JP2006212451A - Virtual pet device and control program recording medium for the same - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To build a specific relationship between a user and a virtual pet to provide a user with a real feeling like the user has an actual pet. <P>SOLUTION: Sound inputted from a microphone 22 is recognized by a recognition section 20, and a person who speaks is identified by referring to a speaker database 12. Predetermined words are registered in the database 12, and events corresponding to the words are stored in an event list 31. When the spoken words meet event generation conditions, an event corresponding to the recognized words (barking, tail wagging, or the like) is performed. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a pet-type game machine using a pet-type toy imitating the shape of an animal or the like, or a display imitating the shape of an animal or the like.

There is a pet-type toy equipped with a mechanism for moving the mouth, tail, etc., inside the stuffed animal body imitating an animal.
There is also known a portable electronic pet toy that is reared by artificially feeding a pet displayed as an image on a display unit.

  However, in the conventional stuffed toy pet toy, there are various forms that operate in response to a specific operation using a battery, a mainspring or the like as a power source, but these mechanically repeat simple reactions. It wasn't too much.

  In addition, in electronic pet toys that have become popular in recent years, there is a sense of unity between the owner and the virtual pet that is inside because of the portable type, but if the toy itself is in the hands of another person, a specific breeder The concept of (owner) is ambiguous.

  In such a conventional pet-type toy, it is not possible to experience the exchange of emotions as seen in the relationship between the actual (real) pet and the owner, and the operator or the owner and the virtual pet There was a problem that a specific relationship was not formed.

  The present invention solves the above-described problems in conventional pet-type toys, forms a specific relationship between the operator or the owner and the virtual pet, and can obtain a sense closer to a real pet. A device is provided.

  The present invention also provides a recording medium storing a control program for the virtual pet device.

  A virtual pet device according to the present invention is a virtual pet device configured to appropriately respond to an external stimulus, and includes an input unit that inputs an external stimulus, and an input from the input unit. Recognizing means for recognizing a stimulus, and control means for controlling the response operation of the apparatus based on the recognition result of the recognizing means, the control means analyzing the inputted stimulus and inputting the stimulus It is characterized by identifying the individual who made it. According to the present invention, by analyzing a stimulus by an operator, an individual who has generated the stimulus is identified, and by establishing a specific relationship between the virtual pet and the operator, a real pet between the two It is possible to produce a dense relationship similar to the relationship between the owner and the owner.

The input means may be a voice input device. In that case, the operator can be identified by a voice, for example, a phrase, uttered by the operator.
The input means can be an image input device. In that case, for example, the operator can be identified by the face image of the operator.

  Furthermore, it is provided so as to be able to identify an unspecified number of stimulus input persons, and has storage means for storing the recognition frequency for each input person corresponding to the input, and performs a response operation to the same input according to the input person and the recognition power It is preferable that the configuration can be changed. With this configuration, it is possible to show different reactions even for the same word or phrase depending on, for example, the recognition frequency of the word or the operator, and a more realistic virtual pet can be provided.

  In addition, it is preferable to have a combination of a plurality of preset basic actions and words corresponding to each action, and to be able to newly set a relationship with words corresponding to a series of actions combining the basic actions. . Thereby, the correspondence between a plurality of combinations of basic actions and a new word / phrase can be set, and a more complex response action becomes possible. For this reason, it is possible to obtain a feeling as if an actual pet is being prepared.

  The virtual pet device according to the present invention is a virtual pet device configured to appropriately respond to an external stimulus, and includes an input unit for inputting an external stimulus, and an input from the input unit. Input person recognizing means for recognizing the input person based on the stimulated stimulus, and input person storage means for storing the input person recognized by the input person recognizing means together with the recognition degree, and based on the recognition degree The event to be executed is changed in response to the input stimulus.

  The virtual pet device according to the present invention is a virtual pet device configured to respond to an external stimulus, and includes a voice input means for inputting voice, and a voice input from the voice input means. Speaker recognition means for recognizing the speaker based on the speaker recognition means, and speaker storage means for storing the speaker recognized by the speaker recognition means together with the recognition frequency, and input based on the recognition frequency The event to be executed is changed according to the voice.

  The virtual pet device according to the present invention is a virtual pet device configured to respond to an external stimulus, and includes an image input means for inputting an image, and an image input from the image input means. Input person recognition means for recognizing the input person based on the input person, and input person storage means for storing the input person recognized by the input person recognition means together with the recognition degree, and input based on the recognition degree The event to be executed is changed in accordance with the image.

  Further, the recording medium storing the virtual pet control program according to the present invention recognizes the signal input from the stimulus input means as a stimulus, controls the response operation based on the recognition result, and the recognized stimulus. A program for controlling to identify an individual who has input the stimulus after analysis is stored.

