of the person from the acquired image, and recognizes an action of the person;
A state estimation unit configured to estimate at least one of the sensory temperature and emotion of the person based on the recognized action;
A human state estimation system comprising: A system for estimating a state of at least one of human body temperature and emotion,
An action recognition unit that acquires an image of a person, extracts at least one of the movement, body shape, moving speed, and clothes of the person's body from the acquired image, and recognizes the action of the person;
A state estimation unit configured to estimate at least one of the sensory temperature and emotion of the person based on the recognized action;
A human state estimation system comprising: What is claimed is: 1. A system for estimating at least one of the temperature and emotion of a group of people, the system comprising:
An action recognition unit that acquires an image of a group of people, extracts a movement of the group of people from the acquired image, and recognizes an action of the group of people;
A state estimation unit that estimates the state of at least one of the temperature and emotion of the group of people based on the recognized action;
A human state estimation system comprising:

Images