JP2008129961A - Examination result recording program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device, a method, and the like for allowing an examination operator to perform an examination and recording by himself/herself in dental examination for improving efficiency of examination processing. <P>SOLUTION: When executing input to a storage device or a storage program for storing dental examination results of respective examination portions is carried out in a computer, conditions of the storage device or the storage program are monitored. Based on the conditions obtained by monitoring, a voice recognition processing method is decided. According to the decided processing method, voice inputted via a voice input device is recognized, and the recognized recognition result is inputted as the examination result into the storage device or the storage program. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a technique for improving convenience when information on teeth is recorded for each tooth in tooth inspection.

  Dental examinations such as dental examinations and periodontal disease examinations are generally carried out by two persons, a person who conducts examination (examiner) and a person who records (recorder). The examiner observes the inside of the oral cavity of the person undergoing the examination (inspected person) and verbally conveys information on each tooth to the record person. Then, the recordr records information about each tooth of the examinee according to the information described by the examiner.

As a method of recording by this recorder, there is a method of writing on a dedicated sheet, but in recent years, a method of inputting to a computer is common for the convenience of digitized data. As a specific example, Patent Document 1 (particularly, paragraphs 0035 to 0037) includes a tooth designating unit for designating which of the 32 teeth provided in the oral cavity is a data input on the display, and a mouse or keyboard is used. A technique for specifying a tooth as a data input target and inputting data by using it is disclosed.
JP 2002-083044 A

  However, the inspection by the inspector and the recordr can realize the efficiency of the inspection, but on the other hand, it increases the cost due to the labor cost of the recordr. For this reason, it can be considered that the inspector himself / herself makes the recording, thereby eliminating the need for the recording person and reducing the cost. However, this method has some problems.

  First, there are problems with the hands of the inspector. Since the object to be inspected is in the oral cavity of the person to be inspected, consideration is given to hygiene when performing the inspection. Specifically, the inspector is required to perform inspection by putting on a sanitary glove such as a vinyl glove. Furthermore, in order to accurately grasp the condition of the teeth in the oral cavity, the inspector carries out an inspection with a tool such as a mirror (dental mirror). For this reason, it is difficult to input quickly and accurately due to the fact that the hand of the inspector is blocked by the tool and wearing gloves.

  Secondly, there are problems regarding the posture and line of sight of the examiner. Since the examiner is in a posture looking into the oral cavity of the examinee, it is difficult to input to the computer in the same posture. In order to input, it was necessary to turn the posture toward the computer by twisting the state of the body. For this reason, it is difficult to perform quick and accurate input.

  SUMMARY OF THE INVENTION An object of the present invention is to provide an apparatus, a method, and the like that enable an inspector to perform inspection and recording independently in tooth inspection and to improve the efficiency of the processing.

  The present invention is a program for causing a computer to perform an input to a recording device or a recording program for recording an inspection result of each inspection site of a tooth, and includes a monitoring step, a determination step, a voice recognition step, an input A program for causing a computer to execute each step. The monitoring step is a step of monitoring the state of the recording device or the recording program. The determining step is a step of determining a speech recognition processing method based on the state obtained by the monitoring step. The speech recognition step is a step of recognizing speech input via the speech input device by the processing method determined by the determination step. The input step is a step of inputting the recognition result recognized in the voice recognition step as a test result to the recording device or the recording program.

  According to the program of the present invention configured as described above, the inspector can record the inspection result only by the operation of speaking the inspection result and the content of the instruction without operating the input device. . This eliminates the need for the recording person and makes it possible to reduce costs.

  Further, the speech recognition processing method is determined according to the state obtained by the monitoring step. The determination of the speech recognition processing method may be realized, for example, by selecting and determining a plurality of acoustic models and language models used for the speech recognition processing, You may implement | achieve by selecting and determining from several. With this configuration, it is possible to improve the accuracy of voice recognition processing in each state.

  The present invention may be specified as each device included in the authentication system described above. The present invention may also be specified as a program for operating a plurality of information processing apparatuses configured to be communicable via a network as the authentication system described above. In addition, the present invention may be specified as a program for causing an information processing apparatus to operate as each apparatus included in the authentication system described above. The present invention may be specified as an authentication method performed by the above-described authentication system. In addition, the present invention may be specified as a method executed by each device included in the authentication system described above.