The recording medium can store a program for controlling to recognize an audio signal as an input signal as a stimulus.
In addition, the recording medium can store a program for controlling to recognize an image signal as an input signal as a stimulus.

  In addition, the recording medium identifies an unspecified number of stimulus input persons by stimulus analysis, stores the recognition frequency for each input person corresponding to the input, and responds to the same input according to the input person and the recognition frequency It is possible to store a program that controls to change the operation.

  In addition, the recording medium has a combination of a plurality of preset basic actions and words corresponding to each action, and a new relationship with words corresponding to a series of actions combining the basic actions can be set. It is possible to store a program for controlling the program.

  The input is, for example, a word or the like if the input means is a voice input device, and a face image or the like if the input means is an image input device.

  FIG. 1 shows the appearance of an example of a virtual pet according to the present invention. As shown in this figure, the virtual pet of this embodiment is configured in the shape of a stuffed dog, and a main body (torso) 1 is provided with eyes 2, a mouth 3, a tail 4, and the like. In the main body 1, a control unit to be described later is incorporated. Further, an operation mechanism (not shown) such as moving the head of the dog or swinging the tail can be provided, and each mechanism can be controlled by the control unit.

  FIG. 2 is a block diagram illustrating a configuration of a control unit provided in the main body 1. As shown in this figure, the control unit in the present embodiment is composed of a voice recognition unit 20, an event execution unit 30, an image recognition unit 40, and the like with a control device 10 at the center.

  In addition to the CPU 11, the control device 10 includes a ROM, a RAM, and the like (not shown). In addition, a speaker database 12 and an image database 13 are provided in the control device 11. These databases can be stored in a memory or other storage means.

  The voice recognition unit 20 is provided with a dictionary 21 for voice recognition. The voice input from the voice input device (microphone) 22 is recognized. The voice recognition unit 20 is connected to the control device 10 and outputs the recognized voice to the control device 10 and receives a control signal from the control device 10. Note that the voice recognition dictionary 21 can be updated by a command from the control device 10. Or you may provide the dictionary 21 for speech recognition so that replacement | exchange is possible. The microphone 22 can be installed at an appropriate position of the stuffed dog in FIG. 1, but is preferably arranged around the head or the front chest.

  The event execution unit 30 stores an event list 31. The event list stores a list of events that can be executed by the virtual pet of this embodiment. As a specific example of the event, it is possible to set an appropriate event such as output of sound (imitation sound) or sound, or an operation via an operation mechanism (not shown), for example, moving the neck or shaking the tail.

  In the present embodiment, the image recognition unit 40 includes an image processing device 41. The image processing device 41 processes the image input from the image input device 42 and sends the output to the control device 10. As the image input device 42, for example, a CCD camera can be used and can be arranged at the position of the eye 2 in FIG.

By the way, the speaker database 12 in this embodiment is for identifying the operator who talks to the main body 1, and the configuration example of the speaker database is shown in Table 1 below.

  In this embodiment, the voice of the operator input from the voice input device 22 is recognized by the voice recognition unit 20, and as shown in Table 1, sex data by gender estimation, age data by age estimation (children or adults) ), Rank-specific data of similarity by comparison with the built-in dictionary 21, recognition frequency data of the corresponding phrase indicating the number of times of recognition of the phrase corresponding to the previously registered words (phrase 1 to phrase 5 here), and the speaker Date and time data representing the last recognized date and time are accumulated and constructed as a database. Table 1 also shows the corresponding image data.

Table 2 below shows a configuration example of an optional image database.

  The image database is constructed based on the image data sent from the image recognition unit 40 to the control device 10 by processing the image input from the image input device 42 with the image processing device 41. Here, as an example, it is composed of data such as the size of the speaker's face, the interval between eyes, the pattern of the relative relationship between eyes, nose and mouth. The speaker image data stored in the image database is stored as an image corresponding to each speaker in the speaker database.

  In the virtual pet according to the present embodiment, when image recognition is used, a specific partner can be recognized (recognized by an image) without being consciously input by the owner, for example, like a pet robot. When there is a real entity, it is possible to perform spontaneous actions from the virtual pet side, such as approaching the side. Furthermore, by combining inputs using switches, sensors, etc., similar to voice recognition, it memorizes the recognition frequency when there is an input, and makes a difference in the response of the virtual pet according to the contact frequency of the operator Is possible. For example, it is possible to make a difference in response, such as spontaneously approaching an operator with a high recognition frequency and barking with an operator with a low recognition frequency.

  By the way, in the virtual pet of this embodiment, in order to generate a specific relationship between the speaker (operator) and the virtual pet, that is, a feeling imitating the exchange between the real pet and the owner, the virtual pet is directed to the virtual pet. By identifying the voices of an unspecified number of speakers who talk to each other individually and determining the speaker with the highest recognition frequency, only the most frequently contacted (talked) speaker is responded. Hereinafter, control for generating a specific relationship between the speaker (operator) and the virtual pet will be described with reference to the flowchart of FIG.