  According to the present invention, in an inspection of a tooth, an inspector can inspect and record independently, and further improve the efficiency of the processing.

〔System configuration〕
FIG. 1 is a diagram illustrating a configuration example of a periodontal inspection result input system 100 according to the present invention. The periodontal examination result input system 100 includes a voice input device 10, a voice output device 20, an input device 30, and a test result input device 40.

  The periodontal examination input system 100 is used in a scene where a periodontal examination is performed. A person who performs a periodontal examination (hereinafter referred to as “inspector”) causes the periodontal examination input system 100 to record the examination result by speaking the result of the periodontal examination. That is, in the periodontal examination result input system 100, the examiner inputs the examination result by voice using the voice input device 10. Then, based on the input voice data, the test result input device 40 performs voice recognition and records the test result. Further, the voice output device 20 outputs the result of voice recognition by the test result input device 40 as a voice for confirmation by the inspector. Hereinafter, each device included in the periodontal examination result input system 100 will be described.

<Voice input device>
The voice input device 10 is a device that converts the voice uttered by the examiner into an electrical signal. The voice input device 10 can be configured using a microphone, and the type thereof is not limited. For example, a microphone such as an inner ear microphone, a bone conduction microphone, a pharyngeal microphone, a headset microphone, or a desktop microphone can be used.

  The voice data converted into an electrical signal by the voice input device 10 is passed to the test result input device 40. At this time, the data delivery method may be realized in any manner regardless of wired or wireless.

<Audio output device>
The audio output device 20 converts audio data (electrical signal) delivered from the test result input device 40 into vibration and outputs audio. The audio output device 20 can be configured using a speaker, and the type thereof is not limited. For example, a speaker that vibrates air or a speaker that vibrates the human body (for example, a bone conduction speaker) can be used. Also, the shape is not limited, and it may be one that can be worn by an inspector (for example, a headphone or headset type), or one that can be installed (for example, a desktop type, a floor type, or a wall type). .

  The method when the audio output device 20 receives the audio data from the test result input device 40 may be realized in any manner regardless of wired or wireless.

<Input device>
The input device 30 is a device for inputting data to the test result input device 40 by a method other than voice input. The input device 30 can be configured using existing input devices such as buttons (keyboard, numeric keypad, etc.), pointing devices (mouse, tablet, etc.), jog dial, touch panel, etc., and the type is not limited. Data input via the input device 30 is passed to the inspection result input device 40. At this time, the data delivery method may be realized in any manner regardless of wired or wireless.

<Inspection result input device>
The examination result input device 40 is a device that records examination results of periodontal examination based on voice data input via the voice input device 10 and data input via the input device 30. The inspection result input device 40 can be configured using an information processing device. Specifically, the inspection result input device 40 includes a CPU (Central Processing Unit), a main storage device, an input / output interface, an auxiliary storage device, and the like connected via a bus. The test result input device 40 operates as a device including a speech recognition unit 41, a speech synthesis unit 42, a result input unit 43, and a storage unit 44 when the program is executed by the CPU. Hereinafter, each functional unit included in the inspection result input device 40 will be described.

<Voice recognition unit>
The voice recognition unit 41 is realized by executing a voice recognition program by the CPU. The voice recognition unit 41 performs voice recognition using voice data input from the voice input device 10 as an input signal. Any conventional speech recognition technology may be applied to the speech recognition unit 41. Below, the specific example of the technique applicable to the speech recognition part 41 is shown.

  The voice recognition unit 41 has an acoustic model and a language model in a storage device (not shown). The acoustic model is obtained by associating a feature amount of an audio signal generated by a person's utterance with a word corresponding to the utterance. For example, an acoustic model can be created by acquiring a feature amount of a speech signal of a specific word pronounced by a human in advance and associating the feature amount with the specific word. The correspondence between the feature quantity and the wording may be one-to-one or one-to-multiple. In the latter case, it is also effective to weight each wording.

  The language model is data representing the appearance probability of each wording. This appearance probability may be given simply as the probability of appearance regardless of the presence of other words. Further, the appearance probability may be given as a probability of appearing next to the word on the assumption that a word is recognized. By using this language model, the recognition efficiency can be improved as compared with the case of voice recognition based only on the acoustic model.