  In the flowchart of FIG. 3, when a speaker speaks to a virtual pet, the voice of the speaker is sampled. That is, the analog sound input from the microphone 22 is subjected to digital conversion by performing preprocessing such as band limitation and noise reduction as necessary (step 1). Then, features necessary for speech recognition and speaker estimation are extracted (step 2). Next, speech recognition processing is performed based on the speech recognition dictionary 21 to determine whether or not the recognition result is valid (step 3). The determination of validity / invalidity here is performed by checking whether or not a valid word / phrase corresponding to the event list 31 is input. If it is determined that the word is invalid, an event for informing the speaker that the word is invalid is executed, and then an input of a new word is awaited. As an event corresponding to an invalid word / phrase, for example, there is an operation of “raising the neck”.

  From the feature data extracted in step 2, estimation of gender and age as a speaker specific reference is performed to calculate a total similarity, which is used as speaker estimation data (step 4). The calculated speaker estimation data is compared with reference to the speaker database 12, and if the speaker does not correspond to an existing speaker in the database, the speaker estimation data is registered in the speaker database 12 as a new speaker. (Step 5).

  Then, the recognition frequency of the corresponding phrase (the phrase corresponding to the spoken word) of the corresponding speaker (new speaker in the case of a new speaker) in the speaker database 12 is updated (step 6).

  If the recognized corresponding phrase satisfies the event generation condition, an event corresponding to the recognized phrase is executed (step 7). Here, as the event occurrence conditions, (1): when the recognition frequency reaches a certain threshold or more, (2): when the maximum recognition frequency is reached among the speakers for each phrase, etc. Set accordingly.

  In the case of the speaker database shown in Table 1 above, when the threshold value is 10 under condition (1), phrase 1, phrase 2 and phrase 4 are valid for speaker 1 (event occurrence condition is met). The phrases 1 and 5 are valid for the speaker 2, and the event of only the phrase 3 is valid for the speaker 3. Similarly, in the case of condition (2), speaker 1 for word 1, speaker 1 for word 2, speaker 3 for word 3, speaker 1 for word 4, word 5 For speaker 2, speaker 2 is enabled.

  These conditions can be set for each word, common to all words, or set to be valid (event occurrence) only when both conditions are satisfied at the same time. .

  Furthermore, since the last recognition date and time for each speaker is recorded in the speaker database 12, the behavior of the virtual pet is changed (the event is changed) when a long time has elapsed with respect to the current time. Etc. can also be set.

  For example, if the recognition frequency of the input person is high overall, normal operation is performed when the elapsed time from the last recognition date is short, and excessive operation when the owner returns home from a short trip when it is long, When it is extremely long, alert actions such as barking and roaring as if the owner has been forgotten.

  Or, when the recognition frequency of the input person is generally low, and when the elapsed time from the last recognition date and time is short, the normal operation is performed as in the above case (when the recognition frequency of the input person is high overall). When the elapsed time is long, a warning operation is performed as if the owner has been forgotten earlier than in the above case, and when the elapsed time is extremely long, an operation of ignoring the input (non-operation) is performed.

In addition, when the input person's recognition frequency is generally high, when the no-input state continues to some extent, it may be operated spontaneously, for example, it sounds lonely.
Note that the final recognition date and time may be stored for each word.

  Also, different events can be set to be executed for the same word or phrase depending on the recognition frequency of the word or the input person. For example, when a low-recognition speaker speaks "Good morning", only "One" is answered, and when a high-recognition speaker speaks "Good morning" "Event" and "One, One, One" can be set.

  Since the recognition frequency area of the speaker database 12 is limited, it will eventually overflow, and if (1) is set as the condition in step 7, no event will actually occur. Also, if all the words exceed the threshold, any person's words (except for condition (2)) will be reacted, and the specific relationship between the speaker and the pet may be lost. To avoid this situation, readjust the recognition frequency of each of these valid events (the recognition frequency of the word that generated the valid event), so that those with a relatively low recognition frequency become invalid again. Alternatively, the speaker database 12 is adjusted such as threshold adjustment (step 8), and the next voice input is awaited.

  By the way, the virtual pet of this embodiment is configured as a dog, and the following correspondence is set in advance as a combination of a basic operation as a pet and a recognition phrase. Phrase: “Hand” = Action: “Raise Hand (Forefoot)” Phrase: “Sit” = Action: “Sit” Phrase: “Turn” = Action: “Turn” Phrase: “One” = Action: “ bark"

  Further, by configuring the correspondence between a plurality of combinations of these basic operations and new words (words), it becomes possible to perform more complicated operations. For example, with respect to the phrase “turn around 3 times”, the “turn” action can be performed after the “turn around” action is performed 3 times. The creation of such new correspondences (words and movements) is not a task of simply setting movements on a traditional dog stuffed toy, but a feeling as if a real pet is being prepared. Can be obtained.