  Next, the process of the voice recognition unit 41 will be described. First, the speech recognition unit 41 extracts a feature amount effective for speech recognition from the input speech signal by performing speech analysis on the input speech signal. Next, the speech recognition unit 41 matches the feature amount obtained by speech analysis with the acoustic model under the constraint of the language model, and determines which speech signal the speech signal input represents. To do.

  When the recognition result (numerical value, character string, etc.) is obtained, the voice recognition unit 41 passes a command based on the recognition result to the result input unit 43. In this case, the voice recognition unit 41 passes a numerical value or a character string as a recognition result to the voice synthesis unit 42 and causes the voice output device 20 to execute voice output. Hereinafter, such processing is referred to as talkback processing. On the other hand, when the recognition result cannot be obtained (in the case of a recognition error), the speech recognition unit 41 causes the speech synthesis unit 42 to output an error.

  The speech recognition unit 41 has a plurality of models (language model, acoustic model). The model used for the speech recognition processing is changed according to the state of the result input unit 43 (input mode, screen state, etc.) and past recognition results. For example, when the state of the result input unit 43 is the inspection result input state, a model associated with the inspection result input state is used. Further, when the state of the result input unit 43 is the name input state of the examinee, a model associated with the name input state is used. When the result input unit 43 is in the inspection result input state and the recognition result “inspection start” is obtained, it is associated with the combination of the inspection result input state and the keyword “inspection start”. The model is used in the subsequent speech recognition process. Such correspondence is stored in a storage device (not shown). The details of the operation of the voice recognition unit 41 will be described in the column of an operation example.

<Speech synthesis unit>
The voice synthesizer 42 is realized by executing a voice synthesis program by the CPU. The speech synthesis unit 42 performs speech synthesis processing based on the recognition result received from the speech recognition unit 41. Any conventional speech synthesis technique may be applied to the speech synthesizer 42. A specific example of technology applicable to the speech synthesizer 42 is shown below.

  The voice synthesizer 42 has a word dictionary and a voice signal dictionary in a storage device (not shown). The word dictionary associates a plurality of words with their readings. The sound signal dictionary is a correspondence between each sound and its sound signal. The speech synthesizer 42 determines a reading corresponding to the recognition result based on the recognition result received from the speech recognition unit 41 and the word dictionary. Then, the speech synthesizer 42 uses the speech signal dictionary to create a speech signal corresponding to the reading that is the determination result.

  When the speech synthesis unit 42 receives a recognition result such as a numerical value or a character string from the speech recognition unit 41, the speech synthesis unit 42 creates a speech signal of the new result. On the other hand, when receiving an error output command from the speech recognition unit 41, the speech synthesis unit 42 creates a predetermined error speech signal.

  The voice synthesizer 42 passes the created voice signal to the voice output device 20 and causes the voice output device 20 to execute voice output.

<Result input section>
The result input unit 43 is realized by executing a result input program by the CPU. The result input unit 43 updates the recorded content of the storage unit 44 based on the recognition result received from the voice recognition unit 41 or the data received from the input device 30.

<< Storage section >>
The storage unit 44 stores the contents of the inspection result according to the processing by the result input unit 43. FIG. 2 is a diagram illustrating a specific example of the contents stored in the storage unit 44. Hereinafter, description will be made based on a specific example shown in FIG.

  The storage unit 44 stores an inspection result table for each subject. Each subject is distinguished by the subject ID. The inspection result table has a numerical value of the pocket depth in the inspection region for each inspection region ID. The inspection part ID indicates a part which is an inspection target in the oral cavity of the subject, and is expressed as “upper right 1”, “upper left 1”, or the like. The pocket depth is obtained by quantifying the depth of the periodontal pocket at the site corresponding to each examination site ID.

  Note that the specific example shown in FIG. 2 is merely an example, and need not be limited to this configuration. Therefore, the storage content of the storage unit 44 may take other configurations depending on the environment to which the periodontal examination result input system 100 is applied.

[Operation example]
FIGS. 3 and 4 are flowcharts showing an operation example of the periodontal examination result input system 100. Hereinafter, an operation example of the periodontal examination result input system 100 will be described using each flowchart.