  As mentioned above, although this invention was demonstrated by embodiment of illustration, this invention is not limited to this embodiment. For example, the appearance of a virtual pet is not limited to a dog, but can be arbitrarily set to animals such as cats and pandas and characters such as cartoons and animations. Moreover, it is possible to configure it as an electronic pet displayed on the display means rather than having a substance such as a stuffed toy. Of course, these electronic pets can take the form of operating on a dedicated game machine or a general-purpose personal computer. In that case, a recording medium storing a control program mainly based on the processing shown in the flowchart of FIG. 3, such as a CD-ROM or a cassette-type ROM, is set in a dedicated game machine or a general-purpose personal computer and stored. A game program can be executed.

  Furthermore, the stimulus given to the virtual pet is not limited to voice, for example, an image (such as an operator's face) by an image input device, contact using a microswitch / sensor, etc., or key input from an operation unit or remote control / It is possible to use various input signals through communication or the like.

  Further, the configuration content of the database for identifying the operator is not limited to the example of the present embodiment, and can be configured with arbitrary content. Of course, the setting of the recognition phrase is not limited. In addition, events executed in response to them are not limited to sound output and operation, and can be arbitrarily set such as light emission and oscillation. Furthermore, in the case of electronic pets, it is possible to consider the occurrence of events in various forms such as changes in form and color, growth, formation of friends and lovers, birth of children, and changes in the living environment of pets.

It is an external view which shows the virtual pet of one Embodiment of this invention. It is a block diagram which shows the structure of the control part provided in the virtual pet main body. It is a flowchart which shows the control for generating the specific relationship between an operator and a virtual pet in embodiment of invention.

Claims (13)

A virtual pet device configured to respond appropriately to external stimuli,
An input means for inputting an external stimulus, a recognition means for recognizing the stimulus input from the input means, and a control means for controlling the response operation of the apparatus based on the recognition result of the recognition means,
The virtual pet device characterized in that the control means analyzes the input stimulus and identifies an individual who has input the stimulus. The virtual pet device according to claim 1, wherein the input unit is a voice input device. The virtual pet apparatus according to claim 1, wherein the input unit is an image input apparatus. It is provided to be able to identify an unspecified number of stimulus input persons, and has a memory means for storing the recognition frequency for each input person corresponding to the input, and the response operation for the same input can be changed according to the input person and the recognition frequency The virtual pet device according to claim 1, wherein the virtual pet device is a device. It has a combination of a plurality of preset basic actions and words corresponding to each action, and a relationship with words corresponding to a series of actions combining the basic actions can be newly set. Item 3. The virtual pet device according to Item 2. A virtual pet device configured to respond appropriately to external stimuli,
An input means for inputting an external stimulus;
An input person recognition means for recognizing an input person based on a stimulus input from the input means;
An input person storage means for storing the input person recognized by the input person recognition means together with the recognition frequency;
An event to be executed corresponding to an input stimulus is changed based on the recognition frequency. A virtual pet device configured to respond to external stimuli,
Voice input means for inputting voice;
Speaker recognition means for recognizing a speaker based on the voice input from the voice input means;
Speaker storage means for storing the speaker recognized by the speaker recognition means together with the recognition frequency;
An event to be executed in response to an input voice is changed based on the recognition frequency. A virtual pet device configured to respond to external stimuli,
An image input means for inputting an image;
An input person recognition means for recognizing an input person based on an image input from the image input means;
An input person storage means for storing the input person recognized by the input person recognition means together with the recognition frequency;
An event to be executed corresponding to an input image is changed based on the recognition frequency. A recording medium storing a control program for a virtual pet that responds appropriately to external stimuli,
Recognizing the signal input from the stimulus input means as a stimulus, controlling the response action based on the recognition result,
A recording medium storing a program for controlling to identify the individual who has input the stimulus by analyzing the recognized stimulus. The recording medium according to claim 9, wherein a program for controlling to recognize a voice signal as the input signal as a stimulus is stored. The recording medium according to claim 9, wherein a program for controlling to recognize an image signal as the input signal as a stimulus is stored. The stimulus analysis identifies an unspecified number of stimulus input persons, stores the recognition frequency for each input person corresponding to the input, and controls to change the response operation for the same input according to the input person and the recognition frequency. The recording medium according to claim 9, wherein the program is stored. A program that has a combination of a plurality of preset basic actions and words corresponding to each action, and that controls to newly set a relationship with words corresponding to a series of actions that combine the basic actions is stored The recording medium according to claim 10.

Images