  First, when the process of the test result input device 40 starts, the voice recognition unit 41 acquires the state of the result input unit 43 from the result input unit 43. Then, the speech recognition unit 41 determines a model to be used based on the state of the result input unit 43 (S01). In the following description, a case where the state of the result input unit 43 is a result input state will be described as an example.

  When the user utters (S02), the voice generated by the utterance is converted into an electrical signal by the voice input device. This electric signal is input to the voice recognition unit 41. The voice recognition unit 41 performs voice recognition processing on the voice data that is the input electrical signal (S03). In this speech recognition process, when a recognition result is not obtained (S04-NO), the speech recognition unit 41 instructs the speech synthesis unit 42 to output an error. Then, the voice synthesizer 42 performs error output by voice via the voice output device 20 (S05). And the process after S02 is repeated.

  In the speech recognition process, when a recognition result is obtained (S04-YES), the subsequent process differs depending on the numerical value or character string that is the recognition result. When the recognition result is “inspection start (Sensaikai)” (S06-inspection start), the speech recognition unit 41 causes the speech synthesis unit 42 to perform talkback (S07). Next, the speech recognition unit 41 changes the model used in the subsequent speech recognition processing based on the recognition result (S08). That is, the model is changed to a model associated with the combination of the inspection result input state and the keyword “inspection start”. And the process after S02 is repeated.

  When the recognition result is “examination completed (S06-examination)”, the speech recognition unit 41 causes the speech synthesis unit 42 to perform talkback (S09). Next, the speech recognition unit 41 changes the use model to a model associated with the combination of the inspection result input state and the keyword “end of inspection” (S10). And the process after S02 is repeated.

  When the recognition result is a numerical value such as “0” or “1”, an inspection item name such as “pocket” or “plaque”, or a part name such as “top right 5” (S06 numerical value / part name / inspection item) Name), the speech recognition unit 41 first cuts the recognition phrase into words (S11). Next, the speech recognition unit 41 inputs a processing command according to each segmented word (input of a pocket depth (numerical value) as an inspection result, movement of an input location to a position specified by a region name, inspection item The change of the inspection item specified by the name, the click of the command button, etc.) is passed to the result input unit 43 (S12). The process of S12 is repeated as many times as the number of isolated words (S13). Then, the result input unit 43 performs writing to the storage unit 44 in accordance with this command. After the repeating process of S13, the speech recognition unit 41 causes the speech synthesis unit 42 to perform talkback (S14). The process of S14 is repeated as many times as the number of words that have been cut (S15). Then, the voice recognition unit 41 determines whether or not the usage model needs to be changed based on the recognition result (S16). The change of the usage model is executed when the state of the result input unit 43 is changed (for example, when the screen for inputting the inspection result is switched) by a processing instruction according to the above word, for example. When the use model needs to be changed (S16-required), the speech recognition unit 41 changes the model to be associated with the combination of the test result input state and the recognition result keyword (S17). On the other hand, when it is not necessary to change the use model (S16-unnecessary) and after the process of S17, the processes after S02 are repeated.

  If the recognition result is “correction input” (S06—correction input), the speech recognition unit 41 causes the speech synthesis unit 42 to perform talkback (S18). Next, the speech recognition unit 41 changes the use model to a model associated with the combination of the test result input state and the keyword “correction input” (S19). Next, the voice recognition unit 41 stores the current input location of the test result (test site ID to which the test result is to be input) (S20). And the process after S02 is repeated.

  When the recognition result is “end of correction” (S06—end of correction), the speech recognition unit 41 causes the speech synthesis unit 42 to perform talkback (S21). Next, the speech recognition unit 41 changes the use model to a model associated with the combination of the test result input state and the keyword “correction completed” (S22). Next, the voice recognition unit 41 moves the current input location of the inspection result to the location stored in the process of S20 (S23). And the process after S02 is repeated.

  FIG. 5 is a flowchart showing an example of the monitoring process performed independently of the processes shown in FIGS. Hereinafter, the operation example of the monitoring process of the periodontal examination result input system 100 will be described using this flowchart.

  The monitoring process is periodically executed. The monitoring process is executed based on, for example, a certain time, a certain number of processes, a certain number of clocks, and the like. Then, when the execution of the monitoring process is started, the voice recognition unit 41 acquires the state from the result input unit 43 (S30). Next, the voice recognition unit 41 determines whether or not the usage model needs to be changed based on the acquired state (S31). This determination may be made based on, for example, whether or not the state acquired last time is temporarily held and whether the held state is the same as the newly acquired state. . In this case, if it is not the same, it is necessary to change the use model. When the usage model needs to be changed (S32-necessary), the speech recognition unit 41 changes the usage model based on the newly acquired state (S33). On the other hand, when it is not necessary to change the use model (S32—unnecessary) and after the process of S33, the speech recognition unit 41 periodically repeats the processes after S30.

[Action / Effect]
In the periodontal examination result input system 100, the examiner records the examination result only by the operation of speaking the examination result and the command content to the voice input device 10 without operating the input device 30 such as a keyboard. Can be done. Since it is not necessary to operate the input device 30 by hand, the inspector can quickly and accurately check the test result by speaking even if he / she has a tool such as a dental mirror or a glove in his / her hand. It becomes possible to input. Similarly, since it is possible to input by utterance without directing the posture toward the input device 30, problems related to the posture and line of sight are also solved. Therefore, by using the periodontal inspection result input system 100, it is not necessary to have a recorder in the inspection, and the cost can be reduced.

  The voice recognition unit 41 periodically monitors the state of the result input unit 43, and changes the acoustic model and language model used for the voice recognition process according to the state. The content that can be spoken by the examiner is limited for each state of the result input unit 43. For example, when the state of the result input unit 43 is a state in which a numerical value (for example, pocket depth) is input as an inspection result, there is no need for anything other than numerical values, “examination of inspection”, and “correction input”. In view of such features, the speech recognition unit 41 uses an acoustic model and a language model that are prepared only for data necessary for speech recognition of content that may be uttered in each state. . With this configuration, it is possible to improve the accuracy of the speech recognition processing of the speech recognition unit 41 in each state.

  In addition, when the examiner speaks the examination results for a plurality of examination parts together, data representing each examination result is recognized by the voice recognition unit 41. And the test result of a some test site | part is memorize | stored in the memory | storage part 44 at the same timing. For example, when the inspector speaks “1, 3, 0” when inputting the inspection result in the upper right 1, the voice recognition unit 41 recognizes three numerical values “1”, “3”, and “0”. To do. The voice recognition unit 41 passes the values “1”, “3”, and “0” to the result input unit 43 as the inspection results of the upper right 2 and the upper right 3 in order from the upper right 1. The result input unit 43 records these three inspection results. It is not necessary to be limited to the above three that can speak in a lump, and may be two or four or more. By being configured in this way, the inspector can input the inspection results collectively, and the efficiency of the inspection can be improved.

[Modification]
The periodontal inspection result input system 100 is a system used for inputting the inspection result of the periodontal inspection, but may be applied to the input of the inspection result of teeth other than the periodontal inspection. For example, finding input such as electronic medical records. In this case, it can be realized by changing the acoustic model and language model of the speech recognition unit 41 and the contents of the inspection result table of the storage unit 44 according to the application.

  The voice recognition unit 41 may include a plurality of voice recognition programs for realizing the voice recognition unit 41, and may be configured to change the voice recognition program instead of changing the model based on the state. . Further, the periodontal examination result input system 100 may be configured to include an image output device instead of the audio output device 20.

It is a figure which shows the structural example of the periodontal examination result input system in this invention. It is a figure which shows the specific example of the memory content of a memory | storage part. It is a flowchart which shows the operation example of the periodontal examination result input system in this invention. It is a flowchart which shows the operation example of the periodontal examination result input system in this invention. It is a flowchart which shows the operation example of the periodontal examination result input system in this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Voice input device 20 Voice output device 30 Input device 40 Test result input device 41 Speech recognition part 42 Speech synthesizer 43 Result input part 44 Storage part 100 Periodontal test result input system

Claims (1)

A program for causing a computer to execute an input to a recording device or a recording program for recording an inspection result of each inspection site of a tooth,
A monitoring step of monitoring the status of the recording device or recording program;
A determination step for determining a processing method of speech recognition based on the state obtained by the monitoring step;
A speech recognition step for recognizing speech input via a speech input device by the processing method determined in the determination step;
A program for causing a computer to execute an input step of inputting a recognition result recognized in the voice recognition step as an inspection result to the recording device or a recording program.